mirror of
https://github.com/storytold/gridcraft.git
synced 2026-11-12 08:53:03 -06:00
Compare commits
1
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
 Dylan M. Taylor
|
17cb3c83d9 |
@@ -27,10 +27,6 @@ jobs:
|
||||
name: Windows installation policy (no app build)
|
||||
runs-on: windows-latest
|
||||
steps:
|
||||
- name: Harden runner
|
||||
uses: step-security/harden-runner@1d88384fd173af6ecb012dfd0264a306b711241a # v2.22.4
|
||||
with:
|
||||
egress-policy: audit
|
||||
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
|
||||
with:
|
||||
persist-credentials: false
|
||||
|
||||
@@ -1,10 +1,6 @@
|
||||
name: Release
|
||||
|
||||
on:
|
||||
# A nightly dry run on main: every release build except macOS, nothing signed or released
|
||||
# (craft-common-actions release.yml `dry-run`). Finds release-only breakage before release day.
|
||||
schedule:
|
||||
- cron: "56 1 * * *"
|
||||
push:
|
||||
branches: [release]
|
||||
workflow_dispatch:
|
||||
@@ -27,7 +23,6 @@ jobs:
|
||||
# Intentionally tracked at @main (not SHA-pinned) so org-wide CI improvements, security fixes, and build optimizations propagate without PR churn across all craftapp repos.
|
||||
uses: storytold/craft-common-actions/.github/workflows/release.yml@main
|
||||
with:
|
||||
dry-run: ${{ github.event_name == 'schedule' }}
|
||||
app-name: gridcraft
|
||||
app-title: GridCraft
|
||||
craft-fonts-ref: b076417b5880bd4154bd22f00695b1d28c6bea83
|
||||
|
||||
@@ -12,7 +12,6 @@ Generated-in-code art is original and has no file to list: the UI icon set (`cra
|
||||
| `epaint_default_fonts::HACK_REGULAR` (dependency font input) | Source Foundry Authors (Hack; DejaVu and Bitstream Vera foundations) | [egui 6d518fc](https://github.com/storytold/egui/tree/6d518fc799681571963842f95b55a6b68431b15a/crates/epaint_default_fonts/fonts), [Hack](https://github.com/source-foundry/Hack) | MIT + Bitstream Vera License (full upstream texts in `NOTICE`) | Hack-Regular.ttf is egui's bundled monospace face and GridCraft's last-resort symbol fallback. Dependency asset unmodified here, no font file copied into this repository. |
|
||||
| `epaint_default_fonts::NOTO_EMOJI_REGULAR` (dependency font input) | The Noto Project Authors | [egui 6d518fc](https://github.com/storytold/egui/tree/6d518fc799681571963842f95b55a6b68431b15a/crates/epaint_default_fonts/fonts), [Noto Emoji](https://fonts.google.com/noto/specimen/Noto+Emoji) | SIL Open Font License 1.1 (full upstream text in `NOTICE`) | NotoEmoji-Regular.ttf is enabled by egui's `monochrome_emoji_fonts` feature and used as the monochrome emoji fallback. Dependency asset unmodified here, no font file copied into this repository. |
|
||||
| `epaint_default_fonts::UBUNTU_LIGHT` (dependency font input) | Canonical Ltd. and the Ubuntu Font Family authors | [egui 6d518fc](https://github.com/storytold/egui/tree/6d518fc799681571963842f95b55a6b68431b15a/crates/epaint_default_fonts/fonts), [Ubuntu font](https://design.ubuntu.com/font) | Ubuntu Font Licence 1.0 (full upstream text in `NOTICE`) | Ubuntu-Light.ttf is egui's default proportional face. Dependency asset unmodified here, no font file copied into this repository. |
|
||||
| Regional format data in `crates/core/src/region.rs` (data, no file of its own) | Unicode, Inc. (Unicode CLDR) | [cldr-json](https://github.com/unicode-org/cldr-json) (`cldr-numbers-full`, `cldr-dates-full`) and ICU's CLDR 48 data | Unicode License v3 (`Unicode-3.0`; full text in `NOTICE`) | Separators, month and weekday names, AM/PM words, currency symbols and date order of the built-in regions. No CLDR file is copied into this repository. |
|
||||
| `docs/brand/artcraft-logo.svg` | ArtCraft Team | ArtCraft brand kit | ArtCraft brand terms (`docs/brand/LICENSE-brand.txt`) | Trademark, not open source |
|
||||
| `docs/brand/artcraft-logo.png` | ArtCraft Team | ArtCraft brand kit | ArtCraft brand terms | Trademark |
|
||||
| `docs/brand/artcraft-logo-white.svg` | ArtCraft Team | ArtCraft brand kit | ArtCraft brand terms | Trademark |
|
||||
|
||||
Generated
+24
-60
@@ -910,12 +910,12 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-fonts-build"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
|
||||
[[package]]
|
||||
name = "craft-i18n"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
dependencies = [
|
||||
"fixed_decimal",
|
||||
"icu_decimal",
|
||||
@@ -938,7 +938,7 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-print"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
dependencies = [
|
||||
"craft-winprint-gdi",
|
||||
"log",
|
||||
@@ -952,7 +952,7 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-ui"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
dependencies = [
|
||||
"craft-fonts-build",
|
||||
"craft-i18n",
|
||||
@@ -972,7 +972,7 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-update"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
dependencies = [
|
||||
"chrono",
|
||||
"dunce",
|
||||
@@ -982,7 +982,6 @@ dependencies = [
|
||||
"image",
|
||||
"log",
|
||||
"plist",
|
||||
"rustix 1.1.5",
|
||||
"semver",
|
||||
"serde",
|
||||
"serde_json",
|
||||
@@ -1019,7 +1018,7 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-window"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=6e03f65bf45181db77ba6229716e6624db9eb25b#6e03f65bf45181db77ba6229716e6624db9eb25b"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=1b34eba41e62138fc2438d2734523b302840b45d#1b34eba41e62138fc2438d2734523b302840b45d"
|
||||
dependencies = [
|
||||
"egui",
|
||||
"objc2 0.6.5",
|
||||
@@ -2023,7 +2022,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"arboard",
|
||||
"chrono",
|
||||
@@ -2035,7 +2034,6 @@ dependencies = [
|
||||
"getrandom 0.4.3",
|
||||
"gridcraft-engine",
|
||||
"gridcraft-mcp",
|
||||
"gridcraft-sandbox",
|
||||
"gridcraft-ui-egui",
|
||||
"image",
|
||||
"log",
|
||||
@@ -2051,7 +2049,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-calc"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-core",
|
||||
"gridcraft-formula",
|
||||
@@ -2063,7 +2061,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-chart"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-calc",
|
||||
"gridcraft-core",
|
||||
@@ -2074,11 +2072,10 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-cli"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-engine",
|
||||
"gridcraft-mcp",
|
||||
"gridcraft-sandbox",
|
||||
"log",
|
||||
"quick-xml 0.41.0",
|
||||
"serde_json",
|
||||
@@ -2087,7 +2084,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-core"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"chrono",
|
||||
"js-sys",
|
||||
@@ -2098,7 +2095,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-engine"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-calc",
|
||||
"gridcraft-chart",
|
||||
@@ -2121,7 +2118,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-formula"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-core",
|
||||
"serde",
|
||||
@@ -2130,7 +2127,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-functions"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-core",
|
||||
"regex",
|
||||
@@ -2138,7 +2135,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-math"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"ratex-font",
|
||||
"ratex-layout",
|
||||
@@ -2148,7 +2145,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-mcp"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-engine",
|
||||
"log",
|
||||
@@ -2157,7 +2154,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-model"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-core",
|
||||
"gridcraft-formula",
|
||||
@@ -2170,29 +2167,18 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-numfmt"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-core",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-pdf"
|
||||
version = "0.19.1"
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-sandbox"
|
||||
version = "0.19.1"
|
||||
dependencies = [
|
||||
"landlock",
|
||||
"libc",
|
||||
"nix",
|
||||
"seccompiler",
|
||||
"thiserror 2.0.21",
|
||||
]
|
||||
version = "0.18.0"
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-ui-egui"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"craft-ui",
|
||||
"craft-update",
|
||||
@@ -2211,7 +2197,6 @@ dependencies = [
|
||||
"skrifa",
|
||||
"sys-locale",
|
||||
"toml",
|
||||
"unicode-bidi",
|
||||
"web-sys",
|
||||
"web-time",
|
||||
"zbus",
|
||||
@@ -2219,7 +2204,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-web"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"craft-web-api",
|
||||
"craft-web-host",
|
||||
@@ -2238,7 +2223,7 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "gridcraft-xlsx"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"aes",
|
||||
"gridcraft-core",
|
||||
@@ -2887,17 +2872,6 @@ dependencies = [
|
||||
"smallvec",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "landlock"
|
||||
version = "0.4.7"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "4cca98e95f35b29d469dade6724c6f96cec9236640f745a0e99b0334ec320ab1"
|
||||
dependencies = [
|
||||
"enumflags2",
|
||||
"libc",
|
||||
"thiserror 2.0.21",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "lazy_static"
|
||||
version = "1.5.1"
|
||||
@@ -4569,15 +4543,6 @@ dependencies = [
|
||||
"tiny-skia",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "seccompiler"
|
||||
version = "0.5.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "a4ae55de56877481d112a559bbc12667635fdaf5e005712fd4e2b2fa50ffc884"
|
||||
dependencies = [
|
||||
"libc",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "security-framework"
|
||||
version = "3.7.0"
|
||||
@@ -6520,7 +6485,7 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "winprint"
|
||||
version = "0.2.1"
|
||||
source = "git+https://github.com/storytold/winprint.rs?rev=bc06a7c48b768eb27a828c7986f6d2868b245ee0#bc06a7c48b768eb27a828c7986f6d2868b245ee0"
|
||||
source = "git+https://github.com/storytold/winprint.rs?rev=51934db382f225454cf152a8657eb03789be63c4#51934db382f225454cf152a8657eb03789be63c4"
|
||||
dependencies = [
|
||||
"flate2",
|
||||
"fmt-derive",
|
||||
@@ -6535,7 +6500,6 @@ dependencies = [
|
||||
"windows",
|
||||
"windows-core",
|
||||
"windows-implement",
|
||||
"windows-numerics",
|
||||
"xml",
|
||||
]
|
||||
|
||||
@@ -6710,7 +6674,7 @@ checksum = "e450f9b2ed1dff33c94c12589a87338689467b9c4f5d8a5710bd09a847d2c8a7"
|
||||
|
||||
[[package]]
|
||||
name = "xtask"
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
dependencies = [
|
||||
"gridcraft-engine",
|
||||
"serde_json",
|
||||
|
||||
+3
-4
@@ -4,7 +4,7 @@ members = ["crates/*", "apps/*", "xtask"]
|
||||
default-members = ["crates/*", "apps/gridcraft", "apps/gridcraft-cli", "xtask"]
|
||||
|
||||
[workspace.package]
|
||||
version = "0.19.1"
|
||||
version = "0.18.0"
|
||||
edition = "2024"
|
||||
license = "MIT OR Apache-2.0"
|
||||
rust-version = "1.95"
|
||||
@@ -14,7 +14,6 @@ repository = "https://github.com/storytold/gridcraft"
|
||||
# internal
|
||||
gridcraft-core = { path = "crates/core" }
|
||||
gridcraft-numfmt = { path = "crates/numfmt" }
|
||||
gridcraft-sandbox = { path = "crates/sandbox" }
|
||||
gridcraft-formula = { path = "crates/formula" }
|
||||
gridcraft-functions = { path = "crates/functions" }
|
||||
gridcraft-model = { path = "crates/model" }
|
||||
@@ -29,7 +28,7 @@ gridcraft-math = { path = "crates/math" }
|
||||
# Every craft-libs crate (craft-update and craft-window here, craft-ui in ui-egui, craft-print in
|
||||
# the app) comes from one craft-libs revision.
|
||||
# craft-update: app-specific pretend versions apply to startup and manual checks.
|
||||
craft-update = { git = "https://github.com/storytold/craft-libs", rev = "6e03f65bf45181db77ba6229716e6624db9eb25b", version = "=0.1.0", features = ["egui"] }
|
||||
craft-update = { git = "https://github.com/storytold/craft-libs", rev = "1b34eba41e62138fc2438d2734523b302840b45d", version = "=0.1.0", features = ["egui"] }
|
||||
|
||||
# foundation
|
||||
thiserror = "2"
|
||||
@@ -72,7 +71,7 @@ skrifa = { version = "0.44", default-features = false, features = ["std"] }
|
||||
epaint_default_fonts = "0.37"
|
||||
# One native window per workbook: the window list, Window menu titles and the immediate-viewport
|
||||
# driver. Same craft-libs revision as craft-update above.
|
||||
craft-window = { version = "0.1.0", git = "https://github.com/storytold/craft-libs", rev = "6e03f65bf45181db77ba6229716e6624db9eb25b", default-features = false, features = ["egui"] }
|
||||
craft-window = { version = "0.1.0", git = "https://github.com/storytold/craft-libs", rev = "1b34eba41e62138fc2438d2734523b302840b45d", default-features = false, features = ["egui"] }
|
||||
|
||||
[workspace.lints.rust]
|
||||
unsafe_code = "forbid"
|
||||
|
||||
@@ -275,55 +275,6 @@ DAMAGES, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
|
||||
FROM, OUT OF THE USE OR INABILITY TO USE THE FONT SOFTWARE OR FROM OTHER
|
||||
DEALINGS IN THE FONT SOFTWARE.
|
||||
|
||||
|
||||
The regional format data in crates/core/src/region.rs (separators, month and weekday names, AM/PM
|
||||
words, currency symbols and date order) follows the Unicode Common Locale Data Repository (CLDR),
|
||||
read from unicode-org/cldr-json and from ICU's CLDR 48 data, under the Unicode License v3.
|
||||
Source: https://github.com/unicode-org/cldr-json
|
||||
The following text is reproduced verbatim from cldr-json's LICENSE:
|
||||
|
||||
UNICODE LICENSE V3
|
||||
|
||||
COPYRIGHT AND PERMISSION NOTICE
|
||||
|
||||
Copyright © 2015-2024 Unicode, Inc.
|
||||
|
||||
NOTICE TO USER: Carefully read the following legal agreement. BY
|
||||
DOWNLOADING, INSTALLING, COPYING OR OTHERWISE USING DATA FILES, AND/OR
|
||||
SOFTWARE, YOU UNEQUIVOCALLY ACCEPT, AND AGREE TO BE BOUND BY, ALL OF THE
|
||||
TERMS AND CONDITIONS OF THIS AGREEMENT. IF YOU DO NOT AGREE, DO NOT
|
||||
DOWNLOAD, INSTALL, COPY, DISTRIBUTE OR USE THE DATA FILES OR SOFTWARE.
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a
|
||||
copy of data files and any associated documentation (the "Data Files") or
|
||||
software and any associated documentation (the "Software") to deal in the
|
||||
Data Files or Software without restriction, including without limitation
|
||||
the rights to use, copy, modify, merge, publish, distribute, and/or sell
|
||||
copies of the Data Files or Software, and to permit persons to whom the
|
||||
Data Files or Software are furnished to do so, provided that either (a)
|
||||
this copyright and permission notice appear with all copies of the Data
|
||||
Files or Software, or (b) this copyright and permission notice appear in
|
||||
associated Documentation.
|
||||
|
||||
THE DATA FILES AND SOFTWARE ARE PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
|
||||
KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
||||
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF
|
||||
THIRD PARTY RIGHTS.
|
||||
|
||||
IN NO EVENT SHALL THE COPYRIGHT HOLDER OR HOLDERS INCLUDED IN THIS NOTICE
|
||||
BE LIABLE FOR ANY CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES,
|
||||
OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
|
||||
WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
|
||||
ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THE DATA
|
||||
FILES OR SOFTWARE.
|
||||
|
||||
Except as contained in this notice, the name of a copyright holder shall
|
||||
not be used in advertising or otherwise to promote the sale, use or other
|
||||
dealings in these Data Files or Software without prior written
|
||||
authorization of the copyright holder.
|
||||
|
||||
SPDX-License-Identifier: Unicode-3.0
|
||||
|
||||
Linux Wayland window decorations
|
||||
-------------------------------
|
||||
sctk-adwaita 0.10.1 (https://github.com/PolyMeilex/sctk-adwaita,
|
||||
|
||||
+4
-12
@@ -2,7 +2,7 @@
|
||||
|
||||
**Stage: alpha** · next: beta, ~25 points of "ready for real work" and ~160–240 h away
|
||||
|
||||
> **Last reviewed:** 2026-11-01 · **Last updated:** 2026-11-05 · **Change:** minor (CJK fallback text shares the Latin baseline; earlier: form controls: buttons, sliders and spin buttons; earlier: RefEdit reference boxes in dialogs: Goal Seek, Solver and Data Analysis take ranges by selection on the grid; earlier: Python mode in the formula bar; earlier: Python cells stored and kept in files; earlier: equations: Insert › Equation, editor, gridcraft-math; earlier: What-if and analysis: Solver Sensitivity report for Simplex LP and GRG Nonlinear; earlier: floating objects: one selection, eight resize handles, keyboard delete; earlier: Chart data tables: live source values beneath column, bar, line, area and compatible combo charts; legend keys and borders; XLSX round-trip and print/PDF; earlier: Formulas › Function Library as labelled buttons with category menus; Recently Used and Insert Function categories; XLSX workbook-view selection, bounded tab ratios and sheet-edit regressions; readiness estimates unchanged; earlier: CSV and text open detects each column's date order, so day-first dates are not silently swapped) · **Target:** Microsoft Excel (Microsoft 365)
|
||||
> **Last reviewed:** 2026-11-01 · **Last updated:** 2026-11-02 · **Change:** minor (form controls: buttons, sliders and spin buttons; earlier: RefEdit reference boxes in dialogs: Goal Seek, Solver and Data Analysis take ranges by selection on the grid; earlier: Python mode in the formula bar; earlier: Python cells stored and kept in files; earlier: equations: Insert › Equation, editor, gridcraft-math; earlier: What-if and analysis: Solver Sensitivity report for Simplex LP and GRG Nonlinear; earlier: floating objects: one selection, eight resize handles, keyboard delete; earlier: Chart data tables: live source values beneath column, bar, line, area and compatible combo charts; legend keys and borders; XLSX round-trip and print/PDF; earlier: Formulas › Function Library as labelled buttons with category menus; Recently Used and Insert Function categories; XLSX workbook-view selection, bounded tab ratios and sheet-edit regressions; readiness estimates unchanged; earlier: CSV and text open detects each column's date order, so day-first dates are not silently swapped) · **Target:** Microsoft Excel (Microsoft 365)
|
||||
|
||||
GridCraft aims at full Microsoft Excel parity, and at being better in places: faster, open (XLSX
|
||||
is the native format), fully scriptable by agents (MCP, control channel, CLI), and on the web.
|
||||
@@ -43,7 +43,8 @@ long `.xls` track, so about half parallelizes (≈30–50 wall-clock hours with
|
||||
|
||||
Why alpha: all six core workflows in the [alpha gate](docs/roadmap.md#alpha-gate) pass end to end on macOS with save and reopen, and the core workflow works end to end (open an XLSX, enter formulas and data, format, sort
|
||||
and filter, chart, pivot, print to PDF, save XLSX), but users still hit blocking gaps on real
|
||||
files: 100 MB workbooks don't open (#175), there's no `.xls` (#121), Devanagari text needs shaping verification (Chinese text still reported in #142), charts lose their type and options on round trip, and
|
||||
files: 100 MB workbooks don't open (#175), there's no `.xls` (#121), Arabic and Devanagari text
|
||||
isn't shaped (Chinese text still reported in #142), charts lose their type and options on round trip, and
|
||||
nobody has yet confirmed that Excel opens our output without a repair prompt across a real
|
||||
corpus. Beta needs those closed (see [beta gates](docs/roadmap.md#beta-gates)).
|
||||
|
||||
@@ -104,7 +105,7 @@ Localized German formulas and numbers are in PR #163. Detail: [localization-pari
|
||||
| Simplified Chinese | zh-Hans | menus only | ~12% |
|
||||
| Spanish | es | none | 0% |
|
||||
| Hindi | hi | none | 0% |
|
||||
| Arabic | ar | catalog pending; shared shaping/bidi works | 0% |
|
||||
| Arabic | ar | none (no RTL layout, no shaping) | 0% |
|
||||
| French | fr | none | 0% |
|
||||
| Portuguese | pt-BR | menus, File view, dialogs | ~43% |
|
||||
| Indonesian | id | none | 0% |
|
||||
@@ -148,9 +149,6 @@ Ranked; estimates in Opus 5.5 agent-hours. Detail and the full beta-gate list in
|
||||
| 2026-10-31 | CSV and text open detects each column's date order: a value that reads only day first (`31/05/2026`) makes the column day first, a value that reads only month first (`12/31/2026`) makes it month first, an all-ambiguous column keeps month first with a warning, and a column with both kinds stays text with a warning (#2765) |
|
||||
| 2026-10-31 | Recently Used lists the functions most recently entered in a formula from any source (ribbon, control channel, MCP, scripts), kept with the preferences and returned by `formulas.recentlyUsed`; the Insert Function dialog takes Excel's title and a working category list (Most Recently Used, All, each category) |
|
||||
| 2026-10-31 | Formulas › Function Library as a big Insert Function button, the AutoSum split button and three columns of labelled small buttons, each category book in its own colour; category menus end with Insert Function… on that category; More Functions cascades into the remaining categories; Home › Editing shares the AutoSum menu; the MR key tip opens Recently Used |
|
||||
| 2026-10-31 | System and embedded CJK fallback text shares the Latin baseline in cells, editors and window titles (#1265, #2621). Symbol and emoji alignment, manual row heights and clipping are unchanged |
|
||||
| 2026-10-26 | Arabic bidi display/editing preserve logical source text; per-cell reading direction is editable and survives undo, XLSX and conditional-format priority |
|
||||
| 2026-10-26 | Arabic text shows in cells and the interface through a system, fontconfig or craft-fonts Arabic face; workbook fonts keep their script fallbacks |
|
||||
| 2026-10-25 | AutoSum over a selection does what Excel does: a selected row gets one total to its right (in its last cell if that is empty), not a one-cell total under each cell; a block whose last column is empty gets each row's total there, one whose last row is empty each column's, and with both the corner gets the total of the block; otherwise totals go under each column as before |
|
||||
| 2026-10-24 | Macro interface ([macros.md](docs/macros.md#in-the-window)): an Automate › Macros group (Record Macro, which turns into Stop Recording, Use Relative References, Run Macro with recent macros); "Recording · N steps" with Stop in the status bar; a Record Macro dialog (checked name, free Ctrl+Shift shortcut, description); the Macros dialog lists macros beside Automate scripts and a VBA project, with Run, Step Into, Edit, Create, Delete and Options; Ctrl+Shift+letter macro shortcuts; questions that offer Run, Overwrite, Shift Down or Shift Right; a step editor with a breakpoint gutter, drag to reorder, a command picker, reference boxes that switch between relative tokens and absolute references, Inputs and Output tabs, a debug bar (Check, Run, Step, Continue, Run to Here, Stop with keep or undo) and grid overlays for inputs, writes, conflicts and the current step |
|
||||
| 2026-10-24 | Go To lists the named ranges of every sheet (`Sheet2!Name` for a name scoped to another sheet; hidden names, constants and formulas left out) and the references entered before, a double-click goes to an entry, and the Reference box takes several names or references separated by commas (`Price, Type`, `B14:C22,F19:G30`) |
|
||||
@@ -199,7 +197,6 @@ Ranked; estimates in Opus 5.5 agent-hours. Detail and the full beta-gate list in
|
||||
| 2026-10-13 | XLSX: `[Content_Types].xml`, `_rels/.rels`, the workbook rels and `xl/workbook.xml` are the first zip entries (document properties last), so libmagic and PHP `finfo` detect the file type even with large properties (#1090) |
|
||||
| 2026-10-13 | Conditional formatting: a rule gathers range statistics only when it reads them (top/bottom, above average, duplicates, scales, bars, icons) and borrows them instead of copying them per cell, so the per-frame lookup of an expression or cell-value rule no longer visits its whole range (#1098, part 1) |
|
||||
| 2026-10-12 | Next/previous sheet navigation stops at the first/last visible worksheet instead of activating a hidden boundary sheet; the current selection is retained |
|
||||
| 2026-10-12 | Number formatting: one table of built-in formats (xlsx); numbers, dates and General render with one scratch buffer instead of per-token strings; output unchanged, checked against digests recorded from the old code (#817) |
|
||||
| 2026-10-21 | Macro engine ([macros.md](docs/macros.md)): typed, declarative macros saved in the workbook (a JSON field; hidden names in xlsx); declared inputs checked with messages that name the cells; tokens (`Input.col(2)`, `Row`, `Output`, `Name:Total`, `{{…}}` in formulas); one undo step with full rollback; a footprint dry run with ask, fail, overwrite, shift down or shift right; a policy that keeps macros away from files and the app, with approval for macros in opened files; Repeat, If and Stop steps; a debugger; a relative recorder; `macro.*` commands beside the unchanged Automate scripts |
|
||||
| 2026-10-13 | Charts: work and memory are bounded by series x points (2,000,000 per chart: every series keeps its first points; 16 trendlines per chart), hairline columns and bars are merged per pixel column, series ranges are clipped before they are evaluated, the moving average is linear and rendering clean data no longer copies it (#1147) |
|
||||
| 2026-10-13 | Charts: one combo plot-kind rule for resolving and drawing; ticks of a scale off the step grid keep their own values (#1146) |
|
||||
@@ -267,7 +264,6 @@ Ranked; estimates in Opus 5.5 agent-hours. Detail and the full beta-gate list in
|
||||
| Date | Change | Summary |
|
||||
|---|---|---|
|
||||
| 2026-11-06 | trivial | Progress log: scroll positions follow their sheet through sheet inserts, deletes and moves (#3485) |
|
||||
| 2026-11-05 | minor | Reconcile upstream form-control progress with the CJK baseline update |
|
||||
| 2026-11-02 | minor | Form controls; catalog 277 / 292 → 281 / 296 (Developer › Controls); progress log row |
|
||||
| 2026-11-02 | minor | Progress log: RefEdit reference boxes in dialogs |
|
||||
| 2026-11-02 | minor | Progress log: Python mode in the formula bar |
|
||||
@@ -280,9 +276,6 @@ Ranked; estimates in Opus 5.5 agent-hours. Detail and the full beta-gate list in
|
||||
| 2026-10-31 | minor | File-page text fields keep the view open when Esc cancels an edit (#3035) |
|
||||
| 2026-10-31 | minor | Progress log: Recently Used and Insert Function categories |
|
||||
| 2026-10-31 | minor | Progress log: Function Library |
|
||||
| 2026-10-31 | minor | Progress log: shared CJK/Latin baselines in cells, editors and window titles |
|
||||
| 2026-10-26 | minor | Progress log: Arabic bidi display and editing |
|
||||
| 2026-10-26 | minor | Progress log: Arabic font fallback |
|
||||
| 2026-10-25 | minor | Progress log: AutoSum over a selected row, and a selection's empty last row and column |
|
||||
| 2026-10-24 | minor | Macro interface on the Automate tab; progress log row |
|
||||
| 2026-10-24 | minor | Macro engine and `macro.*` commands; catalog 275 / 291 → 276 / 292 (View › Macros › Record Macro); progress log row |
|
||||
@@ -313,7 +306,6 @@ Ranked; estimates in Opus 5.5 agent-hours. Detail and the full beta-gate list in
|
||||
| 2026-10-13 | trivial | Progress log: shared pictures written once on XLSX export (#985) |
|
||||
| 2026-10-13 | minor | Progress log: conditional-format lookups no longer scan the range for rules without statistics (#1098) |
|
||||
| 2026-10-12 | minor | Progress log: hidden-sheet navigation at workbook boundaries |
|
||||
| 2026-10-12 | minor | Numfmt cleanup: one built-in format table (gridcraft-xlsx), fewer allocations rendering numbers, dates and General, equality test against the previous output (#817); progress log row |
|
||||
| 2026-10-12 | minor | The control server requires an auth token and closes bad connections (#777); progress log row |
|
||||
| 2026-10-12 | minor | Rows auto-fit oversized fonts on open and whole-row/column formatting (#833); progress log row |
|
||||
| 2026-10-12 | minor | Progress log: multi-area row/column delete (#758) |
|
||||
|
||||
@@ -14,7 +14,6 @@ name = "gridcraft-cli"
|
||||
path = "src/main.rs"
|
||||
|
||||
[dependencies]
|
||||
gridcraft-sandbox = { workspace = true }
|
||||
gridcraft-engine = { workspace = true }
|
||||
gridcraft-mcp = { workspace = true }
|
||||
serde_json = { workspace = true }
|
||||
|
||||
@@ -214,7 +214,7 @@ pub fn main(args: &[String]) -> ExitCode {
|
||||
Err(_) => return usage(format!("--max-report needs a whole number, not `{n}`")),
|
||||
}
|
||||
}
|
||||
let Discovery { files, failures: mut reports, .. } = collect_inputs(&a.positional);
|
||||
let Discovery { files, failures: mut reports } = collect_inputs(&a.positional);
|
||||
reports.extend(files.iter().map(|p| {
|
||||
let path = p.to_string_lossy().into_owned();
|
||||
match gridcraft_engine::io::read_file(&path) {
|
||||
@@ -233,9 +233,6 @@ pub fn main(args: &[String]) -> ExitCode {
|
||||
#[derive(Default)]
|
||||
struct Discovery {
|
||||
files: Vec<PathBuf>,
|
||||
/// Identity (dev+inode) of each collected file, so a rescan reached by a shallower path doesn't add it twice.
|
||||
#[cfg(unix)]
|
||||
file_ids: std::collections::HashSet<(u64, u64)>,
|
||||
failures: Vec<FileReport>,
|
||||
}
|
||||
|
||||
@@ -260,11 +257,11 @@ impl Discovery {
|
||||
/// (`~$…`) are skipped.
|
||||
fn collect_inputs(inputs: &[String]) -> Discovery {
|
||||
let mut out = Discovery::default();
|
||||
let mut seen = std::collections::HashMap::new();
|
||||
let mut seen = std::collections::HashSet::new();
|
||||
for i in inputs {
|
||||
let p = PathBuf::from(i);
|
||||
if p.is_dir() {
|
||||
if !walk(&p, 0, false, &mut out, &mut seen) {
|
||||
if !walk(&p, 0, &mut out, &mut seen) {
|
||||
break;
|
||||
}
|
||||
} else if !out.add_file(p) {
|
||||
@@ -280,29 +277,19 @@ fn collect_inputs(inputs: &[String]) -> Discovery {
|
||||
out
|
||||
}
|
||||
|
||||
/// Returns false only when the global file budget stops traversal. `seen` holds the shallowest
|
||||
/// depth each directory was scanned from; `rescan` is set below a directory reached again from a
|
||||
/// shallower path (its depth-limited subtree may now fit), where already collected files are skipped.
|
||||
#[cfg_attr(not(unix), allow(unused_mut))]
|
||||
fn walk(dir: &Path, depth: usize, mut rescan: bool, out: &mut Discovery, seen: &mut std::collections::HashMap<(u64, u64), usize>) -> bool {
|
||||
// A directory is visited once by identity (dev+inode on unix) unless reached from a shallower
|
||||
// path: several links to the same directory are aliases, not exponential branches of the scan
|
||||
// (each directory is scanned at most MAX_DEPTH + 1 times). Other platforms keep the depth cap only.
|
||||
/// Returns false only when the global file budget stops traversal.
|
||||
fn walk(dir: &Path, depth: usize, out: &mut Discovery, seen: &mut std::collections::HashSet<u64>) -> bool {
|
||||
// A directory is visited once by identity (dev+inode on unix): several links to the same
|
||||
// directory are aliases, not exponential branches of the scan. Other platforms keep the
|
||||
// depth cap only.
|
||||
#[cfg(unix)]
|
||||
{
|
||||
use std::os::unix::fs::MetadataExt as _;
|
||||
if let Ok(m) = std::fs::metadata(dir) {
|
||||
let identity = (m.dev(), m.ino());
|
||||
// A directory past the depth limit is not scanned, so it is not recorded: asked for
|
||||
// by name later, it is walked from depth 0.
|
||||
match seen.get(&identity) {
|
||||
Some(&shallowest) if shallowest <= depth => return true,
|
||||
Some(_) => rescan = true,
|
||||
None => {}
|
||||
}
|
||||
if depth <= MAX_DEPTH {
|
||||
seen.insert(identity, depth);
|
||||
}
|
||||
if let Ok(m) = std::fs::metadata(dir)
|
||||
&& let Some(id) = m.dev().checked_mul(1 << 60).and_then(|d| d.checked_add(m.ino()))
|
||||
&& !seen.insert(id)
|
||||
{
|
||||
return true;
|
||||
}
|
||||
}
|
||||
if depth > MAX_DEPTH {
|
||||
@@ -337,20 +324,13 @@ fn walk(dir: &Path, depth: usize, mut rescan: bool, out: &mut Discovery, seen: &
|
||||
}
|
||||
};
|
||||
if metadata.is_dir() {
|
||||
if !walk(&p, depth + 1, rescan, out, seen) {
|
||||
return false;
|
||||
}
|
||||
} else if p.extension().and_then(std::ffi::OsStr::to_str).is_some_and(|e| e.eq_ignore_ascii_case("xlsx") || e.eq_ignore_ascii_case("xlsm")) {
|
||||
#[cfg(unix)]
|
||||
{
|
||||
use std::os::unix::fs::MetadataExt as _;
|
||||
if !out.file_ids.insert((metadata.dev(), metadata.ino())) && rescan {
|
||||
continue;
|
||||
}
|
||||
}
|
||||
if !out.add_file(p) {
|
||||
if !walk(&p, depth + 1, out, seen) {
|
||||
return false;
|
||||
}
|
||||
} else if p.extension().and_then(std::ffi::OsStr::to_str).is_some_and(|e| e.eq_ignore_ascii_case("xlsx") || e.eq_ignore_ascii_case("xlsm"))
|
||||
&& !out.add_file(p)
|
||||
{
|
||||
return false;
|
||||
}
|
||||
}
|
||||
true
|
||||
|
||||
@@ -11,8 +11,8 @@ fn discovery_reports_a_directory_read_error() {
|
||||
let path = std::env::temp_dir().join(format!("gridcraft-check-not-a-directory-{}", std::process::id()));
|
||||
std::fs::write(&path, "not a directory").unwrap();
|
||||
let mut discovery = Discovery::default();
|
||||
let mut seen = std::collections::HashMap::new();
|
||||
assert!(walk(&path, 0, false, &mut discovery, &mut seen));
|
||||
let mut seen = std::collections::HashSet::new();
|
||||
assert!(walk(&path, 0, &mut discovery, &mut seen));
|
||||
assert!(discovery.files.is_empty());
|
||||
assert_eq!(discovery.failures.len(), 1);
|
||||
assert_eq!(discovery.failures[0].path, path.to_string_lossy());
|
||||
@@ -31,47 +31,6 @@ fn discovery_reports_omissions_only_after_the_file_limit() {
|
||||
assert!(discovery.failures[0].error.as_ref().unwrap().contains("file limit"));
|
||||
}
|
||||
|
||||
/// A new directory under the temp dir: `create_dir` fails on a name somebody else made first.
|
||||
#[cfg(unix)]
|
||||
fn fresh_dir(tag: &str) -> PathBuf {
|
||||
let mut n = 0u32;
|
||||
loop {
|
||||
let dir = std::env::temp_dir().join(format!("gridcraft-check-{tag}-{}-{n}", std::process::id()));
|
||||
match std::fs::create_dir(&dir) {
|
||||
Ok(()) => return dir,
|
||||
Err(e) if e.kind() == std::io::ErrorKind::AlreadyExists => n += 1,
|
||||
Err(e) => panic!("{e}"),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// A directory named again by a deeper path, or by a link back to a parent, lists each workbook once.
|
||||
#[cfg(unix)]
|
||||
#[test]
|
||||
fn discovery_lists_each_workbook_once_whatever_path_reaches_it() {
|
||||
let dir = fresh_dir("once");
|
||||
let sub = dir.join("sub");
|
||||
std::fs::create_dir(&sub).unwrap();
|
||||
std::fs::write(dir.join("a.xlsx"), b"x").unwrap();
|
||||
std::fs::write(sub.join("b.xlsx"), b"x").unwrap();
|
||||
std::os::unix::fs::symlink(&dir, sub.join("loop")).unwrap();
|
||||
let names = |d: &Discovery| {
|
||||
let mut v: Vec<String> = d.files.iter().map(|p| p.file_name().unwrap().to_string_lossy().into_owned()).collect();
|
||||
v.sort();
|
||||
v
|
||||
};
|
||||
// The link back to the top is one directory, not another branch of the scan.
|
||||
let found = collect_inputs(&[dir.to_string_lossy().into_owned()]);
|
||||
assert_eq!(names(&found), ["a.xlsx", "b.xlsx"]);
|
||||
// `sub` named after its parent is reached from a shallower depth: scanned again, nothing listed twice.
|
||||
let found = collect_inputs(&[dir.to_string_lossy().into_owned(), sub.to_string_lossy().into_owned()]);
|
||||
assert_eq!(names(&found), ["a.xlsx", "b.xlsx"]);
|
||||
// Named first, the same `sub` is the shallowest path and the parent's scan skips it.
|
||||
let found = collect_inputs(&[sub.to_string_lossy().into_owned(), dir.to_string_lossy().into_owned()]);
|
||||
assert_eq!(names(&found), ["a.xlsx", "b.xlsx"]);
|
||||
let _ = std::fs::remove_dir_all(dir);
|
||||
}
|
||||
|
||||
fn at(a1: &str) -> CellRef {
|
||||
CellRef::parse(a1).unwrap()
|
||||
}
|
||||
|
||||
@@ -11,7 +11,6 @@
|
||||
//! gridcraft-cli functions [--json] [--search X] [--category C]
|
||||
//! gridcraft-cli mcp [--connect PORT] [--in FILE | --sample NAME]
|
||||
//! gridcraft-cli send <port> <method> [json]
|
||||
//! gridcraft-cli sandbox-check [--read PATH]... [--exec PATH]... [-- PROGRAM ARGS...]
|
||||
//! gridcraft-cli version | --version
|
||||
//! ```
|
||||
#![deny(clippy::unwrap_used, clippy::expect_used, clippy::panic, clippy::unimplemented, clippy::todo, clippy::unreachable)]
|
||||
@@ -31,7 +30,6 @@ macro_rules! outln {
|
||||
}
|
||||
|
||||
mod check;
|
||||
mod sandbox_check;
|
||||
|
||||
use gridcraft_engine::Session;
|
||||
use serde_json::{Value, json};
|
||||
@@ -51,8 +49,6 @@ const USAGE: &str = "usage:
|
||||
gridcraft-cli functions [--json] [--search X] [--category C]
|
||||
gridcraft-cli mcp [--connect PORT] [--in FILE | --sample NAME]
|
||||
gridcraft-cli send <port> <method> [json]
|
||||
gridcraft-cli sandbox-check [--read PATH]... [--exec PATH]... [-- PROGRAM ARGS...]
|
||||
whether this system confines programs (no program: a confined self-test)
|
||||
gridcraft-cli version";
|
||||
|
||||
/// stdout went away. A reader that stopped early (a closed pipe) ends the program quietly; any
|
||||
@@ -66,11 +62,6 @@ fn stdout_failed(e: std::io::Error) -> ! {
|
||||
}
|
||||
|
||||
fn main() -> ExitCode {
|
||||
// A confined launcher run (and its probes) comes first, before anything else starts.
|
||||
let raw: Vec<std::ffi::OsString> = std::env::args_os().skip(1).collect();
|
||||
if let Some(code) = gridcraft_sandbox::handle_args(&raw) {
|
||||
return ExitCode::from(u8::try_from(code).unwrap_or(1));
|
||||
}
|
||||
let args: Vec<String> = std::env::args().skip(1).collect();
|
||||
let rest = args.get(1..).unwrap_or_default();
|
||||
let r = match args.first().map(String::as_str) {
|
||||
@@ -88,7 +79,6 @@ fn main() -> ExitCode {
|
||||
Some("cat") => cat(rest),
|
||||
Some("run") => run(rest),
|
||||
Some("check") => return check::main(rest),
|
||||
Some("sandbox-check") => return sandbox_check::main(rest),
|
||||
Some("commands") => commands(rest),
|
||||
Some("functions") => functions(rest),
|
||||
Some("mcp") => mcp(rest),
|
||||
|
||||
@@ -1,98 +0,0 @@
|
||||
//! `gridcraft-cli sandbox-check [--read PATH]... [--exec PATH]... [-- PROGRAM ARGS...]`: whether
|
||||
//! this system confines programs, for a person or a reviewer checking a machine.
|
||||
//!
|
||||
//! Without a program it starts a confined copy of itself that tries what the sandbox must deny
|
||||
//! (reading outside the policy, writing, the network, starting a process) and what it must still
|
||||
//! allow (a thread, reading an allowed file), and prints one line per attempt. With a program it
|
||||
//! runs that program confined, with the given read and executable paths and the system's shared
|
||||
//! libraries, and passes its output and exit status through; its own lines then go to standard
|
||||
//! error. Every path is absolute. Exit status: 0 when every attempt is
|
||||
//! as required (or the program succeeded), 1 otherwise (a refused launch included), 2 when GridCraft
|
||||
//! has no sandbox for this kind of system.
|
||||
|
||||
use std::ffi::OsString;
|
||||
use std::path::PathBuf;
|
||||
use std::process::ExitCode;
|
||||
|
||||
use gridcraft_sandbox::{Canary, Policy, REFUSED_EXIT_CODE, REPORT_FLAG, Support};
|
||||
|
||||
pub fn main(args: &[String]) -> ExitCode {
|
||||
match run(args) {
|
||||
Ok(code) => code,
|
||||
Err(e) => {
|
||||
eprintln!("gridcraft-cli: sandbox-check: {e}");
|
||||
ExitCode::FAILURE
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn run(args: &[String]) -> Result<ExitCode, String> {
|
||||
let mut policy = Policy::new().with_system_libraries();
|
||||
let mut it = args.iter();
|
||||
let mut program: Option<(PathBuf, Vec<OsString>)> = None;
|
||||
while let Some(a) = it.next() {
|
||||
match a.as_str() {
|
||||
"--read" => policy = policy.read(it.next().ok_or("--read needs a path")?),
|
||||
"--exec" => policy = policy.exec(it.next().ok_or("--exec needs a path")?),
|
||||
"--" => {
|
||||
let p = it.next().ok_or("a program follows `--`")?;
|
||||
program = Some((PathBuf::from(p), it.by_ref().map(OsString::from).collect()));
|
||||
}
|
||||
other => return Err(format!("unknown option `{other}`")),
|
||||
}
|
||||
}
|
||||
// With a program, its standard output is passed through untouched, so this command's own
|
||||
// lines go to standard error; the self-test's report is the output, on standard output.
|
||||
let report = program.is_none();
|
||||
let say = |line: String| {
|
||||
if report {
|
||||
outln!("{line}");
|
||||
} else {
|
||||
eprintln!("{line}");
|
||||
}
|
||||
};
|
||||
match gridcraft_sandbox::support() {
|
||||
Support::Available(how) => say(format!("sandbox: available ({how})")),
|
||||
Support::Unsupported(why) => {
|
||||
say(format!("sandbox: unsupported ({why})"));
|
||||
return Ok(ExitCode::from(2));
|
||||
}
|
||||
}
|
||||
let canary = Canary::create().map_err(|e| e.to_string())?;
|
||||
let exe = std::env::current_exe().map_err(|e| e.to_string())?;
|
||||
let (program, program_args) = match program {
|
||||
Some((p, a)) => (p, a),
|
||||
// A confined copy of this program, reporting each attempt; it reads its own executable as
|
||||
// the allowed file.
|
||||
None => {
|
||||
let home = std::env::var_os("HOME").unwrap_or_default();
|
||||
(exe.clone(), vec![REPORT_FLAG.into(), canary.file().into(), exe.clone().into(), home])
|
||||
}
|
||||
};
|
||||
let policy = policy.exec(program.clone());
|
||||
let mut cmd = gridcraft_sandbox::launch_command(&policy, &canary, &program, &program_args).map_err(|e| e.to_string())?;
|
||||
let status = cmd.status().map_err(|e| format!("couldn't start the launcher: {e}"))?;
|
||||
drop(canary);
|
||||
match status.code() {
|
||||
Some(0) => {
|
||||
say(if report { "result: every attempt is as the sandbox requires" } else { "result: the program ran confined and succeeded" }.into());
|
||||
Ok(ExitCode::SUCCESS)
|
||||
}
|
||||
Some(REFUSED_EXIT_CODE) => {
|
||||
// The launcher's refusal and a program's own exit share the code; the launcher's reason,
|
||||
// when it refused, is the `gridcraft-sandbox: refused:` line above.
|
||||
say(format!(
|
||||
"result: exit code {REFUSED_EXIT_CODE}: the launcher refused to run the program (its reason is above), or the program itself exited with {REFUSED_EXIT_CODE}"
|
||||
));
|
||||
Ok(ExitCode::FAILURE)
|
||||
}
|
||||
other => {
|
||||
say(format!(
|
||||
"result: {} ({})",
|
||||
if report { "some attempts are not as the sandbox requires" } else { "the program failed" },
|
||||
other.map_or("no exit code".into(), |c| format!("exit code {c}"))
|
||||
));
|
||||
Ok(ExitCode::FAILURE)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -205,44 +205,6 @@ fn check_reports_a_depth_limited_scan_while_checking_readable_files() {
|
||||
let _ = std::fs::remove_dir_all(dir);
|
||||
}
|
||||
|
||||
#[cfg(unix)]
|
||||
#[test]
|
||||
fn a_directory_past_the_depth_limit_is_still_scanned_when_named_as_an_input() {
|
||||
let dir = tmpdir("check-depth-named");
|
||||
// The first folder past the limit (33 levels down): the scan from the top reports it unreached.
|
||||
let mut deep = dir.clone();
|
||||
for _ in 0..33 {
|
||||
deep.push("d");
|
||||
}
|
||||
std::fs::create_dir_all(&deep).unwrap();
|
||||
ok(&["run", "--out", deep.join("deep.xlsx").to_str().unwrap(), "--quiet"]);
|
||||
let out = run(&["check", dir.to_str().unwrap(), deep.to_str().unwrap(), "--json"]);
|
||||
let report: serde_json::Value = serde_json::from_slice(&out.stdout).unwrap();
|
||||
// The first scan reports the subtree it could not reach; the named directory is checked.
|
||||
assert_eq!(report["totals"]["opened"], 1, "{report}");
|
||||
let _ = std::fs::remove_dir_all(dir);
|
||||
}
|
||||
|
||||
#[cfg(unix)]
|
||||
#[test]
|
||||
fn a_directory_first_reached_at_the_depth_limit_is_completed_when_named_as_an_input() {
|
||||
let dir = tmpdir("check-depth-shallower");
|
||||
// The 32nd folder is scanned from the top, but its child (and the workbook in it) is past the limit.
|
||||
let mut limit = dir.clone();
|
||||
for _ in 0..32 {
|
||||
limit.push("d");
|
||||
}
|
||||
let child = limit.join("c");
|
||||
std::fs::create_dir_all(&child).unwrap();
|
||||
ok(&["run", "--out", limit.join("at_limit.xlsx").to_str().unwrap(), "--quiet"]);
|
||||
ok(&["run", "--out", child.join("below.xlsx").to_str().unwrap(), "--quiet"]);
|
||||
let out = run(&["check", dir.to_str().unwrap(), limit.to_str().unwrap(), "--json"]);
|
||||
let report: serde_json::Value = serde_json::from_slice(&out.stdout).unwrap();
|
||||
// Named from depth 0 the subtree is read: the workbook below the limit is checked, the one at it only once.
|
||||
assert_eq!(report["totals"]["opened"], 2, "{report}");
|
||||
let _ = std::fs::remove_dir_all(dir);
|
||||
}
|
||||
|
||||
/// A script line whose parameters are under a key `run` does not read used to run the command
|
||||
/// with no parameters at all, and the run still exited 0: three mistaken lines made Sheet2,
|
||||
/// Sheet3 and Sheet4 instead of the named sheets, and nothing said so (#2256).
|
||||
|
||||
@@ -1,4 +1,3 @@
|
||||
//! The crate's integration tests, one module per file, built as one test binary.
|
||||
mod cli;
|
||||
mod sandbox;
|
||||
mod text_to_columns_quotes;
|
||||
|
||||
@@ -1,384 +0,0 @@
|
||||
//! The confined launcher, through the real binary: what a confined program can't do, what it
|
||||
//! still can, and that a launch whose check fails runs nothing. Linux and macOS confine; other
|
||||
//! systems report that they can't.
|
||||
|
||||
use std::path::{Path, PathBuf};
|
||||
use std::process::{Command, Output};
|
||||
|
||||
/// Whether this machine confines programs: a supported system whose self-test passes. On a Linux
|
||||
/// or macOS machine that reports no sandbox (another architecture, no `sandbox-exec`) or whose
|
||||
/// kernel doesn't enforce Landlock, the confinement tests have nothing to check and pass with a
|
||||
/// note. On CI (`CI` set) the machine must confine, so a failure there is never skipped.
|
||||
fn confines() -> bool {
|
||||
static CONFINES: std::sync::OnceLock<bool> = std::sync::OnceLock::new();
|
||||
*CONFINES.get_or_init(|| {
|
||||
if !cfg!(any(target_os = "linux", target_os = "macos")) {
|
||||
return false;
|
||||
}
|
||||
let o = cli(&["sandbox-check"]);
|
||||
let out = text(&o);
|
||||
let missing = out.contains("sandbox: unsupported") || out.contains("Landlock is not enforced");
|
||||
if missing && std::env::var_os("CI").is_none() {
|
||||
eprintln!("this machine doesn't confine programs; skipping:\n{out}");
|
||||
return false;
|
||||
}
|
||||
true
|
||||
})
|
||||
}
|
||||
|
||||
fn cli(args: &[&str]) -> Output {
|
||||
Command::new(env!("CARGO_BIN_EXE_gridcraft-cli")).args(args).output().unwrap()
|
||||
}
|
||||
|
||||
fn text(o: &Output) -> String {
|
||||
format!("{}{}", String::from_utf8_lossy(&o.stdout), String::from_utf8_lossy(&o.stderr))
|
||||
}
|
||||
|
||||
/// A directory of the test's own in the temporary directory, with one file.
|
||||
fn scratch(name: &str) -> PathBuf {
|
||||
let dir = std::env::temp_dir().canonicalize().unwrap().join(format!("gridcraft-sandbox-it-{name}-{}", std::process::id()));
|
||||
std::fs::create_dir_all(&dir).unwrap();
|
||||
std::fs::write(dir.join("allowed.txt"), "allowed text\n").unwrap();
|
||||
dir
|
||||
}
|
||||
|
||||
/// `cat` as this system names it.
|
||||
fn cat() -> &'static str {
|
||||
if Path::new("/bin/cat").exists() { "/bin/cat" } else { "/usr/bin/cat" }
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_self_test_finds_every_denial() {
|
||||
// A home directory of the test's own, so the home-directory attempt always runs.
|
||||
let home = scratch("home");
|
||||
let o = Command::new(env!("CARGO_BIN_EXE_gridcraft-cli")).arg("sandbox-check").env("HOME", &home).output().unwrap();
|
||||
let out = text(&o);
|
||||
if !cfg!(any(target_os = "linux", target_os = "macos")) {
|
||||
assert_eq!(o.status.code(), Some(2), "{out}");
|
||||
assert!(out.contains("sandbox: unsupported"), "{out}");
|
||||
return;
|
||||
}
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
assert!(o.status.success(), "{out}");
|
||||
for line in [
|
||||
"denied: read the canary",
|
||||
"denied: change the canary's permissions",
|
||||
"denied: truncate the canary",
|
||||
"denied: write next to the canary",
|
||||
"denied: write in the temporary directory",
|
||||
"denied: TCP connection",
|
||||
"denied: start a process",
|
||||
"denied: signal another process",
|
||||
"denied: UDP socket",
|
||||
"denied: list the home directory",
|
||||
"works: start a thread",
|
||||
"works: read an allowed file",
|
||||
] {
|
||||
assert!(out.contains(line), "missing `{line}` in:\n{out}");
|
||||
}
|
||||
std::fs::remove_dir_all(home).unwrap();
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_confined_program_reads_only_what_it_is_given() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let dir = scratch("read");
|
||||
let outside = scratch("outside");
|
||||
let d = dir.to_string_lossy().into_owned();
|
||||
let allowed = dir.join("allowed.txt").to_string_lossy().into_owned();
|
||||
let o = cli(&["sandbox-check", "--read", &d, "--", cat(), &allowed]);
|
||||
assert!(o.status.success(), "{}", text(&o));
|
||||
// The program's output is passed through untouched; this command's own lines go to stderr.
|
||||
assert_eq!(String::from_utf8_lossy(&o.stdout), "allowed text\n", "{}", text(&o));
|
||||
assert!(String::from_utf8_lossy(&o.stderr).contains("result: the program ran confined"), "{}", text(&o));
|
||||
// The same program, a file outside the policy: the read is refused and nothing is printed.
|
||||
let other = outside.join("allowed.txt").to_string_lossy().into_owned();
|
||||
let o = cli(&["sandbox-check", "--read", &d, "--", cat(), &other]);
|
||||
assert!(!o.status.success(), "{}", text(&o));
|
||||
assert!(!String::from_utf8_lossy(&o.stdout).contains("allowed text"), "{}", text(&o));
|
||||
std::fs::remove_dir_all(dir).unwrap();
|
||||
std::fs::remove_dir_all(outside).unwrap();
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_launch_whose_check_fails_runs_nothing() {
|
||||
// A run that says it is already confined and isn't finds the canary readable and must refuse.
|
||||
// On macOS that is the second stage started without `sandbox-exec`; on Linux the launcher never
|
||||
// uses that flag, and the same refusal must hold.
|
||||
let canary = scratch("canary");
|
||||
let echo = if Path::new("/bin/echo").exists() { "/bin/echo" } else { "/usr/bin/echo" };
|
||||
let probe = canary.join("allowed.txt").to_string_lossy().into_owned();
|
||||
// A dynamically linked `echo` needs the system's shared libraries and its loader.
|
||||
// The second stage takes the process the first looked at: here, the parent of what `cli` starts.
|
||||
let mut launch = vec![
|
||||
"--sandbox-confined".to_string(),
|
||||
std::process::id().to_string(),
|
||||
"--exec".into(),
|
||||
echo.into(),
|
||||
"--exec".into(),
|
||||
env!("CARGO_BIN_EXE_gridcraft-cli").into(),
|
||||
];
|
||||
for lib in ["/lib", "/lib64", "/usr/lib", "/usr/lib64"].into_iter().filter(|d| Path::new(d).exists()) {
|
||||
launch.extend(["--exec".to_string(), lib.into()]);
|
||||
}
|
||||
launch.extend(["--probe".to_string(), probe.clone(), "--".into(), echo.into(), "the program ran".into()]);
|
||||
let o = cli(&launch.iter().map(String::as_str).collect::<Vec<_>>());
|
||||
let out = text(&o);
|
||||
assert_eq!(o.status.code(), Some(97), "{out}");
|
||||
assert!(!out.contains("the program ran"), "{out}");
|
||||
assert!(out.contains("gridcraft-sandbox: refused"), "{out}");
|
||||
assert!(out.contains("read the canary"), "the check refused, not the arguments: {out}");
|
||||
// Without that process, the second stage has nothing to try a signal on.
|
||||
let without: Vec<&str> = std::iter::once("--sandbox-confined").chain(launch[2..].iter().map(String::as_str)).collect();
|
||||
let o = cli(&without);
|
||||
assert_eq!(o.status.code(), Some(97), "{}", text(&o));
|
||||
assert!(text(&o).contains("needs the process"), "{}", text(&o));
|
||||
// The normal path: the launcher confines itself first, and the program runs.
|
||||
if confines() {
|
||||
launch.splice(0..2, ["--sandbox-launch".to_string()]);
|
||||
// The launcher may be under a folder in the list, on either system.
|
||||
let mut by_folder = launch.clone();
|
||||
by_folder[4] = Path::new(env!("CARGO_BIN_EXE_gridcraft-cli")).parent().unwrap().to_string_lossy().into_owned();
|
||||
let o = cli(&by_folder.iter().map(String::as_str).collect::<Vec<_>>());
|
||||
assert!(o.status.success(), "{}", text(&o));
|
||||
assert!(text(&o).contains("the program ran"), "{}", text(&o));
|
||||
let o = cli(&launch.iter().map(String::as_str).collect::<Vec<_>>());
|
||||
let out = text(&o);
|
||||
assert!(o.status.success() && out.contains("the program ran"), "{out}");
|
||||
}
|
||||
// A canary inside an allowed path is refused before anything runs, everywhere.
|
||||
let dir = canary.to_string_lossy().into_owned();
|
||||
let o = cli(&["--sandbox-launch", "--read", &dir, "--exec", echo, "--probe", &probe, "--", echo, "the program ran"]);
|
||||
assert_eq!(o.status.code(), Some(97), "{}", text(&o));
|
||||
assert!(!text(&o).contains("the program ran"));
|
||||
std::fs::remove_dir_all(canary).unwrap();
|
||||
}
|
||||
|
||||
/// A Python found on this system: the interpreter to run confined, and its installation, the path it
|
||||
/// needs to read. `None` when there is no `python3`.
|
||||
fn system_python() -> Option<(String, String)> {
|
||||
// On macOS `/usr/bin/python3` is a stub that starts Xcode's Python, outside any read path given
|
||||
// here; an installed Python comes first.
|
||||
let candidates = if cfg!(target_os = "macos") {
|
||||
["/opt/homebrew/bin/python3", "/usr/local/bin/python3"]
|
||||
} else {
|
||||
["/usr/bin/python3", "/usr/local/bin/python3"]
|
||||
};
|
||||
let python = candidates.into_iter().find(|p| Path::new(p).exists())?;
|
||||
let prefix = Command::new(python).args(["-c", "import sys; print(sys.base_prefix)"]).output().unwrap();
|
||||
let prefix = String::from_utf8_lossy(&prefix.stdout).trim().to_string();
|
||||
// A framework build's `bin/python3` only starts `Python.app`'s interpreter as a new process,
|
||||
// which the sandbox refuses; the interpreter itself is run instead.
|
||||
let framework = Path::new(&prefix).join("Resources/Python.app/Contents/MacOS/Python");
|
||||
let real = if framework.exists() { framework } else { std::fs::canonicalize(python).unwrap() };
|
||||
Some((real.to_string_lossy().into_owned(), prefix))
|
||||
}
|
||||
|
||||
/// Confined, a program can't signal another process (this test, signal 0 delivering nothing), but can
|
||||
/// still signal itself and its own threads. On Linux, where the filter names them, it also can't
|
||||
/// change another process's limits or priority, or type into a terminal: no call here reaches a real
|
||||
/// terminal, since its standard input isn't one.
|
||||
#[test]
|
||||
fn a_confined_program_leaves_other_processes_alone() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let Some((python, prefix)) = system_python() else {
|
||||
return;
|
||||
};
|
||||
let code = r#"
|
||||
import os, signal, sys, threading
|
||||
other = int(sys.argv[1])
|
||||
def refused(call, *args):
|
||||
try:
|
||||
call(*args)
|
||||
except PermissionError:
|
||||
return True
|
||||
except OSError as e:
|
||||
print("other error", call, args, e)
|
||||
return False
|
||||
assert refused(os.kill, other, 0), "a signal to another process"
|
||||
os.kill(os.getpid(), 0)
|
||||
seen = []
|
||||
signal.signal(signal.SIGUSR1, lambda *_: seen.append(1))
|
||||
signal.raise_signal(signal.SIGUSR1)
|
||||
worker = threading.Thread(target=lambda: signal.pthread_kill(threading.main_thread().ident, signal.SIGUSR1))
|
||||
worker.start()
|
||||
worker.join()
|
||||
assert len(seen) >= 1, "a signal to itself"
|
||||
if sys.platform == "linux":
|
||||
import ctypes, errno, fcntl, resource, termios
|
||||
libc = ctypes.CDLL(None, use_errno=True)
|
||||
assert libc.shmget(0, 4096, 0o600) == -1 and ctypes.get_errno() == errno.EPERM, "System V shared memory"
|
||||
assert libc.msgget(0, 0o600) == -1 and ctypes.get_errno() == errno.EPERM, "a System V message queue"
|
||||
assert libc.semget(0, 1, 0o600) == -1 and ctypes.get_errno() == errno.EPERM, "a System V semaphore"
|
||||
if hasattr(libc, "mq_open"):
|
||||
assert libc.mq_open(b"/gridcraft-check", os.O_RDONLY) == -1 and ctypes.get_errno() == errno.EPERM, "a POSIX message queue"
|
||||
# macOS answers these for the processes the profile lets it signal, which are only itself.
|
||||
assert refused(os.kill, -1, 0), "a signal to every process"
|
||||
assert refused(os.kill, 0, 0), "a signal to the process group"
|
||||
assert refused(resource.prlimit, other, resource.RLIMIT_NOFILE), "another process's limits"
|
||||
assert refused(os.setpriority, os.PRIO_PROCESS, other, 0), "another process's priority"
|
||||
assert refused(os.setpriority, os.PRIO_USER, 0, 0), "every process's priority"
|
||||
assert refused(os.sched_setaffinity, other, {0}), "another process's processors"
|
||||
assert refused(os.pidfd_open, other), "a handle on another process"
|
||||
resource.prlimit(0, resource.RLIMIT_NOFILE)
|
||||
resource.prlimit(os.getpid(), resource.RLIMIT_NOFILE)
|
||||
os.setpriority(os.PRIO_PROCESS, 0, os.getpriority(os.PRIO_PROCESS, 0))
|
||||
assert refused(fcntl.ioctl, 0, termios.TIOCSTI, b"x"), "typing into a terminal"
|
||||
assert refused(fcntl.ioctl, 0, termios.TIOCSTI | (1 << 32), b"x"), "the same with high bits set"
|
||||
assert refused(fcntl.ioctl, 0, 0x541C, b"x"), "TIOCLINUX"
|
||||
print("left alone")
|
||||
"#;
|
||||
let me = std::process::id().to_string();
|
||||
let o = cli(&["sandbox-check", "--read", &prefix, "--", &python, "-I", "-c", code, &me]);
|
||||
assert!(o.status.success(), "{}", text(&o));
|
||||
assert!(String::from_utf8_lossy(&o.stdout).contains("left alone"), "{}", text(&o));
|
||||
}
|
||||
|
||||
/// Confined, a program changes no file, whether the policy lets it read that file or not, by path or
|
||||
/// by a descriptor: not its contents, permissions, owner, extended attributes, times or flags.
|
||||
/// Skipped when there is no `python3`.
|
||||
#[test]
|
||||
fn a_confined_program_changes_no_file() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let Some((python, prefix)) = system_python() else {
|
||||
return;
|
||||
};
|
||||
let code = r#"
|
||||
import errno, fcntl, os, struct, sys
|
||||
readable, outside = sys.argv[1:3]
|
||||
def refused(what, call, *args):
|
||||
try:
|
||||
call(*args)
|
||||
except PermissionError:
|
||||
return
|
||||
except OSError as e:
|
||||
sys.exit(f"{what}: another error, not a refusal: {e}")
|
||||
sys.exit(f"{what}: it worked")
|
||||
fd = os.open(readable, os.O_RDONLY)
|
||||
for target in (readable, outside, fd):
|
||||
refused(f"permissions of {target}", os.chmod, target, 0)
|
||||
refused(f"owner of {target}", os.chown, target, os.getuid(), os.getgid())
|
||||
refused(f"times of {target}", os.utime, target, (0, 0))
|
||||
if sys.platform == "linux":
|
||||
refused(f"setting an attribute of {target}", os.setxattr, target, "user.gridcraft", b"1")
|
||||
refused(f"removing an attribute of {target}", os.removexattr, target, "user.gridcraft")
|
||||
for path in (readable, outside):
|
||||
refused(f"truncate {path}", os.truncate, path, 0)
|
||||
refused(f"open {path} with O_TRUNC", os.open, path, os.O_RDONLY | os.O_TRUNC)
|
||||
refused(f"open {path} to write", os.open, path, os.O_WRONLY)
|
||||
# A descriptor opened to read can't be truncated either way; the file's contents are checked after.
|
||||
try:
|
||||
os.ftruncate(fd, 0)
|
||||
except OSError:
|
||||
pass
|
||||
if sys.platform == "linux":
|
||||
import ctypes
|
||||
libc = ctypes.CDLL(None, use_errno=True)
|
||||
# Newer calls, which the kernel running this may not have: the filter answers before it would.
|
||||
for name, number in (("fchmodat2", 452), ("setxattrat", 463), ("removexattrat", 466), ("file_setattr", 469)):
|
||||
ctypes.set_errno(0)
|
||||
result = libc.syscall(number, -100, readable.encode(), 0, 0)
|
||||
assert result == -1 and ctypes.get_errno() == errno.EPERM, f"{name}: {result}, {ctypes.get_errno()}"
|
||||
# On a descriptor opened only to read, these change the file's flags and attributes.
|
||||
refused("FS_IOC_SETFLAGS", fcntl.ioctl, fd, 0x40086602, struct.pack("l", 0))
|
||||
refused("FS_IOC_FSSETXATTR", fcntl.ioctl, fd, 0x401C5820, bytes(28))
|
||||
refused("FS_IOC_SET_ENCRYPTION_POLICY", fcntl.ioctl, fd, 0x800C6613, bytes(12))
|
||||
# What a program still needs: its own descriptors' flags, and a file of its own in memory.
|
||||
os.set_inheritable(fd, True)
|
||||
os.set_inheritable(fd, False)
|
||||
os.set_blocking(fd, False)
|
||||
own = os.memfd_create("gridcraft-check")
|
||||
os.ftruncate(own, 16)
|
||||
assert os.fstat(own).st_size == 16
|
||||
print("unchanged")
|
||||
"#;
|
||||
let inside = scratch("unchanged-inside");
|
||||
let elsewhere = scratch("unchanged-outside");
|
||||
let (readable, outside) = (inside.join("allowed.txt"), elsewhere.join("allowed.txt"));
|
||||
let state = |file: &Path| {
|
||||
let m = std::fs::metadata(file).unwrap();
|
||||
(std::fs::read(file).unwrap(), m.permissions(), m.modified().unwrap())
|
||||
};
|
||||
let before = (state(&readable), state(&outside));
|
||||
let d = inside.to_string_lossy().into_owned();
|
||||
let (r, x) = (readable.to_string_lossy().into_owned(), outside.to_string_lossy().into_owned());
|
||||
let o = cli(&["sandbox-check", "--read", &prefix, "--read", &d, "--", &python, "-I", "-c", code, &r, &x]);
|
||||
assert!(o.status.success(), "{}", text(&o));
|
||||
assert!(String::from_utf8_lossy(&o.stdout).contains("unchanged"), "{}", text(&o));
|
||||
assert_eq!((state(&readable), state(&outside)), before);
|
||||
std::fs::remove_dir_all(inside).unwrap();
|
||||
std::fs::remove_dir_all(elsewhere).unwrap();
|
||||
}
|
||||
|
||||
/// A Python found on this system starts and imports its standard library while confined, with
|
||||
/// read access to its installation only. Skipped when there is no `python3`.
|
||||
#[test]
|
||||
fn python_starts_confined() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let Some((real, prefix)) = system_python() else {
|
||||
return;
|
||||
};
|
||||
let o =
|
||||
cli(&["sandbox-check", "--read", &prefix, "--", &real, "-I", "-c", "import json, math, datetime, decimal; print(json.dumps([math.pi > 3]))"]);
|
||||
assert!(o.status.success(), "{}", text(&o));
|
||||
assert!(String::from_utf8_lossy(&o.stdout).contains("[true]"), "{}", text(&o));
|
||||
// Confined Python can't open the network or a file outside its installation.
|
||||
let o = cli(&["sandbox-check", "--read", &prefix, "--", &real, "-I", "-c", "import socket; socket.create_connection(('127.0.0.1', 9), 2)"]);
|
||||
assert!(!o.status.success(), "{}", text(&o));
|
||||
assert!(text(&o).contains("PermissionError") || text(&o).contains("Operation not permitted"), "{}", text(&o));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_confined_program_inherits_no_secrets() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let env = if Path::new("/usr/bin/env").exists() { "/usr/bin/env" } else { "/bin/env" };
|
||||
let o = Command::new(env!("CARGO_BIN_EXE_gridcraft-cli"))
|
||||
.args(["sandbox-check", "--", env])
|
||||
.env("GRIDCRAFT_CONTROL_TOKEN", "secret-token")
|
||||
.env("AWS_SECRET_ACCESS_KEY", "secret-key")
|
||||
.env("LANG", "en_US.UTF-8")
|
||||
.output()
|
||||
.unwrap();
|
||||
let out = text(&o);
|
||||
assert!(o.status.success(), "{out}");
|
||||
assert!(!out.contains("secret-"), "the environment reached the program:\n{out}");
|
||||
assert!(out.contains("LANG=en_US.UTF-8"), "the locale is kept:\n{out}");
|
||||
assert!(out.contains("PATH=/usr/bin:/bin"), "{out}");
|
||||
}
|
||||
|
||||
/// A descriptor the parent leaves open (here fd 9, opened by a shell on a file outside any policy)
|
||||
/// would survive `exec` with access the sandbox doesn't revoke, so the launcher closes it: the
|
||||
/// program runs, and can't read through it. (A desktop app's system libraries leave such files open,
|
||||
/// so refusing to run would block Python there.)
|
||||
#[test]
|
||||
fn an_inherited_descriptor_never_reaches_the_program() {
|
||||
if !confines() {
|
||||
return;
|
||||
}
|
||||
let outside = scratch("leak");
|
||||
let leaked = outside.join("allowed.txt").to_string_lossy().into_owned();
|
||||
let cli = env!("CARGO_BIN_EXE_gridcraft-cli");
|
||||
// The shell opens fd 9 (not close-on-exec) and then becomes the command, which inherits it. The
|
||||
// confined shell then tries to read through it, with a built-in: starting `cat` is refused. Not
|
||||
// fd 3: on macOS the system's libraries hold a socket of their own there under the sandbox.
|
||||
let bash = if cfg!(target_os = "macos") { "--exec /bin/bash" } else { "" };
|
||||
let program = "read line <&9 && echo \"leaked: $line\"; echo finished";
|
||||
let script = format!("exec 9<'{leaked}'; exec '{cli}' sandbox-check {bash} -- /bin/sh -c '{program}'");
|
||||
let o = Command::new("/bin/sh").args(["-c", &script]).output().unwrap();
|
||||
let out = text(&o);
|
||||
assert!(out.contains("finished"), "the program ran:\n{out}");
|
||||
assert!(!out.contains("leaked:"), "the program read through the descriptor:\n{out}");
|
||||
std::fs::remove_dir_all(outside).unwrap();
|
||||
}
|
||||
@@ -13,7 +13,6 @@ workspace = true
|
||||
[dependencies]
|
||||
gridcraft-engine = { workspace = true }
|
||||
gridcraft-ui-egui = { workspace = true }
|
||||
gridcraft-sandbox = { workspace = true }
|
||||
eframe = { workspace = true, features = ["default_fonts", "wgpu", "accesskit", "wayland", "x11", "clipboard"] }
|
||||
egui = { workspace = true }
|
||||
serde_json = { workspace = true }
|
||||
@@ -35,7 +34,7 @@ winit = { version = "0.30", default-features = false, features = ["x11", "waylan
|
||||
[target.'cfg(not(target_arch = "wasm32"))'.dependencies]
|
||||
# Same craft-libs revision as the workspace's craft-update and craft-window (see the root
|
||||
# Cargo.toml).
|
||||
craft-print = { git = "https://github.com/storytold/craft-libs", rev = "6e03f65bf45181db77ba6229716e6624db9eb25b" }
|
||||
craft-print = { git = "https://github.com/storytold/craft-libs", rev = "1b34eba41e62138fc2438d2734523b302840b45d" }
|
||||
craft-update = { workspace = true }
|
||||
rfd = { workspace = true }
|
||||
arboard = { workspace = true, features = ["image-data"] }
|
||||
|
||||
@@ -241,12 +241,6 @@ fn icon() -> Option<egui::IconData> {
|
||||
}
|
||||
|
||||
fn main() -> eframe::Result<()> {
|
||||
// A confined launcher run (`gridcraft_sandbox::launch_command`) comes before everything else:
|
||||
// it opens no window, writes no log, and becomes the program it confines.
|
||||
let raw: Vec<std::ffi::OsString> = std::env::args_os().skip(1).collect();
|
||||
if let Some(code) = gridcraft_sandbox::handle_args(&raw) {
|
||||
std::process::exit(code);
|
||||
}
|
||||
// First, so the panic hook and every start-up warning are recorded (`logging`).
|
||||
let logger = logging::install();
|
||||
// Commands catch their panics (`Session::execute`), but the default hook only prints to
|
||||
|
||||
@@ -110,6 +110,3 @@ sync_github_name = true
|
||||
|
||||
[people.maxm11]
|
||||
sync_github_name = true
|
||||
|
||||
[people.RockerHX]
|
||||
sync_github_name = true
|
||||
|
||||
+8
-215
@@ -39,15 +39,6 @@ const MAX_DEPTH: usize = 400;
|
||||
/// every legitimate call; hostile work that grows past one dimension (a LAMBDA that recurses
|
||||
/// or nest `MAP` into `MAP`) then stops at the same bound instead of squaring.
|
||||
const LAMBDA_FUEL: u64 = MAX_CELLS;
|
||||
/// Work one cell's formula may do: one unit per expression evaluated, per cell read from a
|
||||
/// reference, per array value handed to a function, per pair of areas intersected and per
|
||||
/// character of a defined name's formula parsed where the name is used. Each limit
|
||||
/// above bounds one kind of step, but a compact formula could still repeat bounded steps
|
||||
/// exponentially often (defined names that each use the one before twice, a computed criterion
|
||||
/// whose database function reruns it, an aggregate over a large array inside every LAMBDA call).
|
||||
/// Past this the formula is `#CALC!`. It is far above what a formula that reads every cell of a
|
||||
/// full-size array a few times needs.
|
||||
const EVAL_WORK: u64 = 16 * MAX_CELLS;
|
||||
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
|
||||
pub struct Area {
|
||||
@@ -154,16 +145,6 @@ pub struct Evaluator<'h> {
|
||||
/// runs out of work after the same number of calls that a legitimately wide one could
|
||||
/// make, instead of an amount of work that squares with array nesting (#2382).
|
||||
fuel: u64,
|
||||
/// Work this cell still has left (see [`EVAL_WORK`]), shared by every branch of the formula
|
||||
/// and by the formulas it computes on the way (names, computed criteria).
|
||||
work: u64,
|
||||
/// A spend found too little work left. A spend that leaves exactly nothing is fine, so this
|
||||
/// can't be told from `work` alone; a database function checks it after its computed
|
||||
/// criteria ran, whose failures it can't otherwise see.
|
||||
exhausted: bool,
|
||||
/// Computed criteria being evaluated, outermost first: one that is reached again from inside
|
||||
/// itself counts as not met instead of repeating its work at every level.
|
||||
active_criteria: Vec<(usize, CellRef)>,
|
||||
}
|
||||
|
||||
struct FnCtx<'a, 'h> {
|
||||
@@ -193,9 +174,6 @@ impl Ctx for FnCtx<'_, '_> {
|
||||
fn computed_criterion(&mut self, row: usize, col: usize, db: &Array, record: usize) -> Option<bool> {
|
||||
let area = self.criteria?;
|
||||
let at = area.range.start.offset(row as i64, col as i64).filter(|c| area.range.contains(*c))?;
|
||||
if self.ev.active_criteria.contains(&(area.sheet, at)) {
|
||||
return Some(false);
|
||||
}
|
||||
let expr = self.ev.host.workbook().sheet(area.sheet)?.cell(at)?.formula.as_ref()?.expr_arc()?;
|
||||
let shifted = gridcraft_formula::adjust::shift_relative((*expr).clone(), record as i64 - 1, 0);
|
||||
let env = (0..db.cols)
|
||||
@@ -215,24 +193,12 @@ impl Ctx for FnCtx<'_, '_> {
|
||||
calls: self.ev.calls,
|
||||
stack_top: self.ev.stack_top,
|
||||
fuel: self.ev.fuel,
|
||||
work: self.ev.work,
|
||||
exhausted: false,
|
||||
active_criteria: std::mem::take(&mut self.ev.active_criteria),
|
||||
};
|
||||
ev.active_criteria.push((area.sheet, at));
|
||||
let v = matches!(ev.value(&shifted).scalar().to_bool(), Ok(true));
|
||||
// The child's spend counts against this cell's budget.
|
||||
self.ev.fuel = ev.fuel;
|
||||
self.ev.work = ev.work;
|
||||
self.ev.exhausted |= ev.exhausted;
|
||||
let mut active = std::mem::take(&mut ev.active_criteria);
|
||||
active.pop();
|
||||
self.ev.active_criteria = active;
|
||||
Some(v)
|
||||
}
|
||||
fn exhausted(&self) -> bool {
|
||||
self.ev.exhausted
|
||||
}
|
||||
fn has_lambda(&self) -> bool {
|
||||
self.lambda.is_some()
|
||||
}
|
||||
@@ -250,42 +216,7 @@ impl Ctx for FnCtx<'_, '_> {
|
||||
|
||||
impl<'h> Evaluator<'h> {
|
||||
pub fn new(host: &'h mut dyn Host, sheet: usize, at: CellRef) -> Self {
|
||||
Evaluator {
|
||||
host,
|
||||
sheet,
|
||||
at,
|
||||
env: Vec::new(),
|
||||
depth: 0,
|
||||
calls: 0,
|
||||
stack_top: 0,
|
||||
fuel: LAMBDA_FUEL,
|
||||
work: EVAL_WORK,
|
||||
exhausted: false,
|
||||
active_criteria: Vec::new(),
|
||||
}
|
||||
}
|
||||
|
||||
/// Spends `units` of this cell's work; `false` (and nothing left) when there isn't enough.
|
||||
pub(crate) fn spend(&mut self, units: u64) -> bool {
|
||||
match self.work.checked_sub(units) {
|
||||
Some(w) => {
|
||||
self.work = w;
|
||||
true
|
||||
}
|
||||
None => {
|
||||
self.work = 0;
|
||||
self.exhausted = true;
|
||||
false
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Charges the values of an array (one unit per value; one for anything else).
|
||||
pub(crate) fn spend_on(&mut self, v: &Value) -> bool {
|
||||
match v {
|
||||
Value::Array(a) => self.spend(a.data.len() as u64),
|
||||
_ => self.spend(1),
|
||||
}
|
||||
Evaluator { host, sheet, at, env: Vec::new(), depth: 0, calls: 0, stack_top: 0, fuel: LAMBDA_FUEL }
|
||||
}
|
||||
|
||||
/// Evaluates to a value (references are dereferenced; multi-cell → array).
|
||||
@@ -335,27 +266,18 @@ impl<'h> Evaluator<'h> {
|
||||
}
|
||||
// A reference keeps its size, so the array has its shape and its blanks count
|
||||
// (`=A1:A10`, `SUMPRODUCT(--(A1:A10=""))`).
|
||||
let range = self.read_range(a);
|
||||
let (h, w) = (range.height() as usize, range.width() as usize);
|
||||
if (h as u64) * (w as u64) > MAX_CELLS {
|
||||
return Value::Error(CellError::Num);
|
||||
}
|
||||
if !self.spend((h as u64) * (w as u64)) {
|
||||
return Value::Error(CellError::Calc);
|
||||
}
|
||||
let data = self.host.range_values(a.sheet, range);
|
||||
Array::new(h, w, data).map(Value::from).unwrap_or(Value::Error(CellError::Value))
|
||||
}
|
||||
|
||||
/// The block that reading `a` gives: `a` itself, or when it is large (or runs to the sheet's
|
||||
/// edge) the part inside the sheet's used range.
|
||||
pub(crate) fn read_range(&mut self, a: &Area) -> RangeRef {
|
||||
let cells = u64::from(a.range.height()) * u64::from(a.range.width());
|
||||
// `A:A`, `1:1` and `A2:A1048576` run to the edge of the sheet: they stand for "the rest of
|
||||
// the column", not for a block of a chosen size.
|
||||
let to_edge = a.range.end.row == MAX_ROWS - 1 || a.range.end.col == MAX_COLS - 1;
|
||||
let full_size = cells <= FULL_SIZE_CELLS || (!to_edge && cells <= FULL_SIZE_BOUNDED_CELLS);
|
||||
if full_size { a.range } else { self.trim(a) }
|
||||
let range = if full_size { a.range } else { self.trim(a) };
|
||||
let (h, w) = (range.height() as usize, range.width() as usize);
|
||||
if (h as u64) * (w as u64) > MAX_CELLS {
|
||||
return Value::Error(CellError::Num);
|
||||
}
|
||||
let data = self.host.range_values(a.sheet, range);
|
||||
Array::new(h, w, data).map(Value::from).unwrap_or(Value::Error(CellError::Value))
|
||||
}
|
||||
|
||||
/// Trims the bottom/right of an area to the sheet's used range (keeps the top-left so
|
||||
@@ -388,9 +310,6 @@ impl<'h> Evaluator<'h> {
|
||||
self.host.stack_exhausted();
|
||||
return err(CellError::Num);
|
||||
}
|
||||
if !self.spend(1) {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
self.depth += 1;
|
||||
if self.depth > MAX_DEPTH {
|
||||
self.depth -= 1;
|
||||
@@ -463,11 +382,6 @@ impl<'h> Evaluator<'h> {
|
||||
if let Some(def) = wb.name(bare, scope_sheet) {
|
||||
let text = def.formula.clone();
|
||||
let scope = def.scope.unwrap_or(self.sheet);
|
||||
// Each use parses and evaluates the definition again: names that each use the one
|
||||
// before twice double the work per name, so it is charged.
|
||||
if !self.spend(text.len() as u64) {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
return match gridcraft_formula::parse_name(&text) {
|
||||
Ok(expr) => {
|
||||
// Relative references in a name are relative to A1: used here, they move by
|
||||
@@ -629,11 +543,6 @@ impl<'h> Evaluator<'h> {
|
||||
if r.height() == 1 && r.start.col <= self.at.col && self.at.col <= r.end.col {
|
||||
return Some(CellRef::new(r.start.row, self.at.col));
|
||||
}
|
||||
// A 2-D range intersects on both the formula's row and column (observed in Excel, on
|
||||
// another sheet too); one that includes the formula's own cell is a circular reference.
|
||||
if r.start.row <= self.at.row && self.at.row <= r.end.row && r.start.col <= self.at.col && self.at.col <= r.end.col {
|
||||
return Some(self.at);
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
@@ -665,10 +574,6 @@ impl<'h> Evaluator<'h> {
|
||||
let (x, y) = (self.eval(a), self.eval(b));
|
||||
match (x, y) {
|
||||
(Ev::R(mut xa), Ev::R(ya)) => {
|
||||
// A union reused through LET doubles at each step: its areas count.
|
||||
if !self.spend((xa.len() as u64).saturating_add(ya.len() as u64)) {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
xa.extend(ya);
|
||||
Ev::R(xa)
|
||||
}
|
||||
@@ -680,11 +585,6 @@ impl<'h> Evaluator<'h> {
|
||||
let (x, y) = (self.eval(a), self.eval(b));
|
||||
match (x, y) {
|
||||
(Ev::R(xa), Ev::R(ya)) => {
|
||||
// Every pair is looked at, and each may be kept: charged before the loop,
|
||||
// so unions squared through LET can't build an enormous list.
|
||||
if !self.spend((xa.len() as u64).saturating_mul(ya.len() as u64)) {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
let mut out = Vec::new();
|
||||
for p in &xa {
|
||||
for q in &ya {
|
||||
@@ -778,113 +678,6 @@ mod fuel_tests {
|
||||
ev.value(&gridcraft_formula::parser::parse(src).unwrap_or_else(|e| panic!("{e}")))
|
||||
}
|
||||
|
||||
/// Evaluates `src` in `wb` with `work` units of work, and the work left.
|
||||
fn eval_work(wb: Workbook, work: u64, src: &str) -> (Value, u64) {
|
||||
let mut host = Empty { wb };
|
||||
let mut ev = Evaluator::new(&mut host, 0, CellRef::new(0, 0));
|
||||
ev.work = work;
|
||||
let v = ev.value(&gridcraft_formula::parser::parse(src).unwrap_or_else(|e| panic!("{e}")));
|
||||
(v, ev.work)
|
||||
}
|
||||
|
||||
fn is_calc(v: &Value) -> bool {
|
||||
matches!(v, Value::Error(CellError::Calc))
|
||||
}
|
||||
|
||||
/// Serves the workbook's stored values, for database functions to read.
|
||||
struct Stored {
|
||||
wb: Workbook,
|
||||
}
|
||||
impl Host for Stored {
|
||||
fn workbook(&self) -> &Workbook {
|
||||
&self.wb
|
||||
}
|
||||
fn cell_value(&mut self, sheet: usize, c: CellRef) -> Value {
|
||||
self.wb.sheet(sheet).and_then(|s| s.cell(c)).map_or(Value::Empty, |cell| cell.value.clone())
|
||||
}
|
||||
fn now_serial(&self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
fn random(&mut self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
}
|
||||
|
||||
/// A computed criterion that runs out of work counts as not met for its record, which the
|
||||
/// database function can't see: it must answer #CALC!, not a count that leaves records out.
|
||||
#[test]
|
||||
fn a_computed_criterion_out_of_work_makes_the_database_function_calc() {
|
||||
use gridcraft_model::{Cell, Formula};
|
||||
let mut wb = Workbook::new();
|
||||
let name = |name: String, formula: String| gridcraft_model::DefinedName { name, scope: None, formula, comment: String::new(), hidden: false };
|
||||
wb.names.push(name("Step0".into(), "1".into()));
|
||||
for i in 1..=60 {
|
||||
wb.names.push(name(format!("Step{i}"), format!("Step{}+Step{}", i - 1, i - 1)));
|
||||
}
|
||||
let at = |s: &str| CellRef::parse(s).unwrap();
|
||||
let sheet = wb.sheet_mut(0).unwrap();
|
||||
for (r, (n, age)) in [("name", Value::text("age")), ("a", Value::number(1.0)), ("b", Value::number(2.0))].into_iter().enumerate() {
|
||||
sheet.cells.set(at(&format!("A{}", r + 1)), Cell::value(Value::text(n)));
|
||||
sheet.cells.set(at(&format!("B{}", r + 1)), Cell::value(age));
|
||||
}
|
||||
sheet.cells.set(at("D1"), Cell::value(Value::text("Formula")));
|
||||
let host_wb = |criterion: &str| {
|
||||
let mut wb = wb.clone();
|
||||
// A criteria cell is read by its value first (a blank one matches everything): give it
|
||||
// the value a recalculation would have stored.
|
||||
let mut cell = Cell::formula(Formula::new(criterion));
|
||||
cell.value = Value::Bool(true);
|
||||
wb.sheet_mut(0).unwrap().cells.set(at("D2"), cell);
|
||||
wb
|
||||
};
|
||||
let run = |criterion: &str| {
|
||||
let mut host = Stored { wb: host_wb(criterion) };
|
||||
let mut ev = Evaluator::new(&mut host, 0, CellRef::new(9, 9));
|
||||
ev.work = 100_000;
|
||||
ev.value(&gridcraft_formula::parser::parse("=DCOUNT(A1:B3,\"age\",D1:D2)").unwrap())
|
||||
};
|
||||
assert_eq!(run("=B2>0"), Value::number(2.0), "a cheap criterion counts every record");
|
||||
assert!(is_calc(&run("=Step60>0")), "{:?}", run("=Step60>0"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn names_that_double_at_each_step_run_out_of_work() {
|
||||
// Step1 = Step0+Step0, Step2 = Step1+Step1, ...: Step60 would evaluate Step0 2^60 times.
|
||||
let mut wb = Workbook::default();
|
||||
let name = |name: String, formula: String| gridcraft_model::DefinedName { name, scope: None, formula, comment: String::new(), hidden: false };
|
||||
wb.names.push(name("Step0".into(), "1".into()));
|
||||
for i in 1..=60 {
|
||||
wb.names.push(name(format!("Step{i}"), format!("Step{}+Step{}", i - 1, i - 1)));
|
||||
}
|
||||
let (v, _) = eval_work(wb.clone(), EVAL_WORK, "=Step5");
|
||||
assert_eq!(v, Value::number(32.0));
|
||||
let (v, left) = eval_work(wb, 100_000, "=Step60");
|
||||
assert!(is_calc(&v), "{v:?}");
|
||||
assert_eq!(left, 0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn intersections_of_unions_are_charged_per_pair() {
|
||||
let wb = Workbook::default();
|
||||
// 4 x 4 pairs, 4 of which meet.
|
||||
let src = "=AREAS((A1,B1,C1,D1) (A1,B1,C1,D1))";
|
||||
let (v, _) = eval_work(wb.clone(), EVAL_WORK, src);
|
||||
assert_eq!(v, Value::number(4.0));
|
||||
let (v, _) = eval_work(wb, 20, src);
|
||||
assert!(is_calc(&v), "{v:?}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn an_aggregate_inside_every_lambda_call_spends_work() {
|
||||
let wb = Workbook::default();
|
||||
// 100 calls, each summing the 10,000 values of a captured array: about a million units.
|
||||
let src = "=LET(big,SEQUENCE(10000),SUM(MAP(SEQUENCE(100),LAMBDA(x,SUM(big)))))";
|
||||
let (v, _) = eval_work(wb.clone(), 2_000_000, src);
|
||||
assert_eq!(v, Value::number(100.0 * 50_005_000.0));
|
||||
let (v, _) = eval_work(wb, 200_000, src);
|
||||
assert!(is_calc(&v), "{v:?}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn lambda_calls_spend_fuel_and_run_out_quietly() {
|
||||
// Three applications, one charge each: 3 inside the budget. (A nested MAP wouldn't do
|
||||
|
||||
@@ -52,16 +52,13 @@ impl<'h> Evaluator<'h> {
|
||||
"ISFORMULA" | "FORMULATEXT" => {
|
||||
let Some(Ev::R(a)) = args.first().map(|a| self.eval(a)) else { return err(CellError::NA) };
|
||||
let Some(a) = a.first() else { return err(CellError::NA) };
|
||||
let wb = self.host.workbook();
|
||||
let f = wb.sheet(a.sheet).and_then(|s| s.cell(a.range.start)).and_then(|c| c.formula.clone());
|
||||
// A formula formatted Hidden on a protected sheet stays hidden from FORMULATEXT too.
|
||||
let hidden = wb.sheet(a.sheet).is_some_and(|s| s.is_protected() && wb.styles.get(s.style_id(a.range.start)).protection.hidden);
|
||||
let f = self.host.workbook().sheet(a.sheet).and_then(|s| s.cell(a.range.start)).and_then(|c| c.formula.clone());
|
||||
if name == "ISFORMULA" {
|
||||
Ev::V(Value::Bool(f.is_some()))
|
||||
} else {
|
||||
match f {
|
||||
Some(f) if !hidden => Ev::V(Value::text(format!("={}", f.text))),
|
||||
_ => err(CellError::NA),
|
||||
Some(f) => Ev::V(Value::text(format!("={}", f.text))),
|
||||
None => err(CellError::NA),
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -206,9 +203,9 @@ impl<'h> Evaluator<'h> {
|
||||
let index = self.value(&args[0]);
|
||||
if let Value::Array(_) = index {
|
||||
let Some(spec) = gridcraft_functions::lookup("CHOOSE") else { return err(CellError::Name) };
|
||||
let mut evaluated = vec![Arg::val(index)];
|
||||
let mut evaluated = vec![Arg { value: index, from_ref: false }];
|
||||
for argument in &args[1..] {
|
||||
evaluated.push(Arg::val(self.value(argument)));
|
||||
evaluated.push(Arg { value: self.value(argument), from_ref: false });
|
||||
}
|
||||
let mut context = FnCtx { ev: self, criteria: None, lambda: None };
|
||||
return gridcraft_functions::call(spec, &evaluated, &mut context).pipe(Ev::V);
|
||||
@@ -354,25 +351,10 @@ impl<'h> Evaluator<'h> {
|
||||
}
|
||||
ev => self.deref(ev),
|
||||
};
|
||||
// Only an argument left out (`f(a,,b)`) is omitted; one that computes a blank is given.
|
||||
let blank_value = !from_ref && value == Value::Empty && lambda_at != Some(i) && !matches!(a, Expr::Missing);
|
||||
evaluated.push(Arg { value, from_ref, blank_value });
|
||||
evaluated.push(Arg { value, from_ref });
|
||||
}
|
||||
// The function may go through every value it is given (a SUM inside each call of a
|
||||
// LAMBDA over a large array): the values count against the cell's work.
|
||||
let values = evaluated.iter().fold(0u64, |n, a| n.saturating_add(if let Value::Array(x) = &a.value { x.data.len() as u64 } else { 1 }));
|
||||
if !self.spend(values) {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
let database = criteria.is_some();
|
||||
let mut ctx = FnCtx { ev: self, criteria, lambda };
|
||||
let v = gridcraft_functions::call(spec, &evaluated, &mut ctx);
|
||||
// A computed criterion that ran out of work counted as not met: the result leaves
|
||||
// records out.
|
||||
if database && self.exhausted {
|
||||
return err(CellError::Calc);
|
||||
}
|
||||
Ev::V(v)
|
||||
gridcraft_functions::call(spec, &evaluated, &mut ctx).pipe(Ev::V)
|
||||
}
|
||||
|
||||
/// Performs aggregation operations without copying its ranges: blocks are folded over the stored
|
||||
@@ -410,12 +392,6 @@ impl<'h> Evaluator<'h> {
|
||||
acc.count = acc.count.saturating_add(count);
|
||||
return Ok(());
|
||||
}
|
||||
// Folded without building an array, but still walked: the part inside the used
|
||||
// range counts against the cell's work.
|
||||
let r = self.read_range(a);
|
||||
if !self.spend(u64::from(r.height()) * u64::from(r.width())) {
|
||||
return Err(err(CellError::Calc));
|
||||
}
|
||||
let mut error = None;
|
||||
let walked = self.host.visit_range(a.sheet, a.range, &mut |v| match acc.add_cell(kind, v) {
|
||||
Ok(()) => true,
|
||||
@@ -435,13 +411,10 @@ impl<'h> Evaluator<'h> {
|
||||
/// Adds one argument's value by calling `SUM` and `COUNT` on it, so coercion and error rules
|
||||
/// stay in one place (`AVERAGE` of an argument is its `SUM` over its `COUNT`).
|
||||
fn aggregate_value(&mut self, kind: Agg, value: Value, from_ref: bool, acc: &mut Acc) -> Result<(), Ev> {
|
||||
if !from_ref && !self.spend_on(&value) {
|
||||
return Err(err(CellError::Calc));
|
||||
}
|
||||
let (Some(sum), Some(count)) = (gridcraft_functions::lookup("SUM"), gridcraft_functions::lookup("COUNT")) else {
|
||||
return Err(err(CellError::Name));
|
||||
};
|
||||
let args = [Arg { value, from_ref, blank_value: false }];
|
||||
let args = [Arg { value, from_ref }];
|
||||
if kind != Agg::Count {
|
||||
match gridcraft_functions::call(sum, &args, &mut FnCtx { ev: self, criteria: None, lambda: None }) {
|
||||
Value::Number(n) => acc.sum += n,
|
||||
|
||||
@@ -118,29 +118,11 @@ impl<'h> Evaluator<'h> {
|
||||
for c in 0..w {
|
||||
let pos = (num(&Some(ra.get_broadcast(r, c))), num(&Some(ca.get_broadcast(r, c))));
|
||||
let v = match pos {
|
||||
// A position that picks several cells is #CALC!. That is decided from the
|
||||
// shape: reading the block first, for every position, let a short formula
|
||||
// read one large range over and over only to discard it.
|
||||
(Ok(row), Ok(col)) if index_picks_several_values(&src, row, col, two) => Value::Error(CellError::Calc),
|
||||
(Ok(row), Ok(col)) => {
|
||||
let one = self.index_one(src.clone(), row, col, area_n, two);
|
||||
let cells = match &one {
|
||||
Ev::R(areas) => match areas.as_slice() {
|
||||
[a] => self.area_cells(a),
|
||||
_ => 0,
|
||||
},
|
||||
_ => 0,
|
||||
};
|
||||
if cells > MAX_CELLS {
|
||||
// As reading it would answer.
|
||||
Value::Error(CellError::Num)
|
||||
} else if cells > 1 {
|
||||
Value::Error(CellError::Calc)
|
||||
} else {
|
||||
match self.deref(one) {
|
||||
Value::Array(a) if a.rows * a.cols > 1 => Value::Error(CellError::Calc),
|
||||
v => v.scalar(),
|
||||
}
|
||||
match self.deref(one) {
|
||||
Value::Array(a) if a.rows * a.cols > 1 => Value::Error(CellError::Calc),
|
||||
v => v.scalar(),
|
||||
}
|
||||
}
|
||||
(Err(e), _) | (_, Err(e)) => Value::Error(e),
|
||||
@@ -157,16 +139,6 @@ impl<'h> Evaluator<'h> {
|
||||
self.index_one(src, row, col, area_n, two)
|
||||
}
|
||||
|
||||
/// Cells that reading `a` gives: its own size, or for a large block the part inside the
|
||||
/// sheet's used range (what [`Evaluator::area_value`] reads).
|
||||
fn area_cells(&mut self, a: &Area) -> u64 {
|
||||
if a.range.is_single() {
|
||||
return 1;
|
||||
}
|
||||
let r = self.read_range(a);
|
||||
u64::from(r.height()) * u64::from(r.width())
|
||||
}
|
||||
|
||||
/// INDEX for one row and column position (0 = the whole column or row).
|
||||
fn index_one(&mut self, src: Ev, row: i64, col: i64, area_n: i64, two_args: bool) -> Ev {
|
||||
if row < 0 || col < 0 {
|
||||
@@ -231,7 +203,7 @@ impl<'h> Evaluator<'h> {
|
||||
let horizontal = l.range.height() == 1 && l.range.width() > 1;
|
||||
let same_len = if vertical { r.range.height() == l.range.height() } else { r.range.width() == l.range.width() };
|
||||
if !matches!(key, Value::Array(_)) && (vertical || horizontal) && same_len {
|
||||
let mut xargs = vec![Arg::val(key), Arg::reference(self.area_value(&l))];
|
||||
let mut xargs = vec![Arg { value: key, from_ref: false }, Arg { value: self.area_value(&l), from_ref: true }];
|
||||
// An empty match_mode before a given search_mode is the default (0), not the end.
|
||||
let given = |i: usize| args.get(i).is_some_and(|e| !matches!(e, Expr::Missing));
|
||||
let last = if given(5) {
|
||||
@@ -243,11 +215,10 @@ impl<'h> Evaluator<'h> {
|
||||
};
|
||||
for i in 4..=last {
|
||||
match args.get(i) {
|
||||
Some(Expr::Missing) | None => xargs.push(Arg::val(Value::Number(0.0))),
|
||||
Some(Expr::Missing) | None => xargs.push(Arg { value: Value::Number(0.0), from_ref: false }),
|
||||
Some(e) => {
|
||||
let value = self.value(e);
|
||||
let blank_value = value == Value::Empty;
|
||||
xargs.push(Arg { value, from_ref: false, blank_value });
|
||||
xargs.push(Arg { value, from_ref: false });
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -436,71 +407,3 @@ impl<'h> Evaluator<'h> {
|
||||
Ev::V(gridcraft_functions::aggregate_values(func, &values))
|
||||
}
|
||||
}
|
||||
|
||||
/// Whether INDEX on an array picks a whole row or column of several values at position
|
||||
/// (`row`, `col`), which the array-of-positions form answers with #CALC!.
|
||||
fn index_picks_several_values(src: &Ev, row: i64, col: i64, two_args: bool) -> bool {
|
||||
let Ev::V(Value::Array(arr)) = src else { return false };
|
||||
let (row, col) = if two_args && arr.rows == 1 { (1, row) } else { (row, col) };
|
||||
if row < 0 || col < 0 || row as u64 > arr.rows as u64 || col as u64 > arr.cols as u64 {
|
||||
return false;
|
||||
}
|
||||
(row == 0 && arr.rows > 1) || (col == 0 && arr.cols > 1)
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::super::Host;
|
||||
use super::*;
|
||||
use gridcraft_model::Workbook;
|
||||
|
||||
/// A sheet of blanks that counts the blocks it is asked for.
|
||||
struct Counting {
|
||||
wb: Workbook,
|
||||
cells_read: usize,
|
||||
}
|
||||
impl Host for Counting {
|
||||
fn workbook(&self) -> &Workbook {
|
||||
&self.wb
|
||||
}
|
||||
fn cell_value(&mut self, _sheet: usize, _c: CellRef) -> Value {
|
||||
Value::Empty
|
||||
}
|
||||
fn now_serial(&self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
fn random(&mut self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
fn range_values(&mut self, _sheet: usize, range: RangeRef) -> Vec<Value> {
|
||||
self.cells_read += range.count() as usize;
|
||||
vec![Value::Empty; range.count() as usize]
|
||||
}
|
||||
}
|
||||
|
||||
fn eval(src: &str) -> (Value, usize) {
|
||||
let mut host = Counting { wb: Workbook::default(), cells_read: 0 };
|
||||
let mut ev = Evaluator::new(&mut host, 0, CellRef::new(0, 0));
|
||||
let v = ev.value(&gridcraft_formula::parser::parse(src).unwrap_or_else(|e| panic!("{e}")));
|
||||
(v, host.cells_read)
|
||||
}
|
||||
|
||||
fn all_calc(v: &Value, n: usize) -> bool {
|
||||
matches!(v, Value::Array(a) if a.data.len() == n && a.data.iter().all(|x| *x == Value::Error(CellError::Calc)))
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn index_positions_that_pick_several_cells_are_calc_without_reading_them() {
|
||||
// Each zero row position picks the whole of A1:A50000: #CALC!, decided from the shape. It
|
||||
// used to read the 50,000 cells once per position before discarding them.
|
||||
let (v, read) = eval("=INDEX(A1:A50000,{0,0,0,0},1)");
|
||||
assert!(all_calc(&v, 4), "{v:?}");
|
||||
assert_eq!(read, 0);
|
||||
// The same for an array source: a whole column of a 3x3 array.
|
||||
let (v, _) = eval("=INDEX({1,2,3;4,5,6;7,8,9},{0,0},2)");
|
||||
assert!(all_calc(&v, 2), "{v:?}");
|
||||
// Single cells are still read.
|
||||
let (v, _) = eval("=INDEX({1,2,3;4,5,6;7,8,9},{1,3},2)");
|
||||
assert!(matches!(&v, Value::Array(a) if a.data == vec![Value::number(2.0), Value::number(8.0)]), "{v:?}");
|
||||
}
|
||||
}
|
||||
|
||||
+2
-112
@@ -1826,54 +1826,6 @@ fn trim_references() {
|
||||
assert_eq!(t.get("K20"), Value::Error(CellError::Calc));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn lookup_of_an_empty_cell_stays_blank() {
|
||||
// #374: B1 (the value for key "k") is empty.
|
||||
let mut t = T::new();
|
||||
t.set("A1", "k");
|
||||
t.set("A2", "j");
|
||||
t.set("B2", "v");
|
||||
for (at, f, want) in [
|
||||
("D2", "=\"x\"&VLOOKUP(\"k\",A1:B2,2,0)&\"y\"", Value::from("xy")),
|
||||
("D3", "=CONCATENATE(\"x\",VLOOKUP(\"k\",A1:B2,2,0),\"y\")", Value::from("xy")),
|
||||
("D4", "=ISBLANK(XLOOKUP(\"k\",A1:A2,B1:B2))", Value::Bool(true)),
|
||||
("D5", "=\"x\"&XLOOKUP(\"k\",A1:A2,B1:B2)&\"y\"", Value::from("xy")),
|
||||
// A bare lookup still shows 0, like a reference to an empty cell.
|
||||
("D8", "=VLOOKUP(\"k\",A1:B2,2,0)", Value::Number(0.0)),
|
||||
// Arithmetic still reads the blank as 0.
|
||||
("D9", "=VLOOKUP(\"k\",A1:B2,2,0)+1", Value::Number(1.0)),
|
||||
// A blank lookup given as an argument is still given, not omitted.
|
||||
("D10", "=FILTER(1,FALSE,VLOOKUP(\"k\",A1:B2,2,FALSE))", Value::Number(0.0)),
|
||||
("D11", "=XLOOKUP(\"z\",A1:A2,B1:B2,VLOOKUP(\"k\",A1:B2,2,0))", Value::Number(0.0)),
|
||||
// The same holds per element when the call is lifted over an array of blank lookups.
|
||||
("D12", "=CEILING.MATH(6.7,VLOOKUP(\"k\",A1:B2,2,0))", Value::Number(0.0)),
|
||||
("D13", "=SUM(CEILING.MATH(6.7,VLOOKUP({\"k\",\"k\"},A1:B2,2,0)))", Value::Number(0.0)),
|
||||
("D14", "=FLOOR.MATH(6.7,VLOOKUP(\"k\",A1:B2,2,0))", Value::Number(0.0)),
|
||||
("D15", "=SUM(FLOOR.MATH(6.7,VLOOKUP({\"k\",\"k\"},A1:B2,2,0)))", Value::Number(0.0)),
|
||||
] {
|
||||
t.set(at, f);
|
||||
assert_eq!(t.get(at), want, "{f}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn blank_lookup_values_match_excel_windows_observation() {
|
||||
// Excel 365 Windows 2609, build 16.0.20430.20032, black-box check 2026-10-13:
|
||||
// A1="k", B1 and A2 genuinely empty.
|
||||
// All three value lookups preserve blankness inside ISBLANK/concatenation.
|
||||
let mut t = T::new();
|
||||
t.set("A1", "k");
|
||||
for (at, f, want) in [
|
||||
("D1", "=ISBLANK(VLOOKUP(\"k\",A1:B1,2,FALSE))", Value::Bool(true)),
|
||||
("D2", "=ISBLANK(HLOOKUP(\"k\",A1:A2,2,FALSE))", Value::Bool(true)),
|
||||
("D3", "=ISBLANK(LOOKUP(\"k\",A1:A1,B1:B1))", Value::Bool(true)),
|
||||
("D4", "=\"x\"&HLOOKUP(\"k\",A1:A2,2,FALSE)&\"y\"", Value::from("xy")),
|
||||
] {
|
||||
t.set(at, f);
|
||||
assert_eq!(t.get(at), want, "{f}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn cell_width_is_the_column_width_dialog_value_rounded() {
|
||||
// #735: Excel 365 values. Widths in px as `home.columnWidth {"chars"}` sets them
|
||||
@@ -1970,54 +1922,6 @@ fn offset_extreme_arguments_return_ref() {
|
||||
}
|
||||
}
|
||||
|
||||
/// `@` intersects a reference with the formula's row (a column range) or column (a row range);
|
||||
/// a 2D range both (circular when it includes the formula's cell); a range the formula does not
|
||||
/// line up with gives #VALUE!. An array gives its top-left value. Legacy formulas read from files use it where Excel intersects implicitly.
|
||||
#[test]
|
||||
fn implicit_intersection_operator() {
|
||||
let mut t = T::new();
|
||||
for (at, v) in [("A1", "1"), ("A2", "2"), ("A3", "3"), ("A5", "4"), ("B5", "5"), ("C5", "6")] {
|
||||
t.set(at, v);
|
||||
}
|
||||
t.set("B2", "=TANH(@A1:A3)");
|
||||
t.set("B7", "=TANH(@A5:C5)");
|
||||
t.set("B3", "=@A1:C3");
|
||||
t.set("B9", "=@A1:A3");
|
||||
t.set("E9", "=@A5:C5");
|
||||
t.set("D1", "=@{7,8;9,10}");
|
||||
t.set("D2", "=@A2");
|
||||
t.set("D3", "=SUM(@A1:A3)");
|
||||
assert_eq!(t.num("B2"), 2f64.tanh());
|
||||
assert_eq!(t.num("B7"), 5f64.tanh());
|
||||
for at in ["B9", "E9"] {
|
||||
assert_eq!(t.get(at), Value::Error(CellError::Value), "{at}");
|
||||
}
|
||||
// A 2D range that includes the formula's own cell is a circular reference (Excel shows 0).
|
||||
assert_eq!(t.get("B3"), Value::Error(CellError::Circ));
|
||||
// From another sheet a 2D range intersects on both row and column; outside it is #VALUE!.
|
||||
t.wb.sheets.push(std::sync::Arc::new(gridcraft_model::Sheet::new("Other")));
|
||||
for (at, input) in [("B2", "=@Sheet1!A1:C3"), ("C1", "=@Sheet1!A1:C3"), ("B9", "=@Sheet1!A1:C3"), ("E2", "=@Sheet1!A1:C3")] {
|
||||
t.wb.sheet_mut(1).unwrap().cells.set(c(at), Cell::formula(Formula::new(input)));
|
||||
}
|
||||
t.calc.recalc_all(&mut t.wb);
|
||||
let other = |at: &str| t.wb.sheet(1).unwrap().value(c(at));
|
||||
assert_eq!(other("B2"), Value::Number(2f64.tanh())); // Sheet1!B2
|
||||
assert_eq!(other("C1"), Value::Number(0.0));
|
||||
assert_eq!(other("B9"), Value::Error(CellError::Value));
|
||||
assert_eq!(other("E2"), Value::Error(CellError::Value));
|
||||
assert_eq!(t.num("D1"), 7.0);
|
||||
assert_eq!(t.num("D2"), 2.0);
|
||||
assert_eq!(t.num("D3"), 3.0);
|
||||
// An array result shows its top-left value (T of a reference reads its top-left cell).
|
||||
t.set("A11", "x");
|
||||
t.set("C11", "y");
|
||||
t.set("E1", "=@T(A11:C11)");
|
||||
assert_eq!(t.get("E1"), Value::text("x"));
|
||||
// Without `@` (a dynamic-array formula) the same formula spills.
|
||||
t.set("F1", "=TANH(A1:A3)");
|
||||
assert_eq!((t.num("F1"), t.num("F2"), t.num("F3")), (1f64.tanh(), 2f64.tanh(), 3f64.tanh()));
|
||||
}
|
||||
|
||||
/// "Set precision as displayed": results are rounded to their cell's format before they are
|
||||
/// stored and before other formulas read them. Values observed in Excel (COM,
|
||||
/// `Workbook.PrecisionAsDisplayed = True`) with the same synthetic formulas.
|
||||
@@ -2083,6 +1987,7 @@ fn precision_as_displayed_rounds_results_by_cell_format() {
|
||||
t.calc.recalc_all(&mut t.wb);
|
||||
assert_eq!(t.num("A2"), (-130.0 + 657.41) * 10.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn analysis_toolpak_value_params_reject_multi_cell_references() {
|
||||
// Excel 365 (observed): EOMONTH, WORKDAY, COUPDAYS, BIN2DEC... give #VALUE! for a reference to more
|
||||
@@ -2167,6 +2072,7 @@ fn analysis_toolpak_value_params_reject_multi_cell_references() {
|
||||
t.set("H5", "");
|
||||
}
|
||||
}
|
||||
|
||||
/// A formula reading one further down (or on a later sheet) is calculated on demand, inside the
|
||||
/// formula that reads it. Chains of thousands of such formulas, common in financial models, used
|
||||
/// to overflow the stack (or give `#CIRC!` past 2,000 levels); they now calculate in full.
|
||||
@@ -2880,19 +2786,3 @@ fn formulas_read_a_dirty_python_cell_after_it_is_laid_out() {
|
||||
assert_eq!((t.num("B1"), t.num("B2")), (5.0, 6.0));
|
||||
assert_eq!((t.num("A1"), t.num("A2"), t.num("A3")), (5.0, 100.0, 111.0));
|
||||
}
|
||||
|
||||
/// A computed criterion whose formula runs a database function over the same criteria counts as
|
||||
/// not met inside itself: before, each record's check ran the function again for every record, at
|
||||
/// every level, so the work grew exponentially with the nesting allowed.
|
||||
#[test]
|
||||
fn computed_criterion_that_reenters_itself_is_not_met() {
|
||||
let mut t = T::new();
|
||||
for (r, (name, age)) in [("name", "age"), ("a", "1"), ("b", "2"), ("c", "3"), ("d", "4"), ("e", "5")].iter().enumerate() {
|
||||
t.set(&format!("F{}", r + 2), name);
|
||||
t.set(&format!("G{}", r + 2), age);
|
||||
}
|
||||
t.set("N46", "Formula");
|
||||
t.set("N47", "=DCOUNT($F$2:$G$7,\"age\",$N$46:$N$47)>0");
|
||||
t.set("B60", "=DCOUNT($F$2:$G$7,\"age\",N46:N47)");
|
||||
assert_eq!(t.num("B60"), 0.0);
|
||||
}
|
||||
|
||||
@@ -7,20 +7,6 @@ use crate::{Prim, Rgba};
|
||||
/// Largest image we allocate (pixels).
|
||||
const MAX_PIXELS: u64 = 64 * 1024 * 1024;
|
||||
const SUB: usize = 4;
|
||||
/// Scan-conversion work one [`rasterize`] call may do: edge tests (an active edge at one
|
||||
/// sub-sample) plus covered pixels blended. Edges are only tested on the rows they cross, but a
|
||||
/// caller can still pass enough geometry to keep the rasteriser busy for minutes on a small
|
||||
/// image; once the allowance is spent the remaining fills are skipped. 2^28 is about a second,
|
||||
/// far beyond any chart.
|
||||
const MAX_SCAN_WORK: u64 = 1 << 28;
|
||||
/// Most points one filled path may have: more is skipped before any edge is built. Chart
|
||||
/// renderers cap and decimate their paths far below this; it only bounds direct callers.
|
||||
const MAX_FILL_POINTS: usize = 1 << 22;
|
||||
|
||||
/// Most points one stroked polyline may have: more is skipped before any segment path is built
|
||||
/// (each segment becomes its own quad). Charts hold at most 100,000 points per series and draw far
|
||||
/// fewer; it only bounds direct callers.
|
||||
const MAX_STROKE_POINTS: usize = 1 << 17;
|
||||
|
||||
type Path = Vec<Vec<[f32; 2]>>;
|
||||
|
||||
@@ -29,8 +15,6 @@ struct Canvas {
|
||||
h: usize,
|
||||
buf: Vec<u8>,
|
||||
acc: Vec<f32>,
|
||||
/// Scan-conversion work left (see [`MAX_SCAN_WORK`]).
|
||||
work: u64,
|
||||
}
|
||||
|
||||
impl Canvas {
|
||||
@@ -57,12 +41,7 @@ impl Canvas {
|
||||
|
||||
/// Fills a multi-contour path with the even-odd rule.
|
||||
fn fill(&mut self, path: &Path, color: Rgba) {
|
||||
// Out of work, or more geometry than the allowance could ever scan: no edge list is built.
|
||||
if self.work == 0 || path.iter().map(Vec::len).fold(0usize, usize::saturating_add) > MAX_FILL_POINTS {
|
||||
return;
|
||||
}
|
||||
// Each edge as [top y, bottom y, x0, y0, x1, y1].
|
||||
let mut edges: Vec<[f32; 6]> = Vec::new();
|
||||
let mut edges: Vec<[f32; 4]> = Vec::new();
|
||||
let (mut minx, mut miny, mut maxx, mut maxy) = (f32::INFINITY, f32::INFINITY, f32::NEG_INFINITY, f32::NEG_INFINITY);
|
||||
for c in path {
|
||||
if c.len() < 3 || !c.iter().all(|p| p[0].is_finite() && p[1].is_finite()) {
|
||||
@@ -75,7 +54,7 @@ impl Canvas {
|
||||
miny = miny.min(p[1]);
|
||||
maxy = maxy.max(p[1]);
|
||||
if p[1] != q[1] {
|
||||
edges.push([p[1].min(q[1]), p[1].max(q[1]), p[0], p[1], q[0], q[1]]);
|
||||
edges.push([p[0], p[1], q[0], q[1]]);
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -93,28 +72,8 @@ impl Canvas {
|
||||
if self.acc.len() < bw + 1 {
|
||||
self.acc.resize(bw + 1, 0.0);
|
||||
}
|
||||
// Active-edge list: edges join when the scan reaches their top and leave past their
|
||||
// bottom, so each row tests only the edges crossing it instead of every edge.
|
||||
edges.sort_by(|a, b| a[0].total_cmp(&b[0]));
|
||||
let mut pending = edges.iter().peekable();
|
||||
let mut active: Vec<[f32; 6]> = Vec::new();
|
||||
let mut xs: Vec<f32> = Vec::new();
|
||||
for py in y0..y1 {
|
||||
while let Some(e) = pending.next_if(|e| e[0] < py as f32 + 1.0) {
|
||||
active.push(*e);
|
||||
}
|
||||
active.retain(|e| e[1] > py as f32);
|
||||
if active.is_empty() {
|
||||
continue;
|
||||
}
|
||||
// The row's edge tests and its `bw` accumulator clears and blends, charged before any of
|
||||
// them, so rows with no coverage still count.
|
||||
let cost = (active.len() * SUB) as u64 + bw as u64;
|
||||
if cost > self.work {
|
||||
self.work = 0;
|
||||
return;
|
||||
}
|
||||
self.work -= cost;
|
||||
for v in self.acc.iter_mut().take(bw) {
|
||||
*v = 0.0;
|
||||
}
|
||||
@@ -122,9 +81,10 @@ impl Canvas {
|
||||
for s in 0..SUB {
|
||||
let sy = py as f32 + (s as f32 + 0.5) / SUB as f32;
|
||||
xs.clear();
|
||||
for e in &active {
|
||||
if sy >= e[0] && sy < e[1] {
|
||||
xs.push(e[2] + (sy - e[3]) * (e[4] - e[2]) / (e[5] - e[3]));
|
||||
for e in &edges {
|
||||
let (ya, yb) = if e[1] < e[3] { (e[1], e[3]) } else { (e[3], e[1]) };
|
||||
if sy >= ya && sy < yb {
|
||||
xs.push(e[0] + (sy - e[1]) * (e[2] - e[0]) / (e[3] - e[1]));
|
||||
}
|
||||
}
|
||||
xs.sort_by(f32::total_cmp);
|
||||
@@ -194,9 +154,6 @@ fn rounded_rect(x: f32, y: f32, w: f32, h: f32, r: f32) -> Vec<[f32; 2]> {
|
||||
|
||||
/// Thick polyline → one quad per segment (each a separate path so overlaps don't cancel).
|
||||
fn stroke_paths(pts: &[[f32; 2]], width: f32, dash: bool, closed: bool) -> Vec<Path> {
|
||||
if pts.len() > MAX_STROKE_POINTS {
|
||||
return Vec::new();
|
||||
}
|
||||
let width = width.clamp(1.0, 50.0);
|
||||
let mut segs: Vec<([f32; 2], [f32; 2])> = pts.windows(2).map(|w| (w[0], w[1])).collect();
|
||||
if closed && let (Some(a), Some(b)) = (pts.last(), pts.first()) {
|
||||
@@ -255,7 +212,7 @@ pub fn rasterize(prims: &[Prim], w: u32, h: u32, background: Rgba) -> Vec<u8> {
|
||||
for _ in 0..w * h {
|
||||
buf.extend_from_slice(&background);
|
||||
}
|
||||
let mut cv = Canvas { w, h, buf, acc: Vec::new(), work: MAX_SCAN_WORK };
|
||||
let mut cv = Canvas { w, h, buf, acc: Vec::new() };
|
||||
for p in prims {
|
||||
match p {
|
||||
Prim::Rect { x, y, w, h, fill, stroke, radius } => {
|
||||
@@ -317,40 +274,6 @@ pub fn rasterize(prims: &[Prim], w: u32, h: u32, background: Rgba) -> Vec<u8> {
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
/// Security: each row tests only the edges crossing it. A polygon of 100,000 tiny edges over
|
||||
/// 4,000 rows tested every edge at every sub-sample of every row (1.6 billion tests).
|
||||
#[test]
|
||||
fn many_short_edges_scan_only_their_rows() {
|
||||
let mut pts: Vec<[f32; 2]> = (0..100_000).map(|i| [if i % 2 == 0 { 0.0 } else { 0.5 }, i as f32 * 0.04]).collect();
|
||||
pts.push([2.0, 4_000.0]);
|
||||
pts.push([2.0, 0.0]);
|
||||
let img = rasterize(&[Prim::Polygon { pts, fill: [255, 0, 0, 255], stroke: None }], 2, 4_000, [0, 0, 0, 0]);
|
||||
// The right-hand pixel of every row is inside the polygon.
|
||||
assert_eq!(img.get(4..8), Some(&[255, 0, 0, 255][..]));
|
||||
assert_eq!(img.get(4 * (2 * 3_999 + 1)..4 * (2 * 3_999 + 2)), Some(&[255, 0, 0, 255][..]));
|
||||
}
|
||||
|
||||
/// Security: scan conversion stops once the call's work allowance is spent.
|
||||
#[test]
|
||||
fn fills_stop_when_the_work_allowance_is_spent() {
|
||||
let square = vec![vec![[0.0, 0.0], [4.0, 0.0], [4.0, 4.0], [0.0, 4.0]]];
|
||||
let mut cv = Canvas { w: 4, h: 4, buf: vec![0; 64], acc: Vec::new(), work: 2 * SUB as u64 * 4 + 16 };
|
||||
cv.fill(&square, [255; 4]);
|
||||
assert!(cv.buf.iter().all(|&b| b == 255));
|
||||
cv.buf = vec![0; 64];
|
||||
cv.fill(&square, [255; 4]);
|
||||
assert!(cv.buf.iter().all(|&b| b == 0));
|
||||
}
|
||||
|
||||
/// Security: a polyline over the point cap builds no segment paths at all, solid or dashed.
|
||||
#[test]
|
||||
fn oversized_polylines_build_no_segment_paths() {
|
||||
let line = |n: usize| -> Vec<[f32; 2]> { (0..n).map(|i| [i as f32, (i % 2) as f32]).collect() };
|
||||
assert!(stroke_paths(&line(MAX_STROKE_POINTS + 1), 1.0, false, false).is_empty());
|
||||
assert!(stroke_paths(&line(MAX_STROKE_POINTS + 1), 1.0, true, false).is_empty());
|
||||
assert_eq!(stroke_paths(&line(MAX_STROKE_POINTS), 1.0, false, false).len(), MAX_STROKE_POINTS - 1);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn dashed_geometry_stays_within_the_dash_budget() {
|
||||
// Forty very long dashed lines, each allowed 10_000 dashes per segment: before the
|
||||
|
||||
@@ -136,13 +136,7 @@ pub(super) fn decimate_area(top: Vec<[f32; 2]>, base: Vec<[f32; 2]>, max: usize)
|
||||
|
||||
/// Draws a polyline split at gaps (`None`), decimated when long. With `smooth` (the plot box),
|
||||
/// each run is interpolated with a Catmull-Rom spline so the line curves through the data points.
|
||||
///
|
||||
/// The points of the smoothed runs (data points included) are budgeted across all the runs of the line, not per run:
|
||||
/// short runs between gaps escape decimation, and 100 sub-segments each would turn a
|
||||
/// 100,000-point series of three-point runs into millions of points. Past the budget the
|
||||
/// remaining runs are drawn through their data points only.
|
||||
pub(super) fn draw_runs(ctx: &mut Ctx, pts: &[Option<[f32; 2]>], color: Rgba, width: f32, smooth: Option<Bx>) {
|
||||
let mut budget = MAX_LINE_PTS * 2;
|
||||
let mut run: Vec<[f32; 2]> = Vec::new();
|
||||
for p in pts.iter().chain(std::iter::once(&None)) {
|
||||
match p {
|
||||
@@ -151,11 +145,7 @@ pub(super) fn draw_runs(ctx: &mut Ctx, pts: &[Option<[f32; 2]>], color: Rgba, wi
|
||||
if run.len() >= 2 {
|
||||
let r = decimate(std::mem::take(&mut run), MAX_LINE_PTS);
|
||||
let r = match smooth {
|
||||
Some(plot) => {
|
||||
let r = smooth_run(r, plot, budget);
|
||||
budget = budget.saturating_sub(r.len());
|
||||
r
|
||||
}
|
||||
Some(plot) => smooth_run(r, plot),
|
||||
None => r,
|
||||
};
|
||||
ctx.line(r, color, width);
|
||||
@@ -168,15 +158,15 @@ pub(super) fn draw_runs(ctx: &mut Ctx, pts: &[Option<[f32; 2]>], color: Rgba, wi
|
||||
}
|
||||
|
||||
/// Catmull-Rom spline through `pts`: every segment gets the same number of sub-segments (up to
|
||||
/// 100), chosen so the whole curve stays within `cap` points (one per data point when `cap` is
|
||||
/// smaller than that). Endpoints are
|
||||
/// 100), chosen so the whole curve stays within `MAX_LINE_PTS * 2` points. Endpoints are
|
||||
/// duplicated so the curve passes through the first and last points, and the curve is clamped
|
||||
/// to the `plot` box (a spline overshoots its data near sharp turns).
|
||||
fn smooth_run(pts: Vec<[f32; 2]>, plot: Bx, cap: usize) -> Vec<[f32; 2]> {
|
||||
fn smooth_run(pts: Vec<[f32; 2]>, plot: Bx) -> Vec<[f32; 2]> {
|
||||
let n = pts.len();
|
||||
if n < 3 {
|
||||
return pts;
|
||||
}
|
||||
let cap = MAX_LINE_PTS * 2;
|
||||
let steps = (cap / n).clamp(1, 100);
|
||||
let mut out: Vec<[f32; 2]> = Vec::with_capacity(n.saturating_mul(steps).min(cap) + 1);
|
||||
let at = |i: isize| -> [f32; 2] {
|
||||
@@ -229,7 +219,7 @@ mod smooth_tests {
|
||||
fn long_runs_are_smoothed_uniformly_within_the_cap() {
|
||||
let plot = Bx { x: 0.0, y: 0.0, w: 3000.0, h: 100.0 };
|
||||
let pts: Vec<[f32; 2]> = (0..1500).map(|i| [i as f32 * 2.0, if i % 2 == 0 { 20.0 } else { 80.0 }]).collect();
|
||||
let out = smooth_run(pts, plot, MAX_LINE_PTS * 2);
|
||||
let out = smooth_run(pts, plot);
|
||||
assert!(out.len() <= MAX_LINE_PTS * 2 + 1, "{}", out.len());
|
||||
// Two sub-segments per segment all the way to the end, not 100 each and then a straight tail.
|
||||
assert_eq!(out.len(), 1499 * 2 + 1);
|
||||
@@ -241,29 +231,11 @@ mod smooth_tests {
|
||||
let plot = Bx { x: 0.0, y: 10.0, w: 300.0, h: 100.0 };
|
||||
// A step right at the plot edges: Catmull-Rom overshoots past both.
|
||||
let pts = vec![[0.0, 110.0], [100.0, 110.0], [110.0, 10.0], [200.0, 10.0], [300.0, 10.0]];
|
||||
let out = smooth_run(pts, plot, MAX_LINE_PTS * 2);
|
||||
let out = smooth_run(pts, plot);
|
||||
assert!(out.len() > 5);
|
||||
assert!(out.iter().all(|p| (0.0..=300.0).contains(&p[0]) && (10.0..=110.0).contains(&p[1])));
|
||||
}
|
||||
|
||||
/// Security: smoothing budgets its added points across every run of a line. A series of
|
||||
/// three-point runs between gaps used to get 100 sub-segments per run, 67 times its data.
|
||||
#[test]
|
||||
fn gap_separated_runs_share_one_smoothing_budget() {
|
||||
let plot = Bx { x: 0.0, y: 0.0, w: 1e6, h: 100.0 };
|
||||
let pts: Vec<Option<[f32; 2]>> =
|
||||
(0..100_000).map(|i| if i % 4 == 3 { None } else { Some([i as f32, if i % 2 == 0 { 20.0 } else { 80.0 }]) }).collect();
|
||||
let m = crate::ApproxMeasure;
|
||||
let mut ctx = Ctx { out: Vec::new(), m: &m };
|
||||
draw_runs(&mut ctx, &pts, WHITE, 1.0, Some(plot));
|
||||
let total: usize = ctx.out.iter().map(|p| if let Prim::Line { pts, .. } = p { pts.len() } else { 0 }).sum();
|
||||
// 75,000 data points, plus at most the budget added by interpolation.
|
||||
assert!(total <= 75_000 + MAX_LINE_PTS * 2, "{total}");
|
||||
assert_eq!(ctx.out.len(), 25_000);
|
||||
// The first runs are still smoothed.
|
||||
assert!(matches!(ctx.out.first(), Some(Prim::Line { pts, .. }) if pts.len() > 3));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn decimate_area_preserves_paired_vertical_strips() {
|
||||
let top: Vec<[f32; 2]> = (0..1000).map(|i| [i as f32 * 0.25, ((i * 37) % 100) as f32]).collect();
|
||||
|
||||
+16
-103
@@ -184,53 +184,6 @@ fn value_text(v: &Value) -> String {
|
||||
}
|
||||
}
|
||||
|
||||
/// Characters kept of one category label: an axis label never shows more (labels are fitted to
|
||||
/// the axis), and copying more only costs memory.
|
||||
const MAX_LABEL_CHARS: usize = 255;
|
||||
/// Bytes of category label text kept per series. One long text broadcast by a short formula
|
||||
/// (`IF(SEQUENCE(100000), long_text)`) is a single shared value, but turning every element into
|
||||
/// its own label copied it 100,000 times (gigabytes). Labels past the budget are empty.
|
||||
const MAX_CATEGORY_BYTES: usize = 8 << 20;
|
||||
|
||||
/// At most [`MAX_LABEL_CHARS`] characters and `room` bytes of `s`, cut at a character boundary.
|
||||
fn clip_label(s: &str, room: usize) -> String {
|
||||
let mut end = 0;
|
||||
for (n, (i, c)) in s.char_indices().enumerate() {
|
||||
if n >= MAX_LABEL_CHARS || i + c.len_utf8() > room {
|
||||
break;
|
||||
}
|
||||
end = i + c.len_utf8();
|
||||
}
|
||||
s.get(..end).unwrap_or_default().to_string()
|
||||
}
|
||||
|
||||
/// The label of category `i` (value `v`) in at most `room` bytes: formatted with its source
|
||||
/// cell's number format when it is a number.
|
||||
fn category_label(wb: &Workbook, src: Option<(&gridcraft_model::Sheet, RangeRef)>, i: usize, v: &Value, room: usize) -> String {
|
||||
if room == 0 {
|
||||
return String::new(); // nothing fits: skip the number formatting
|
||||
}
|
||||
if let (Some((s, r)), Value::Number(_)) = (src, v) {
|
||||
let w = r.width().max(1) as usize;
|
||||
let at = r.start.offset((i / w) as i64, (i % w) as i64);
|
||||
if let Some(c) = at {
|
||||
let code = &wb.styles.get(s.style_id(c)).num_fmt.0;
|
||||
if !code.eq_ignore_ascii_case("general") {
|
||||
return clip_label(&format_value(v, &NumberFormat::parse_cached(code), wb.date_system).text, room);
|
||||
}
|
||||
}
|
||||
}
|
||||
// Text is clipped before it is copied: the source may be one long shared string.
|
||||
let mut v = v;
|
||||
while let Value::Array(a) = v {
|
||||
v = a.get(0, 0).unwrap_or(&Value::Empty);
|
||||
}
|
||||
match v {
|
||||
Value::Text(t) => clip_label(t, room),
|
||||
_ => clip_label(&value_text(v), room),
|
||||
}
|
||||
}
|
||||
|
||||
/// Category labels, formatted with each source cell's number format (dates, currency…), and
|
||||
/// the outer levels of a multi-level axis: a category range wider than a column of values (or
|
||||
/// taller than a row of values) groups its innermost labels, year beside month. Each outer level
|
||||
@@ -268,29 +221,29 @@ fn category_table(wb: &Workbook, sheet: usize, formula: &str, values: &str, limi
|
||||
None => eval_list_in(wb, sheet, formula, limit, 0),
|
||||
};
|
||||
let src = cells.and_then(|(sh, r)| wb.sheet(sh).map(|s| (s, r)));
|
||||
// The retained levels are known before any label is charged: a cell on a discarded level
|
||||
// gets no label and spends none of the budget. `None`: a flat list (every cell is a label).
|
||||
let layout = src.and_then(|(_, r)| {
|
||||
let levels = levels?;
|
||||
let (w, h) = (r.width().max(1) as usize, r.height().max(1) as usize);
|
||||
(vals.len() == w.saturating_mul(h)).then_some((levels, w, h)) // else truncated or not a plain block
|
||||
});
|
||||
let kept =
|
||||
|i: usize| layout.is_none_or(|(l, w, h)| if l.by_columns { i % w + 1 + l.outer as usize >= w } else { i / w + 1 + l.outer as usize >= h });
|
||||
let mut room = MAX_CATEGORY_BYTES;
|
||||
let texts: Vec<String> = vals
|
||||
.iter()
|
||||
.enumerate()
|
||||
.map(|(i, v)| {
|
||||
if !kept(i) {
|
||||
return String::new();
|
||||
if let (Some((s, r)), Value::Number(_)) = (src, v) {
|
||||
let w = r.width().max(1) as usize;
|
||||
let at = r.start.offset((i / w) as i64, (i % w) as i64);
|
||||
if let Some(c) = at {
|
||||
let code = &wb.styles.get(s.style_id(c)).num_fmt.0;
|
||||
if !code.eq_ignore_ascii_case("general") {
|
||||
return format_value(v, &NumberFormat::parse_cached(code), wb.date_system).text;
|
||||
}
|
||||
}
|
||||
}
|
||||
let label = category_label(wb, src, i, v, room);
|
||||
room = room.saturating_sub(label.len());
|
||||
label
|
||||
value_text(v)
|
||||
})
|
||||
.collect();
|
||||
let Some((levels, w, h)) = layout else { return (texts, vec![]) };
|
||||
let Some((_, r)) = src else { return (texts, vec![]) };
|
||||
let Some(levels) = levels else { return (texts, vec![]) };
|
||||
let (w, h) = (r.width().max(1) as usize, r.height().max(1) as usize);
|
||||
if texts.len() != w.saturating_mul(h) {
|
||||
return (texts, vec![]); // truncated or not a plain block: a flat list
|
||||
}
|
||||
let at = |row: usize, col: usize| texts.get(row * w + col).cloned().unwrap_or_default();
|
||||
let column = |c: usize| (0..h).map(|row| at(row, c)).collect::<Vec<String>>();
|
||||
let row = |rr: usize| (0..w).map(|c| at(rr, c)).collect::<Vec<String>>();
|
||||
@@ -416,46 +369,6 @@ pub fn resolve(wb: &Workbook, sheet: usize, chart: &Chart) -> ChartData {
|
||||
ChartData { categories, category_levels, series, palette }
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod label_budget_tests {
|
||||
use gridcraft_model::Workbook;
|
||||
|
||||
/// Security: category labels are clipped before they are copied, with one byte budget per
|
||||
/// series. A formula broadcasting one long text to 100,000 categories used to copy it
|
||||
/// 100,000 times (3.3 GB).
|
||||
#[test]
|
||||
fn category_labels_share_one_byte_budget() {
|
||||
let wb = Workbook::new();
|
||||
let (texts, _) = super::category_table(&wb, 0, r#"=IF(SEQUENCE(100000),REPT("x",32767))"#, "", super::MAX_POINTS);
|
||||
assert_eq!(texts.len(), 100_000);
|
||||
assert_eq!(texts.first().map(String::len), Some(super::MAX_LABEL_CHARS));
|
||||
assert!(texts.iter().map(String::len).sum::<usize>() <= super::MAX_CATEGORY_BYTES);
|
||||
// Labels are cut at character boundaries.
|
||||
assert_eq!(super::clip_label("aé", 2), "a");
|
||||
assert_eq!(super::clip_label("aé", 3), "aé");
|
||||
}
|
||||
|
||||
/// Cells on discarded outer levels (only the innermost labels and three outer levels are
|
||||
/// kept) spend none of the byte budget.
|
||||
#[test]
|
||||
fn discarded_levels_do_not_spend_the_label_budget() {
|
||||
let mut wb = Workbook::new();
|
||||
let rows = 8000u32;
|
||||
if let Some(sh) = wb.sheet_mut(0) {
|
||||
for r in 0..rows {
|
||||
for c in 0..5 {
|
||||
sh.set_value(gridcraft_core::CellRef::new(r, c), gridcraft_core::Value::text("x".repeat(255)));
|
||||
}
|
||||
}
|
||||
}
|
||||
// Five columns: column A is a fourth outer level and is discarded.
|
||||
let (inner, outer) = super::category_table(&wb, 0, "=A1:E8000", "", super::MAX_POINTS);
|
||||
assert_eq!(outer.len(), 3);
|
||||
assert_eq!(inner.last().map(String::len), Some(super::MAX_LABEL_CHARS));
|
||||
assert_eq!(outer.first().and_then(|l| l.last()).map(String::len), Some(super::MAX_LABEL_CHARS));
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use gridcraft_model::Workbook;
|
||||
|
||||
@@ -147,10 +147,10 @@ impl RangeRef {
|
||||
RangeRef::new(CellRef::new(0, 0), CellRef::new(MAX_ROWS - 1, MAX_COLS - 1))
|
||||
}
|
||||
pub fn height(&self) -> u32 {
|
||||
self.end.row.saturating_sub(self.start.row).saturating_add(1)
|
||||
self.end.row - self.start.row + 1
|
||||
}
|
||||
pub fn width(&self) -> u32 {
|
||||
self.end.col.saturating_sub(self.start.col).saturating_add(1)
|
||||
self.end.col - self.start.col + 1
|
||||
}
|
||||
pub fn count(&self) -> u64 {
|
||||
self.height() as u64 * self.width() as u64
|
||||
@@ -321,13 +321,6 @@ pub fn split_areas(reference: &str) -> Vec<&str> {
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn hostile_range_dimensions_saturate_instead_of_wrapping() {
|
||||
let r = RangeRef { start: CellRef::new(0, 0), end: CellRef::new(u32::MAX, u32::MAX) };
|
||||
assert_eq!((r.height(), r.width()), (u32::MAX, u32::MAX));
|
||||
assert_eq!(r.count(), u64::from(u32::MAX) * u64::from(u32::MAX));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn areas_split_only_at_commas_outside_quoted_sheet_names() {
|
||||
assert_eq!(split_areas("B2:B4,D1"), ["B2:B4", "D1"]);
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
//! GridCraft core types: cell addresses, values and Excel's coercion rules, error values,
|
||||
//! date serials, regional formats and typed-input parsing. No dependencies on the rest of the workspace.
|
||||
//! date serials and typed-input parsing. No dependencies on the rest of the workspace.
|
||||
#![deny(clippy::unwrap_used, clippy::expect_used, clippy::panic, clippy::unimplemented, clippy::todo, clippy::unreachable)]
|
||||
#![forbid(unsafe_code)]
|
||||
|
||||
@@ -7,7 +7,6 @@ pub mod addr;
|
||||
pub mod color;
|
||||
pub mod date;
|
||||
pub mod parse;
|
||||
pub mod region;
|
||||
pub mod value;
|
||||
|
||||
pub use addr::{
|
||||
@@ -15,5 +14,4 @@ pub use addr::{
|
||||
split_areas, split_sheet, unquote_sheet,
|
||||
};
|
||||
pub use date::DateSystem;
|
||||
pub use region::Region;
|
||||
pub use value::{Array, CellError, Value, compare, compare_numbers, compare_text, number_to_text, sort_compare};
|
||||
|
||||
@@ -1,883 +0,0 @@
|
||||
//! Regional formats: how a region writes numbers, dates, times and money.
|
||||
//!
|
||||
//! Excel takes these from the operating system's regional settings; they are independent of the
|
||||
//! interface language and of the formula language. A [`Region`] changes what people *type* into
|
||||
//! a cell (`1.234,56`, `25/09/2026`) and what cells *show* (separators, month and day names,
|
||||
//! AM/PM, the system short and long dates). It never changes what is stored: values are numbers,
|
||||
//! and the number format codes saved in a file stay in their canonical en-US spelling.
|
||||
//!
|
||||
//! [`EN_US`] is the default and reproduces GridCraft's original behaviour exactly.
|
||||
//!
|
||||
//! # Where the values come from
|
||||
//! The separators, month and weekday names, AM/PM words, currency symbols and the order of day,
|
||||
//! month and year are the Unicode CLDR conventions (`unicode-org/cldr-json`:
|
||||
//! `cldr-numbers-full/main/<locale>/numbers.json` and `cldr-dates-full/main/<locale>/ca-gregorian.json`),
|
||||
//! which is also what macOS, Android and ICU ship. They were read from those files for en-GB,
|
||||
//! en-IN, cs and ru, read from ICU's CLDR 48 data for the Swiss regions and nl-NL, and written
|
||||
//! from the same conventions for the other regions. CLDR data is under the Unicode License v3,
|
||||
//! reproduced in `NOTICE` (row in `ATTRIBUTION.md`).
|
||||
//! Deliberate choices where operating systems differ from CLDR:
|
||||
//! - Short dates use a four-digit year (`dd/mm/yyyy`, where CLDR has `dd/MM/y` or `dd/MM/yy`):
|
||||
//! Excel's short date is the system's, Windows writes four digits, and a two-digit year would
|
||||
//! not read back to the same date.
|
||||
//! - The long date is CLDR's *full* date (with the weekday), like Windows en-US, except where
|
||||
//! Windows' long date differs: Russian, Japanese and Chinese have no weekday (`23 августа
|
||||
//! 2005 г.`, `2005年8月23日`), as in Microsoft's National Language Support tables.
|
||||
//! - Korean writes the short date `yyyy-mm-dd` (CLDR: `yy. M. d.`), also as in Microsoft's
|
||||
//! National Language Support tables.
|
||||
//! - Month names for `mmm` are CLDR's abbreviated names (so en-GB has `Sept`), except
|
||||
//! Portuguese, which drops CLDR's period (`jan`, not `jan.`) as Excel in Brazil shows it.
|
||||
//! - French and Portuguese (Portugal) group digits with U+00A0 (Windows), not CLDR's U+202F.
|
||||
//! Parsing accepts every space-like grouping character in those regions.
|
||||
//! - The Swiss regions (en-CH, de-CH, fr-CH, it-CH; Liechtenstein uses de-CH) group digits with
|
||||
//! CLDR's ASCII apostrophe (`1'234.50`); parsing also accepts the typographic `’` that glibc
|
||||
//! uses. French (Switzerland) writes a decimal point like the other Swiss regions and glibc's
|
||||
//! `fr_CH`, where CLDR has a comma.
|
||||
//!
|
||||
//! Regions this module doesn't list (`de-AT`, `es-MX`, `en-CA`…) resolve to the main region of
|
||||
//! their language through [`lookup`]; their own conventions (a Mexican decimal point) are not
|
||||
//! modelled.
|
||||
|
||||
/// How digits are grouped.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
|
||||
pub enum Grouping {
|
||||
/// Groups of three (`1,234,567`).
|
||||
Thousands,
|
||||
/// The Indian system: three digits, then groups of two (`12,34,567`).
|
||||
Indian,
|
||||
}
|
||||
|
||||
/// The order of the parts of a short date.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
|
||||
pub enum DateOrder {
|
||||
Mdy,
|
||||
Dmy,
|
||||
Ymd,
|
||||
}
|
||||
|
||||
/// One region's conventions. See the module documentation.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
|
||||
pub struct Region {
|
||||
/// BCP 47 tag, `pt-BR`.
|
||||
pub tag: &'static str,
|
||||
/// The region's name in its own language.
|
||||
pub name: &'static str,
|
||||
/// Windows locale id, used by `[$-409]`-style format tags.
|
||||
pub lcid: u16,
|
||||
pub decimal: char,
|
||||
pub group: char,
|
||||
pub grouping: Grouping,
|
||||
pub date_order: DateOrder,
|
||||
pub date_sep: char,
|
||||
pub time_sep: char,
|
||||
/// The system short date as a canonical format code (`dd/mm/yyyy`): built-in format 14.
|
||||
pub short_date: &'static str,
|
||||
/// The system long date (`[$-F800]`).
|
||||
pub long_date: &'static str,
|
||||
/// The system short time. It contains `AM/PM` in regions that use a 12-hour clock.
|
||||
pub short_time: &'static str,
|
||||
/// The system long time (`[$-F400]`).
|
||||
pub long_time: &'static str,
|
||||
/// Built-in format 22: the short date and a 24-hour time.
|
||||
pub datetime: &'static str,
|
||||
pub months: [&'static str; 12],
|
||||
pub months_abbr: [&'static str; 12],
|
||||
/// The month names a date with a day uses (Russian `1 января`, Czech `1. ledna`), where they
|
||||
/// differ from `months`.
|
||||
pub months_genitive: Option<[&'static str; 12]>,
|
||||
/// Sunday first.
|
||||
pub weekdays: [&'static str; 7],
|
||||
pub weekdays_abbr: [&'static str; 7],
|
||||
pub am_pm: [&'static str; 2],
|
||||
pub currency: &'static str,
|
||||
/// The currency format with the currency's usual decimals (two, or none for the yen and the
|
||||
/// won), as a canonical code. Where it starts with a quote the symbol comes before the number,
|
||||
/// otherwise after.
|
||||
pub currency_format: &'static str,
|
||||
/// The same with no decimals (for a typed whole amount).
|
||||
pub currency_format_int: &'static str,
|
||||
}
|
||||
|
||||
const EN_MONTHS: [&str; 12] =
|
||||
["January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December"];
|
||||
const EN_MONTHS_ABBR: [&str; 12] = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"];
|
||||
const EN_DAYS: [&str; 7] = ["Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday"];
|
||||
const EN_DAYS_ABBR: [&str; 7] = ["Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"];
|
||||
const AM_PM: [&str; 2] = ["AM", "PM"];
|
||||
|
||||
/// English (United States): the default. Identical to the formats before regions existed.
|
||||
pub static EN_US: Region = Region {
|
||||
tag: "en-US",
|
||||
name: "English (United States)",
|
||||
lcid: 0x0409,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Mdy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "m/d/yyyy",
|
||||
long_date: "dddd, mmmm dd, yyyy",
|
||||
short_time: "h:mm AM/PM",
|
||||
long_time: "h:mm:ss AM/PM",
|
||||
datetime: "m/d/yyyy h:mm",
|
||||
months: EN_MONTHS,
|
||||
months_abbr: EN_MONTHS_ABBR,
|
||||
months_genitive: None,
|
||||
weekdays: EN_DAYS,
|
||||
weekdays_abbr: EN_DAYS_ABBR,
|
||||
am_pm: AM_PM,
|
||||
currency: "$",
|
||||
currency_format: "\"$\"#,##0.00",
|
||||
currency_format_int: "\"$\"#,##0",
|
||||
};
|
||||
|
||||
static EN_GB: Region = Region {
|
||||
tag: "en-GB",
|
||||
name: "English (United Kingdom)",
|
||||
lcid: 0x0809,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd, d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd/mm/yyyy hh:mm",
|
||||
months: EN_MONTHS,
|
||||
months_abbr: ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sept", "Oct", "Nov", "Dec"],
|
||||
months_genitive: None,
|
||||
weekdays: EN_DAYS,
|
||||
weekdays_abbr: EN_DAYS_ABBR,
|
||||
am_pm: AM_PM,
|
||||
currency: "£",
|
||||
currency_format: "\"£\"#,##0.00",
|
||||
currency_format_int: "\"£\"#,##0",
|
||||
};
|
||||
|
||||
static EN_IN: Region = Region {
|
||||
tag: "en-IN",
|
||||
name: "English (India)",
|
||||
lcid: 0x4009,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Indian,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd, d mmmm yyyy",
|
||||
short_time: "h:mm AM/PM",
|
||||
long_time: "h:mm:ss AM/PM",
|
||||
datetime: "dd/mm/yyyy h:mm",
|
||||
months: EN_MONTHS,
|
||||
months_abbr: ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sept", "Oct", "Nov", "Dec"],
|
||||
months_genitive: None,
|
||||
weekdays: EN_DAYS,
|
||||
weekdays_abbr: EN_DAYS_ABBR,
|
||||
am_pm: AM_PM,
|
||||
currency: "₹",
|
||||
currency_format: "\"₹\"#,##0.00",
|
||||
currency_format_int: "\"₹\"#,##0",
|
||||
};
|
||||
|
||||
static EN_CH: Region = Region {
|
||||
tag: "en-CH",
|
||||
name: "English (Switzerland)",
|
||||
// English (Switzerland) has no LCID of its own: Windows reports LOCALE_CUSTOM_UNSPECIFIED.
|
||||
lcid: 0x1000,
|
||||
decimal: '.',
|
||||
group: '\'',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd, d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: EN_MONTHS,
|
||||
months_abbr: ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sept", "Oct", "Nov", "Dec"],
|
||||
months_genitive: None,
|
||||
weekdays: EN_DAYS,
|
||||
weekdays_abbr: EN_DAYS_ABBR,
|
||||
am_pm: AM_PM,
|
||||
currency: "CHF",
|
||||
currency_format: "\"CHF\" #,##0.00",
|
||||
currency_format_int: "\"CHF\" #,##0",
|
||||
};
|
||||
|
||||
static ES_ES: Region = Region {
|
||||
tag: "es-ES",
|
||||
name: "Español (España)",
|
||||
lcid: 0x0C0A,
|
||||
decimal: ',',
|
||||
group: '.',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd, d \"de\" mmmm \"de\" yyyy",
|
||||
short_time: "h:mm",
|
||||
long_time: "h:mm:ss",
|
||||
datetime: "dd/mm/yyyy h:mm",
|
||||
months: ["enero", "febrero", "marzo", "abril", "mayo", "junio", "julio", "agosto", "septiembre", "octubre", "noviembre", "diciembre"],
|
||||
months_abbr: ["ene", "feb", "mar", "abr", "may", "jun", "jul", "ago", "sep", "oct", "nov", "dic"],
|
||||
months_genitive: None,
|
||||
weekdays: ["domingo", "lunes", "martes", "miércoles", "jueves", "viernes", "sábado"],
|
||||
weekdays_abbr: ["dom", "lun", "mar", "mié", "jue", "vie", "sáb"],
|
||||
am_pm: ["a. m.", "p. m."],
|
||||
currency: "€",
|
||||
currency_format: "#,##0.00 \"€\"",
|
||||
currency_format_int: "#,##0 \"€\"",
|
||||
};
|
||||
|
||||
static DE_DE: Region = Region {
|
||||
tag: "de-DE",
|
||||
name: "Deutsch (Deutschland)",
|
||||
lcid: 0x0407,
|
||||
decimal: ',',
|
||||
group: '.',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd, d. mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: ["Januar", "Februar", "März", "April", "Mai", "Juni", "Juli", "August", "September", "Oktober", "November", "Dezember"],
|
||||
months_abbr: ["Jan", "Feb", "Mär", "Apr", "Mai", "Jun", "Jul", "Aug", "Sep", "Okt", "Nov", "Dez"],
|
||||
months_genitive: None,
|
||||
weekdays: ["Sonntag", "Montag", "Dienstag", "Mittwoch", "Donnerstag", "Freitag", "Samstag"],
|
||||
weekdays_abbr: ["So", "Mo", "Di", "Mi", "Do", "Fr", "Sa"],
|
||||
am_pm: AM_PM,
|
||||
currency: "€",
|
||||
currency_format: "#,##0.00 \"€\"",
|
||||
currency_format_int: "#,##0 \"€\"",
|
||||
};
|
||||
|
||||
static DE_CH: Region = Region {
|
||||
tag: "de-CH",
|
||||
name: "Deutsch (Schweiz)",
|
||||
lcid: 0x0807,
|
||||
decimal: '.',
|
||||
group: '\'',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd, d. mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: ["Januar", "Februar", "März", "April", "Mai", "Juni", "Juli", "August", "September", "Oktober", "November", "Dezember"],
|
||||
months_abbr: ["Jan", "Feb", "Mär", "Apr", "Mai", "Jun", "Jul", "Aug", "Sep", "Okt", "Nov", "Dez"],
|
||||
months_genitive: None,
|
||||
weekdays: ["Sonntag", "Montag", "Dienstag", "Mittwoch", "Donnerstag", "Freitag", "Samstag"],
|
||||
weekdays_abbr: ["So", "Mo", "Di", "Mi", "Do", "Fr", "Sa"],
|
||||
am_pm: AM_PM,
|
||||
currency: "CHF",
|
||||
currency_format: "\"CHF\" #,##0.00",
|
||||
currency_format_int: "\"CHF\" #,##0",
|
||||
};
|
||||
|
||||
static FR_FR: Region = Region {
|
||||
tag: "fr-FR",
|
||||
name: "Français (France)",
|
||||
lcid: 0x040C,
|
||||
decimal: ',',
|
||||
group: '\u{a0}',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd/mm/yyyy hh:mm",
|
||||
months: ["janvier", "février", "mars", "avril", "mai", "juin", "juillet", "août", "septembre", "octobre", "novembre", "décembre"],
|
||||
months_abbr: ["janv.", "févr.", "mars", "avr.", "mai", "juin", "juil.", "août", "sept.", "oct.", "nov.", "déc."],
|
||||
months_genitive: None,
|
||||
weekdays: ["dimanche", "lundi", "mardi", "mercredi", "jeudi", "vendredi", "samedi"],
|
||||
weekdays_abbr: ["dim.", "lun.", "mar.", "mer.", "jeu.", "ven.", "sam."],
|
||||
am_pm: AM_PM,
|
||||
currency: "€",
|
||||
currency_format: "#,##0.00 \"€\"",
|
||||
currency_format_int: "#,##0 \"€\"",
|
||||
};
|
||||
|
||||
static FR_CH: Region = Region {
|
||||
tag: "fr-CH",
|
||||
name: "Français (Suisse)",
|
||||
lcid: 0x100C,
|
||||
decimal: '.',
|
||||
group: '\'',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd, d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: ["janvier", "février", "mars", "avril", "mai", "juin", "juillet", "août", "septembre", "octobre", "novembre", "décembre"],
|
||||
months_abbr: ["janv.", "févr.", "mars", "avr.", "mai", "juin", "juil.", "août", "sept.", "oct.", "nov.", "déc."],
|
||||
months_genitive: None,
|
||||
weekdays: ["dimanche", "lundi", "mardi", "mercredi", "jeudi", "vendredi", "samedi"],
|
||||
weekdays_abbr: ["dim.", "lun.", "mar.", "mer.", "jeu.", "ven.", "sam."],
|
||||
am_pm: AM_PM,
|
||||
currency: "CHF",
|
||||
currency_format: "#,##0.00 \"CHF\"",
|
||||
currency_format_int: "#,##0 \"CHF\"",
|
||||
};
|
||||
|
||||
static IT_IT: Region = Region {
|
||||
tag: "it-IT",
|
||||
name: "Italiano (Italia)",
|
||||
lcid: 0x0410,
|
||||
decimal: ',',
|
||||
group: '.',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd/mm/yyyy hh:mm",
|
||||
months: ["gennaio", "febbraio", "marzo", "aprile", "maggio", "giugno", "luglio", "agosto", "settembre", "ottobre", "novembre", "dicembre"],
|
||||
months_abbr: ["gen", "feb", "mar", "apr", "mag", "giu", "lug", "ago", "set", "ott", "nov", "dic"],
|
||||
months_genitive: None,
|
||||
weekdays: ["domenica", "lunedì", "martedì", "mercoledì", "giovedì", "venerdì", "sabato"],
|
||||
weekdays_abbr: ["dom", "lun", "mar", "mer", "gio", "ven", "sab"],
|
||||
am_pm: AM_PM,
|
||||
currency: "€",
|
||||
currency_format: "#,##0.00 \"€\"",
|
||||
currency_format_int: "#,##0 \"€\"",
|
||||
};
|
||||
|
||||
static IT_CH: Region = Region {
|
||||
tag: "it-CH",
|
||||
name: "Italiano (Svizzera)",
|
||||
lcid: 0x0810,
|
||||
decimal: '.',
|
||||
group: '\'',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd, d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: ["gennaio", "febbraio", "marzo", "aprile", "maggio", "giugno", "luglio", "agosto", "settembre", "ottobre", "novembre", "dicembre"],
|
||||
months_abbr: ["gen", "feb", "mar", "apr", "mag", "giu", "lug", "ago", "set", "ott", "nov", "dic"],
|
||||
months_genitive: None,
|
||||
weekdays: ["domenica", "lunedì", "martedì", "mercoledì", "giovedì", "venerdì", "sabato"],
|
||||
weekdays_abbr: ["dom", "lun", "mar", "mer", "gio", "ven", "sab"],
|
||||
am_pm: AM_PM,
|
||||
currency: "CHF",
|
||||
currency_format: "\"CHF\" #,##0.00",
|
||||
currency_format_int: "\"CHF\" #,##0",
|
||||
};
|
||||
|
||||
static NL_NL: Region = Region {
|
||||
tag: "nl-NL",
|
||||
name: "Nederlands (Nederland)",
|
||||
lcid: 0x0413,
|
||||
decimal: ',',
|
||||
group: '.',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '-',
|
||||
time_sep: ':',
|
||||
short_date: "dd-mm-yyyy",
|
||||
long_date: "dddd d mmmm yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd-mm-yyyy hh:mm",
|
||||
months: ["januari", "februari", "maart", "april", "mei", "juni", "juli", "augustus", "september", "oktober", "november", "december"],
|
||||
months_abbr: ["jan", "feb", "mrt", "apr", "mei", "jun", "jul", "aug", "sep", "okt", "nov", "dec"],
|
||||
months_genitive: None,
|
||||
weekdays: ["zondag", "maandag", "dinsdag", "woensdag", "donderdag", "vrijdag", "zaterdag"],
|
||||
weekdays_abbr: ["zo", "ma", "di", "wo", "do", "vr", "za"],
|
||||
am_pm: AM_PM,
|
||||
currency: "€",
|
||||
currency_format: "\"€\" #,##0.00",
|
||||
currency_format_int: "\"€\" #,##0",
|
||||
};
|
||||
|
||||
static PT_BR: Region = Region {
|
||||
tag: "pt-BR",
|
||||
name: "Português (Brasil)",
|
||||
lcid: 0x0416,
|
||||
decimal: ',',
|
||||
group: '.',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd, d \"de\" mmmm \"de\" yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd/mm/yyyy hh:mm",
|
||||
months: ["janeiro", "fevereiro", "março", "abril", "maio", "junho", "julho", "agosto", "setembro", "outubro", "novembro", "dezembro"],
|
||||
months_abbr: ["jan", "fev", "mar", "abr", "mai", "jun", "jul", "ago", "set", "out", "nov", "dez"],
|
||||
months_genitive: None,
|
||||
weekdays: ["domingo", "segunda-feira", "terça-feira", "quarta-feira", "quinta-feira", "sexta-feira", "sábado"],
|
||||
weekdays_abbr: ["dom", "seg", "ter", "qua", "qui", "sex", "sáb"],
|
||||
am_pm: AM_PM,
|
||||
currency: "R$",
|
||||
currency_format: "\"R$\" #,##0.00",
|
||||
currency_format_int: "\"R$\" #,##0",
|
||||
};
|
||||
|
||||
static PT_PT: Region = Region {
|
||||
tag: "pt-PT",
|
||||
name: "Português (Portugal)",
|
||||
lcid: 0x0816,
|
||||
decimal: ',',
|
||||
group: '\u{a0}',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "dd/mm/yyyy",
|
||||
long_date: "dddd, d \"de\" mmmm \"de\" yyyy",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd/mm/yyyy hh:mm",
|
||||
months: ["janeiro", "fevereiro", "março", "abril", "maio", "junho", "julho", "agosto", "setembro", "outubro", "novembro", "dezembro"],
|
||||
months_abbr: ["jan", "fev", "mar", "abr", "mai", "jun", "jul", "ago", "set", "out", "nov", "dez"],
|
||||
months_genitive: None,
|
||||
weekdays: ["domingo", "segunda-feira", "terça-feira", "quarta-feira", "quinta-feira", "sexta-feira", "sábado"],
|
||||
weekdays_abbr: ["dom", "seg", "ter", "qua", "qui", "sex", "sáb"],
|
||||
am_pm: AM_PM,
|
||||
currency: "€",
|
||||
currency_format: "#,##0.00 \"€\"",
|
||||
currency_format_int: "#,##0 \"€\"",
|
||||
};
|
||||
|
||||
static RU_RU: Region = Region {
|
||||
tag: "ru-RU",
|
||||
name: "Русский (Россия)",
|
||||
lcid: 0x0419,
|
||||
decimal: ',',
|
||||
group: '\u{a0}',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "d mmmm yyyy\" г.\"",
|
||||
short_time: "hh:mm",
|
||||
long_time: "hh:mm:ss",
|
||||
datetime: "dd.mm.yyyy hh:mm",
|
||||
months: ["январь", "февраль", "март", "апрель", "май", "июнь", "июль", "август", "сентябрь", "октябрь", "ноябрь", "декабрь"],
|
||||
months_abbr: ["янв.", "февр.", "март", "апр.", "май", "июнь", "июль", "авг.", "сент.", "окт.", "нояб.", "дек."],
|
||||
months_genitive: Some(["января", "февраля", "марта", "апреля", "мая", "июня", "июля", "августа", "сентября", "октября", "ноября", "декабря"]),
|
||||
weekdays: ["воскресенье", "понедельник", "вторник", "среда", "четверг", "пятница", "суббота"],
|
||||
weekdays_abbr: ["вс", "пн", "вт", "ср", "чт", "пт", "сб"],
|
||||
am_pm: AM_PM,
|
||||
currency: "₽",
|
||||
currency_format: "#,##0.00\u{a0}\"₽\"",
|
||||
currency_format_int: "#,##0\u{a0}\"₽\"",
|
||||
};
|
||||
|
||||
static CS_CZ: Region = Region {
|
||||
tag: "cs-CZ",
|
||||
name: "Čeština (Česko)",
|
||||
lcid: 0x0405,
|
||||
decimal: ',',
|
||||
group: '\u{a0}',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Dmy,
|
||||
date_sep: '.',
|
||||
time_sep: ':',
|
||||
short_date: "dd.mm.yyyy",
|
||||
long_date: "dddd d. mmmm yyyy",
|
||||
short_time: "h:mm",
|
||||
long_time: "h:mm:ss",
|
||||
datetime: "dd.mm.yyyy h:mm",
|
||||
months: ["leden", "únor", "březen", "duben", "květen", "červen", "červenec", "srpen", "září", "říjen", "listopad", "prosinec"],
|
||||
months_abbr: ["led", "úno", "bře", "dub", "kvě", "čvn", "čvc", "srp", "zář", "říj", "lis", "pro"],
|
||||
months_genitive: Some(["ledna", "února", "března", "dubna", "května", "června", "července", "srpna", "září", "října", "listopadu", "prosince"]),
|
||||
weekdays: ["neděle", "pondělí", "úterý", "středa", "čtvrtek", "pátek", "sobota"],
|
||||
weekdays_abbr: ["ne", "po", "út", "st", "čt", "pá", "so"],
|
||||
am_pm: ["dop.", "odp."],
|
||||
currency: "Kč",
|
||||
currency_format: "#,##0.00 \"Kč\"",
|
||||
currency_format_int: "#,##0 \"Kč\"",
|
||||
};
|
||||
|
||||
static JA_JP: Region = Region {
|
||||
tag: "ja-JP",
|
||||
name: "日本語 (日本)",
|
||||
lcid: 0x0411,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Ymd,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "yyyy/m/d",
|
||||
long_date: "yyyy\"年\"m\"月\"d\"日\"",
|
||||
short_time: "h:mm",
|
||||
long_time: "h:mm:ss",
|
||||
datetime: "yyyy/m/d h:mm",
|
||||
months: ["1月", "2月", "3月", "4月", "5月", "6月", "7月", "8月", "9月", "10月", "11月", "12月"],
|
||||
months_abbr: ["1月", "2月", "3月", "4月", "5月", "6月", "7月", "8月", "9月", "10月", "11月", "12月"],
|
||||
months_genitive: None,
|
||||
weekdays: ["日曜日", "月曜日", "火曜日", "水曜日", "木曜日", "金曜日", "土曜日"],
|
||||
weekdays_abbr: ["日", "月", "火", "水", "木", "金", "土"],
|
||||
am_pm: ["午前", "午後"],
|
||||
currency: "¥",
|
||||
currency_format: "\"¥\"#,##0",
|
||||
currency_format_int: "\"¥\"#,##0",
|
||||
};
|
||||
|
||||
static KO_KR: Region = Region {
|
||||
tag: "ko-KR",
|
||||
name: "한국어 (대한민국)",
|
||||
lcid: 0x0412,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Ymd,
|
||||
date_sep: '-',
|
||||
time_sep: ':',
|
||||
short_date: "yyyy-mm-dd",
|
||||
long_date: "yyyy\"년\" m\"월\" d\"일\" dddd",
|
||||
short_time: "AM/PM h:mm",
|
||||
long_time: "AM/PM h:mm:ss",
|
||||
datetime: "yyyy-mm-dd hh:mm",
|
||||
months: ["1월", "2월", "3월", "4월", "5월", "6월", "7월", "8월", "9월", "10월", "11월", "12월"],
|
||||
months_abbr: ["1월", "2월", "3월", "4월", "5월", "6월", "7월", "8월", "9월", "10월", "11월", "12월"],
|
||||
months_genitive: None,
|
||||
weekdays: ["일요일", "월요일", "화요일", "수요일", "목요일", "금요일", "토요일"],
|
||||
weekdays_abbr: ["일", "월", "화", "수", "목", "금", "토"],
|
||||
am_pm: ["오전", "오후"],
|
||||
currency: "₩",
|
||||
currency_format: "\"₩\"#,##0",
|
||||
currency_format_int: "\"₩\"#,##0",
|
||||
};
|
||||
|
||||
static ZH_CN: Region = Region {
|
||||
tag: "zh-CN",
|
||||
name: "中文 (中国)",
|
||||
lcid: 0x0804,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Ymd,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "yyyy/m/d",
|
||||
long_date: "yyyy\"年\"m\"月\"d\"日\"",
|
||||
short_time: "h:mm",
|
||||
long_time: "h:mm:ss",
|
||||
datetime: "yyyy/m/d h:mm",
|
||||
months: ["一月", "二月", "三月", "四月", "五月", "六月", "七月", "八月", "九月", "十月", "十一月", "十二月"],
|
||||
months_abbr: ["1月", "2月", "3月", "4月", "5月", "6月", "7月", "8月", "9月", "10月", "11月", "12月"],
|
||||
months_genitive: None,
|
||||
weekdays: ["星期日", "星期一", "星期二", "星期三", "星期四", "星期五", "星期六"],
|
||||
weekdays_abbr: ["周日", "周一", "周二", "周三", "周四", "周五", "周六"],
|
||||
am_pm: ["上午", "下午"],
|
||||
currency: "¥",
|
||||
currency_format: "\"¥\"#,##0.00",
|
||||
currency_format_int: "\"¥\"#,##0",
|
||||
};
|
||||
|
||||
static ZH_TW: Region = Region {
|
||||
tag: "zh-TW",
|
||||
name: "中文 (台灣)",
|
||||
lcid: 0x0404,
|
||||
decimal: '.',
|
||||
group: ',',
|
||||
grouping: Grouping::Thousands,
|
||||
date_order: DateOrder::Ymd,
|
||||
date_sep: '/',
|
||||
time_sep: ':',
|
||||
short_date: "yyyy/m/d",
|
||||
long_date: "yyyy\"年\"m\"月\"d\"日\"",
|
||||
short_time: "AM/PM hh:mm",
|
||||
long_time: "AM/PM hh:mm:ss",
|
||||
datetime: "yyyy/m/d hh:mm",
|
||||
months: ["一月", "二月", "三月", "四月", "五月", "六月", "七月", "八月", "九月", "十月", "十一月", "十二月"],
|
||||
months_abbr: ["1月", "2月", "3月", "4月", "5月", "6月", "7月", "8月", "9月", "10月", "11月", "12月"],
|
||||
months_genitive: None,
|
||||
weekdays: ["星期日", "星期一", "星期二", "星期三", "星期四", "星期五", "星期六"],
|
||||
weekdays_abbr: ["週日", "週一", "週二", "週三", "週四", "週五", "週六"],
|
||||
am_pm: ["上午", "下午"],
|
||||
currency: "NT$",
|
||||
currency_format: "\"NT$\"#,##0.00",
|
||||
currency_format_int: "\"NT$\"#,##0",
|
||||
};
|
||||
|
||||
static ALL: [&Region; 20] = [
|
||||
&EN_US, &EN_GB, &EN_IN, &EN_CH, &ES_ES, &DE_DE, &DE_CH, &FR_FR, &FR_CH, &IT_IT, &IT_CH, &NL_NL, &PT_BR, &PT_PT, &RU_RU, &CS_CZ, &JA_JP, &KO_KR,
|
||||
&ZH_CN, &ZH_TW,
|
||||
];
|
||||
|
||||
/// Every region, English (United States) first.
|
||||
pub fn all() -> &'static [&'static Region] {
|
||||
&ALL
|
||||
}
|
||||
|
||||
/// The default region, English (United States).
|
||||
pub fn default_region() -> &'static Region {
|
||||
&EN_US
|
||||
}
|
||||
|
||||
/// The region with exactly this tag (`pt-BR`; case and `_` don't matter).
|
||||
pub fn find(tag: &str) -> Option<&'static Region> {
|
||||
let t = tag.trim().replace('_', "-");
|
||||
ALL.iter().copied().find(|r| r.tag.eq_ignore_ascii_case(&t))
|
||||
}
|
||||
|
||||
/// The region for a tag as operating systems and browsers write it: `pt-BR`, `pt_BR.UTF-8`,
|
||||
/// `de`, `en-AU`, `zh-Hant`, `sr_RS@latin`. An exact match wins; otherwise the language's main
|
||||
/// region (with the usual variants: `en-AU`→`en-GB`, `pt-AO`→`pt-PT`, `zh-Hant`→`zh-TW`).
|
||||
/// An unknown language is English (United States).
|
||||
pub fn lookup(tag: &str) -> &'static Region {
|
||||
if let Some(r) = find(tag) {
|
||||
return r;
|
||||
}
|
||||
let base = tag.split(['.', '@']).next().unwrap_or("");
|
||||
let norm = base.trim().replace('_', "-");
|
||||
if let Some(r) = find(&norm) {
|
||||
return r;
|
||||
}
|
||||
let mut parts = norm.split('-').filter(|p| !p.is_empty());
|
||||
let lang = parts.next().unwrap_or("").to_ascii_lowercase();
|
||||
let mut script = String::new();
|
||||
let mut country = String::new();
|
||||
for p in parts {
|
||||
// An extension or private-use singleton (`-u-ca-gregory`, `-x-…`) ends the region part.
|
||||
if p.len() == 1 {
|
||||
break;
|
||||
}
|
||||
if p.len() == 4 && p.chars().all(|c| c.is_ascii_alphabetic()) {
|
||||
script = p.to_ascii_lowercase();
|
||||
} else if (p.len() == 2 && p.chars().all(|c| c.is_ascii_alphabetic())) || (p.len() == 3 && p.chars().all(|c| c.is_ascii_digit())) {
|
||||
country = p.to_ascii_uppercase();
|
||||
}
|
||||
}
|
||||
let tag = match lang.as_str() {
|
||||
"en" => match country.as_str() {
|
||||
"GB" | "AU" | "NZ" | "IE" => "en-GB",
|
||||
"IN" => "en-IN",
|
||||
"CH" | "LI" => "en-CH",
|
||||
_ => "en-US",
|
||||
},
|
||||
"pt" => match country.as_str() {
|
||||
"AO" | "CH" | "CV" | "FR" | "GQ" | "GW" | "LU" | "MO" | "MZ" | "ST" | "TL" | "PT" => "pt-PT",
|
||||
_ => "pt-BR",
|
||||
},
|
||||
"zh" => match country.as_str() {
|
||||
// An explicit territory wins over the script (`zh-Hant-CN` is China's conventions).
|
||||
"CN" | "SG" => "zh-CN",
|
||||
"TW" | "HK" | "MO" => "zh-TW",
|
||||
_ if script == "hant" => "zh-TW",
|
||||
_ => "zh-CN",
|
||||
},
|
||||
"es" => "es-ES",
|
||||
"de" => match country.as_str() {
|
||||
"CH" | "LI" => "de-CH",
|
||||
_ => "de-DE",
|
||||
},
|
||||
"fr" => match country.as_str() {
|
||||
"CH" | "LI" => "fr-CH",
|
||||
_ => "fr-FR",
|
||||
},
|
||||
"it" => match country.as_str() {
|
||||
"CH" | "LI" => "it-CH",
|
||||
_ => "it-IT",
|
||||
},
|
||||
"nl" => "nl-NL",
|
||||
"ru" => "ru-RU",
|
||||
"cs" => "cs-CZ",
|
||||
"ja" => "ja-JP",
|
||||
"ko" => "ko-KR",
|
||||
_ => "en-US",
|
||||
};
|
||||
find(tag).unwrap_or(&EN_US)
|
||||
}
|
||||
|
||||
impl Region {
|
||||
/// True for the characters that separate groups of digits in this region. A region that
|
||||
/// groups with a space also accepts the other space-like characters (NBSP, narrow NBSP, thin
|
||||
/// space), as people type and paste any of them; one that groups with an apostrophe accepts
|
||||
/// both `'` and `’`.
|
||||
pub fn is_group(&self, c: char) -> bool {
|
||||
c == self.group || (is_space_like(self.group) && is_space_like(c)) || (is_apostrophe_like(self.group) && is_apostrophe_like(c))
|
||||
}
|
||||
|
||||
/// True when the currency symbol comes before the number.
|
||||
pub fn currency_first(&self) -> bool {
|
||||
self.currency_format.starts_with('"')
|
||||
}
|
||||
|
||||
/// Month number (1–12) for a month name as typed: the full name, the abbreviation (with or
|
||||
/// without its period) or any prefix of the full name of three or more letters (`Sept`),
|
||||
/// ignoring case. `word` is one word: the genitive form used after a day (`января` in
|
||||
/// `1 января`) matches too, and the caller reads the day itself.
|
||||
pub fn month_from_name(&self, word: &str) -> Option<u32> {
|
||||
let w = word.trim().to_lowercase();
|
||||
// Where the abbreviations end in a period (`сент.`, `janv.`), any abbreviation may be typed
|
||||
// with one (`мар.`, `juin.`). Elsewhere a period is not part of a month name, so en-US
|
||||
// `Mar. 4, 2026` stays text as before.
|
||||
let dotted = self.months_abbr.iter().any(|a| a.ends_with('.'));
|
||||
let stem = if dotted { w.trim_end_matches('.') } else { w.as_str() };
|
||||
if stem.is_empty() {
|
||||
return None;
|
||||
}
|
||||
let long_enough = stem.chars().count() >= 3;
|
||||
for (i, full) in self.months.iter().enumerate() {
|
||||
let full_l = full.to_lowercase();
|
||||
let abbr_l = self.months_abbr.get(i).map(|a| a.to_lowercase()).unwrap_or_default();
|
||||
let gen_l = self.months_genitive.as_ref().and_then(|g| g.get(i)).map(|g| g.to_lowercase());
|
||||
let hit = w == full_l
|
||||
|| w == abbr_l
|
||||
|| stem == abbr_l.trim_end_matches('.')
|
||||
|| gen_l.as_deref() == Some(w.as_str())
|
||||
|| (long_enough && full_l.starts_with(stem))
|
||||
|| (long_enough && gen_l.as_deref().is_some_and(|g| g.starts_with(stem)));
|
||||
if hit {
|
||||
return Some(i as u32 + 1);
|
||||
}
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
/// The name of month `month` (1–12) for `mmmm`; `with_day` selects the genitive form where
|
||||
/// the region has one.
|
||||
pub fn month_name(&self, month: u32, with_day: bool) -> &'static str {
|
||||
let i = (month as usize).wrapping_sub(1);
|
||||
if with_day && let Some(g) = &self.months_genitive {
|
||||
return g.get(i).copied().unwrap_or("");
|
||||
}
|
||||
self.months.get(i).copied().unwrap_or("")
|
||||
}
|
||||
|
||||
/// True when the short time uses a 12-hour clock.
|
||||
pub fn is_12_hour(&self) -> bool {
|
||||
self.short_time.contains("AM/PM")
|
||||
}
|
||||
}
|
||||
|
||||
fn is_space_like(c: char) -> bool {
|
||||
matches!(c, ' ' | '\u{a0}' | '\u{202f}' | '\u{2009}')
|
||||
}
|
||||
|
||||
fn is_apostrophe_like(c: char) -> bool {
|
||||
matches!(c, '\'' | '\u{2019}')
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn lookup_accepts_system_and_browser_tags() {
|
||||
let cases = [
|
||||
("pt-BR", "pt-BR"),
|
||||
("pt_BR.UTF-8", "pt-BR"),
|
||||
("pt", "pt-BR"),
|
||||
("pt-PT", "pt-PT"),
|
||||
("pt-AO", "pt-PT"),
|
||||
("de", "de-DE"),
|
||||
("de-AT", "de-DE"),
|
||||
("de_CH.UTF-8", "de-CH"),
|
||||
("de-LI", "de-CH"),
|
||||
("fr-CH", "fr-CH"),
|
||||
("fr-LI", "fr-CH"),
|
||||
("fr-BE", "fr-FR"),
|
||||
("it_CH.UTF-8", "it-CH"),
|
||||
("it-LI", "it-CH"),
|
||||
("en-CH", "en-CH"),
|
||||
("nl_NL.UTF-8", "nl-NL"),
|
||||
("nl", "nl-NL"),
|
||||
("en-AU", "en-GB"),
|
||||
("en_GB.UTF-8", "en-GB"),
|
||||
("en-IN", "en-IN"),
|
||||
("en-CA", "en-US"),
|
||||
("zh-Hant", "zh-TW"),
|
||||
("zh-HK", "zh-TW"),
|
||||
("zh", "zh-CN"),
|
||||
// Extensions don't replace the region or script; an explicit territory beats the script.
|
||||
("en-GB-u-ca-gregory", "en-GB"),
|
||||
("zh-Hant-u-nu-latn", "zh-TW"),
|
||||
("zh-Hant-CN", "zh-CN"),
|
||||
("zh-Hans-TW", "zh-TW"),
|
||||
("cs_CZ.UTF-8", "cs-CZ"),
|
||||
("C", "en-US"),
|
||||
("", "en-US"),
|
||||
("xx-YY", "en-US"),
|
||||
];
|
||||
for (tag, expected) in cases {
|
||||
assert_eq!(lookup(tag).tag, expected, "{tag}");
|
||||
}
|
||||
assert_eq!(find("DE-de").map(|r| r.tag), Some("de-DE"));
|
||||
assert!(find("xx").is_none());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn every_region_is_consistent() {
|
||||
let mut tags = std::collections::HashSet::new();
|
||||
for r in all() {
|
||||
assert!(tags.insert(r.tag), "duplicate {}", r.tag);
|
||||
assert_ne!(r.decimal, r.group, "{}", r.tag);
|
||||
assert!(r.currency_format.contains(r.currency), "{}", r.tag);
|
||||
assert!(r.months.iter().all(|m| !m.is_empty()) && r.months_abbr.iter().all(|m| !m.is_empty()), "{}", r.tag);
|
||||
assert!(r.weekdays.iter().all(|m| !m.is_empty()) && r.weekdays_abbr.iter().all(|m| !m.is_empty()), "{}", r.tag);
|
||||
assert_eq!(find(r.tag).map(|x| x.tag), Some(r.tag));
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn month_names_match_loosely() {
|
||||
let de = find("de-DE").unwrap();
|
||||
assert_eq!(de.month_from_name("März"), Some(3));
|
||||
assert_eq!(de.month_from_name("mär"), Some(3));
|
||||
assert_eq!(de.month_from_name("DEZ"), Some(12));
|
||||
assert_eq!(de.month_from_name("de"), None);
|
||||
let gb = find("en-GB").unwrap();
|
||||
assert_eq!(gb.month_from_name("Sept"), Some(9));
|
||||
assert_eq!(gb.month_from_name("sep"), Some(9));
|
||||
let ru = find("ru-RU").unwrap();
|
||||
assert_eq!(ru.month_from_name("сентября"), Some(9));
|
||||
assert_eq!(ru.month_from_name("сент."), Some(9));
|
||||
// Abbreviations typed with a period the table doesn't carry.
|
||||
assert_eq!(ru.month_from_name("мар."), Some(3));
|
||||
assert_eq!(ru.month_from_name("июн."), Some(6));
|
||||
assert_eq!(find("fr-FR").unwrap().month_from_name("juin."), Some(6));
|
||||
assert_eq!(EN_US.month_from_name("Mar."), None);
|
||||
assert_eq!(de.month_from_name("."), None);
|
||||
assert_eq!(EN_US.month_from_name("Ma"), None);
|
||||
assert_eq!(EN_US.month_from_name("May"), Some(5));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn space_like_and_apostrophe_groups_are_interchangeable() {
|
||||
let fr = find("fr-FR").unwrap();
|
||||
assert!(fr.is_group('\u{a0}') && fr.is_group(' ') && fr.is_group('\u{202f}'));
|
||||
assert!(!find("de-DE").unwrap().is_group(' '));
|
||||
let ch = find("de-CH").unwrap();
|
||||
assert!(ch.is_group('\'') && ch.is_group('\u{2019}'));
|
||||
assert!(!ch.is_group(' ') && !find("de-DE").unwrap().is_group('\''));
|
||||
}
|
||||
}
|
||||
+10
-258
@@ -346,91 +346,19 @@ pub fn compare_numbers(a: f64, b: f64) -> Ordering {
|
||||
round15(a).partial_cmp(&round15(b)).unwrap_or(Ordering::Equal)
|
||||
}
|
||||
|
||||
/// Case-insensitive text order in three levels, as Excel's collation sorts: first by letters and
|
||||
/// digits alone (Arabic marks, tatweel and invisible format characters skipped, Arabic-Indic digits
|
||||
/// read as digits); then, letter by letter, by the marks on each letter (none first, then in
|
||||
/// Excel's mark order) and the digit script (ASCII, Arabic-Indic, Extended Arabic-Indic); then, to
|
||||
/// break a remaining tie, by every character (Unicode lowercase). Only text that is equal at the
|
||||
/// last level compares equal.
|
||||
/// Case-insensitive text order (Unicode lowercase).
|
||||
pub fn compare_text(a: &str, b: &str) -> Ordering {
|
||||
if !(a.is_ascii() && b.is_ascii()) {
|
||||
let primary = primary_chars(a).cmp(primary_chars(b));
|
||||
if primary != Ordering::Equal {
|
||||
return primary;
|
||||
}
|
||||
let secondary = secondary_weights(a).cmp(secondary_weights(b));
|
||||
if secondary != Ordering::Equal {
|
||||
return secondary;
|
||||
let mut ai = a.chars().flat_map(char::to_lowercase);
|
||||
let mut bi = b.chars().flat_map(char::to_lowercase);
|
||||
loop {
|
||||
match (ai.next(), bi.next()) {
|
||||
(None, None) => return Ordering::Equal,
|
||||
(None, Some(_)) => return Ordering::Less,
|
||||
(Some(_), None) => return Ordering::Greater,
|
||||
(Some(x), Some(y)) if x != y => return x.cmp(&y),
|
||||
_ => {}
|
||||
}
|
||||
}
|
||||
lowercase_chars(a).cmp(lowercase_chars(b))
|
||||
}
|
||||
|
||||
fn lowercase_chars(s: &str) -> impl Iterator<Item = char> + '_ {
|
||||
s.chars().flat_map(char::to_lowercase)
|
||||
}
|
||||
|
||||
/// Characters Excel's text order skips entirely: tatweel and invisible format characters (soft
|
||||
/// hyphen, combining grapheme joiner, zero-width (non-)joiner, bidi marks, embeddings, overrides
|
||||
/// and isolates, word joiner, byte order mark). The zero-width space is not one.
|
||||
fn ignorable(c: char) -> bool {
|
||||
matches!(
|
||||
c,
|
||||
'\u{00AD}'
|
||||
| '\u{034F}'
|
||||
| '\u{061C}'
|
||||
| '\u{0640}'
|
||||
| '\u{200C}'..='\u{200F}'
|
||||
| '\u{202A}'..='\u{202E}'
|
||||
| '\u{2060}'
|
||||
| '\u{2066}'..='\u{2069}'
|
||||
| '\u{FEFF}'
|
||||
)
|
||||
}
|
||||
|
||||
/// Arabic combining marks in the order Excel sorts a letter carrying them; an unmarked letter
|
||||
/// sorts before all of them. Measured in Excel 16.0 (build 20527, en-US) with one mark on a letter.
|
||||
/// U+06D6 and U+06D7 sort as letters in Excel, so they are not listed.
|
||||
const ARABIC_MARK_ORDER: [char; 48] = [
|
||||
'\u{06EC}', '\u{0611}', '\u{0612}', '\u{0613}', '\u{0614}', '\u{06DB}', '\u{0656}', '\u{0657}', '\u{0610}', '\u{06EB}', '\u{0615}', '\u{06D9}',
|
||||
'\u{06E4}', '\u{06E1}', '\u{06DA}', '\u{06DC}', '\u{06EA}', '\u{06E3}', '\u{06D8}', '\u{0655}', '\u{065F}', '\u{0654}', '\u{06E7}', '\u{0653}',
|
||||
'\u{06E8}', '\u{06ED}', '\u{06E2}', '\u{0658}', '\u{064B}', '\u{064C}', '\u{064D}', '\u{064E}', '\u{064F}', '\u{0650}', '\u{0652}', '\u{0651}',
|
||||
'\u{0670}', '\u{0659}', '\u{065A}', '\u{065B}', '\u{065C}', '\u{065D}', '\u{065E}', '\u{0616}', '\u{0617}', '\u{0618}', '\u{0619}', '\u{061A}',
|
||||
];
|
||||
|
||||
/// Rank of an Arabic mark that letter order skips (its index in `ARABIC_MARK_ORDER`).
|
||||
fn arabic_mark_rank(c: char) -> Option<u16> {
|
||||
ARABIC_MARK_ORDER.iter().position(|&m| m == c).and_then(|i| u16::try_from(i).ok())
|
||||
}
|
||||
|
||||
/// Digit script of `c` (0 ASCII, 1 Arabic-Indic, 2 Extended Arabic-Indic) and its ASCII digit.
|
||||
fn digit(c: char) -> Option<(u16, char)> {
|
||||
let (script, zero) = match c {
|
||||
'0'..='9' => return Some((0, c)),
|
||||
'\u{0660}'..='\u{0669}' => (1, 0x0660),
|
||||
'\u{06F0}'..='\u{06F9}' => (2, 0x06F0),
|
||||
_ => return None,
|
||||
};
|
||||
char::from_u32(u32::from(c) - zero + u32::from('0')).map(|d| (script, d))
|
||||
}
|
||||
|
||||
/// The characters that decide text order before any tie-break.
|
||||
fn primary_chars(s: &str) -> impl Iterator<Item = char> + '_ {
|
||||
lowercase_chars(s).filter_map(|c| if ignorable(c) || arabic_mark_rank(c).is_some() { None } else { Some(digit(c).map_or(c, |(_, d)| d)) })
|
||||
}
|
||||
|
||||
/// The second level, for text whose primary characters are equal: per letter, its digit script and
|
||||
/// then its marks, the letters' lists closed by 0 so that an unmarked letter sorts first. Marks
|
||||
/// before the first letter form a list of their own.
|
||||
fn secondary_weights(s: &str) -> impl Iterator<Item = u16> + '_ {
|
||||
lowercase_chars(s)
|
||||
.filter(|&c| !ignorable(c))
|
||||
.flat_map(|c| match arabic_mark_rank(c) {
|
||||
Some(rank) => [Some(rank.saturating_add(1)), None],
|
||||
None => [Some(0), Some(digit(c).map_or(0, |(script, _)| script))],
|
||||
})
|
||||
.flatten()
|
||||
.chain(std::iter::once(0))
|
||||
}
|
||||
|
||||
/// Sorting order used by Sort (ascending): numbers, text, booleans, errors, then blanks last.
|
||||
@@ -529,182 +457,6 @@ mod tests {
|
||||
assert_eq!(sort_compare(&Value::Empty, &Value::Number(1.0)), Ordering::Greater);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_diacritics_sort_with_their_base_letter() {
|
||||
assert_eq!(compare_text("بَاب", "بيت"), Ordering::Less);
|
||||
assert_eq!(compare_text("بيت", "بَاب"), Ordering::Greater);
|
||||
assert_eq!(compare_text("كِتاب", "كتب"), Ordering::Less);
|
||||
assert_eq!(compare_text("رحمٰن", "رحمي"), Ordering::Less);
|
||||
assert_eq!(compare(&Value::from("بَاب"), &Value::from("بيت")), Ordering::Less);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_indic_digits_sort_with_digits() {
|
||||
assert_eq!(compare_text("١٢", "ب"), Ordering::Less);
|
||||
assert_eq!(compare_text("۱۲", "ب"), Ordering::Less);
|
||||
assert_eq!(compare_text("١", "a"), Ordering::Less);
|
||||
assert_eq!(compare_text("١٢", "13"), Ordering::Less);
|
||||
assert_eq!(compare_text("۱۴", "13"), Ordering::Greater);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn tatweel_and_bidi_marks_do_not_move_sort_order() {
|
||||
assert_eq!(compare_text("بـاب", "بت"), Ordering::Less);
|
||||
assert_eq!(compare_text("\u{200F}apple", "banana"), Ordering::Less);
|
||||
assert_eq!(compare_text("\u{200E}zebra", "apple"), Ordering::Greater);
|
||||
assert_eq!(compare_text("\u{061C}a", "b"), Ordering::Less);
|
||||
assert_eq!(compare_text("a\u{200C}b", "ac"), Ordering::Less);
|
||||
assert_eq!(compare_text("a\u{200D}b", "ac"), Ordering::Less);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn primary_ties_keep_equality_unchanged() {
|
||||
assert_eq!(compare_text("ABC", "abc"), Ordering::Equal);
|
||||
assert_eq!(compare_text("باب", "بَاب"), Ordering::Less);
|
||||
assert_eq!(compare_text("12", "١٢"), Ordering::Less);
|
||||
assert_eq!(compare_text("a", "\u{200F}a"), Ordering::Less);
|
||||
let samples = [
|
||||
"",
|
||||
"abc",
|
||||
"ABC",
|
||||
"abd",
|
||||
"ab",
|
||||
"باب",
|
||||
"بَاب",
|
||||
"بِاب",
|
||||
"بـاب",
|
||||
"بيت",
|
||||
"١٢",
|
||||
"۱۲",
|
||||
"12",
|
||||
"\u{200F}12",
|
||||
"\u{200F}a",
|
||||
"a",
|
||||
"A\u{200C}",
|
||||
"a\u{200F}",
|
||||
"b",
|
||||
"Straße",
|
||||
"STRASSE",
|
||||
"ΣΑΣ",
|
||||
"σας",
|
||||
"İ",
|
||||
"i̇",
|
||||
];
|
||||
let lowered = |s: &str| s.chars().flat_map(char::to_lowercase).collect::<String>();
|
||||
for a in samples {
|
||||
for b in samples {
|
||||
let order = compare_text(a, b);
|
||||
assert_eq!(order.is_eq(), lowered(a) == lowered(b), "{a:?} vs {b:?}");
|
||||
assert_eq!(order, compare_text(b, a).reverse(), "{a:?} vs {b:?}");
|
||||
for c in samples {
|
||||
if order.is_le() && compare_text(b, c).is_le() {
|
||||
assert!(compare_text(a, c).is_le(), "{a:?} <= {b:?} <= {c:?}");
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Sorts `expected` from reversed and rotated starts and checks each comes back as `expected`.
|
||||
#[track_caller]
|
||||
fn assert_sorts_to(expected: &[&str]) {
|
||||
let mut reversed: Vec<&str> = expected.iter().rev().copied().collect();
|
||||
reversed.sort_by(|a, b| compare_text(a, b));
|
||||
assert_eq!(reversed, expected);
|
||||
let mut rotated: Vec<&str> = expected.to_vec();
|
||||
rotated.rotate_left(expected.len() / 2);
|
||||
rotated.sort_by(|a, b| compare_text(a, b));
|
||||
assert_eq!(rotated, expected);
|
||||
}
|
||||
|
||||
// The expected orders below are Excel's, observed with Range.Sort, =SORT() and the `<`
|
||||
// operator in Excel 16.0 (build 20527, en-US). Strings Excel ranks equal are left out.
|
||||
|
||||
#[test]
|
||||
fn excel_orders_a_letter_by_its_mark() {
|
||||
assert_sorts_to(&[
|
||||
"باب",
|
||||
"ب\u{0611}اب",
|
||||
"ب\u{0610}اب",
|
||||
"ب\u{0654}اب",
|
||||
"ب\u{064B}اب",
|
||||
"ب\u{064C}اب",
|
||||
"ب\u{064D}اب",
|
||||
"ب\u{064E}اب",
|
||||
"ب\u{064F}اب",
|
||||
"ب\u{0650}اب",
|
||||
"ب\u{0652}اب",
|
||||
"ب\u{0651}اب",
|
||||
"ب\u{0670}اب",
|
||||
"ب\u{065E}اب",
|
||||
"ب\u{061A}اب",
|
||||
]);
|
||||
// Quranic marks are skipped by letter order like harakat (hamza sorts before alef).
|
||||
assert_eq!(compare_text("ب\u{0610}ء", "باب"), Ordering::Less);
|
||||
assert_eq!(compare_text("ب\u{06E1}ء", "باب"), Ordering::Less);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn excel_compares_marks_letter_by_letter() {
|
||||
assert_sorts_to(&[
|
||||
"بت",
|
||||
"بت\u{064E}",
|
||||
"بت\u{064F}",
|
||||
"بت\u{0650}",
|
||||
"ب\u{064E}ت",
|
||||
"ب\u{064E}ت\u{064E}",
|
||||
"ب\u{064E}ت\u{064F}",
|
||||
"ب\u{064F}ت",
|
||||
"ب\u{064F}ت\u{064E}",
|
||||
"ب\u{0650}ت\u{0650}",
|
||||
]);
|
||||
assert_sorts_to(&["مـحـمـد", "محم\u{0651}د", "م\u{064E}حمد", "م\u{064F}ح\u{064E}م\u{064E}\u{0651}د", "محمدي"]);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn excel_weighs_marks_and_digit_scripts_by_position() {
|
||||
// A mark on an earlier letter outranks the digit script or letter after it, whatever the
|
||||
// code point of that next character.
|
||||
assert_sorts_to(&[
|
||||
"ب",
|
||||
"ب\u{064E}",
|
||||
"ب1",
|
||||
"ب١",
|
||||
"ب۱",
|
||||
"ب\u{064E}1",
|
||||
"ب\u{064E}١",
|
||||
"ب\u{064E}۱",
|
||||
"بa",
|
||||
"ب\u{064E}a",
|
||||
"با",
|
||||
"ب\u{064E}ا",
|
||||
"ب\u{067E}",
|
||||
"ب\u{064E}\u{067E}",
|
||||
"ب\u{06CC}",
|
||||
"ب\u{064E}\u{06CC}",
|
||||
]);
|
||||
// And an earlier digit script outranks a later mark.
|
||||
assert_sorts_to(&["12", "1٢", "١2", "١٢", "1ب", "1ب\u{064E}", "1ب\u{0650}", "١ب", "١ب\u{064E}", "١ب\u{0650}", "۱ب", "۱ب\u{064E}"]);
|
||||
assert_sorts_to(&["ب1ب", "ب1ب\u{064E}", "ب١ب", "ب\u{064E}ب", "بب1", "بب١", "ب\u{064E}ب1"]);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn excel_skips_invisible_format_characters() {
|
||||
for c in ['\u{00AD}', '\u{034F}', '\u{061C}', '\u{200C}', '\u{200D}', '\u{200E}', '\u{200F}', '\u{202A}', '\u{202B}']
|
||||
.into_iter()
|
||||
.chain(['\u{202C}', '\u{202D}', '\u{202E}', '\u{2060}', '\u{2066}', '\u{2067}', '\u{2068}', '\u{2069}', '\u{FEFF}'])
|
||||
{
|
||||
let s = format!("{c}b");
|
||||
assert_eq!(compare_text(&s, "a"), Ordering::Greater, "{s:?}");
|
||||
assert_eq!(compare_text(&s, "ba"), Ordering::Less, "{s:?}");
|
||||
assert_eq!(compare_text(&format!("ب{c}\u{064E}اب"), "باب\u{064B}"), Ordering::Greater, "{c:?}");
|
||||
}
|
||||
// Not even as a tie-break: a leading mark doesn't push text past marked or other-script forms.
|
||||
assert_sorts_to(&["\u{200F}باب", "ب\u{064B}اب", "ب\u{064E}اب", "ب\u{0670}اب"]);
|
||||
assert_sorts_to(&["\u{200F}12", "١٢", "۱۲"]);
|
||||
assert_sorts_to(&["ب\u{200F}ب\u{064E}", "\u{200F}ب\u{064E}ب"]);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn errors() {
|
||||
for e in CellError::ALL {
|
||||
|
||||
@@ -118,7 +118,7 @@ pub const CATALOG: &[(&str, &str, &str, &str)] = &[
|
||||
("insert.picture", "Insert", "Illustrations", "Pictures"),
|
||||
("insert.shape", "Insert", "Illustrations", "Shapes"),
|
||||
("insert.icons", "Insert", "Illustrations", "Icons"),
|
||||
("insert.smartArt", "Insert", "Illustrations", "Diagram"),
|
||||
("insert.smartArt", "Insert", "Illustrations", "SmartArt"),
|
||||
("insert.screenshot", "Insert", "Illustrations", "Screenshot"),
|
||||
("insert.recommendedCharts", "Insert", "Charts", "Recommended Charts"),
|
||||
("insert.chart", "Insert", "Charts", "Charts"),
|
||||
@@ -131,7 +131,7 @@ pub const CATALOG: &[(&str, &str, &str, &str)] = &[
|
||||
("review.newComment", "Insert", "Comments", "Comment"),
|
||||
("insert.textBox", "Insert", "Text", "Text Box"),
|
||||
("pageLayout.headerFooter", "Insert", "Text", "Header & Footer"),
|
||||
("insert.wordArt", "Insert", "Text", "Decorative Text"),
|
||||
("insert.wordArt", "Insert", "Text", "WordArt"),
|
||||
("insert.signatureLine", "Insert", "Text", "Signature Line"),
|
||||
("insert.object", "Insert", "Text", "Object"),
|
||||
("insert.equation", "Insert", "Symbols", "Equation"),
|
||||
@@ -340,15 +340,3 @@ pub fn parity() -> (usize, usize) {
|
||||
let done = ids.iter().filter(|id| crate::find_command(id).is_some()).count();
|
||||
(done, ids.len())
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod feature_name_tests {
|
||||
#[test]
|
||||
fn own_feature_labels_keep_command_ids() {
|
||||
for (id, expected) in [("insert.smartArt", "Diagram"), ("insert.wordArt", "Decorative Text")] {
|
||||
let entries: Vec<_> = super::CATALOG.iter().filter(|entry| entry.0 == id).collect();
|
||||
assert_eq!(entries.len(), 1, "{id} appears once");
|
||||
assert_eq!(entries[0].3, expected);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
+4
-158
@@ -1,7 +1,7 @@
|
||||
//! Conditional formatting evaluation: what extra style, data bar, colour scale or icon a cell
|
||||
//! gets.
|
||||
|
||||
use std::collections::{HashMap, HashSet};
|
||||
use std::collections::HashMap;
|
||||
|
||||
use gridcraft_core::{CellRef, RangeRef, Value, compare};
|
||||
use gridcraft_model::{CfOperator, CfRule, CfValueKind, Color, CondFormat, PatternType, Sheet, Style, Underline, Workbook, default_icon_thresholds};
|
||||
@@ -47,8 +47,6 @@ enum StatsNeed {
|
||||
fn stats_need(rule: &CfRule) -> Option<StatsNeed> {
|
||||
match rule {
|
||||
CfRule::Duplicate { .. } | CfRule::Unique { .. } => Some(StatsNeed::Counts),
|
||||
// only two- and three-colour scales paint: no statistics for the others
|
||||
CfRule::ColorScale { stops } if !matches!(stops.len(), 2 | 3) => None,
|
||||
CfRule::Top10 { .. } | CfRule::AboveAverage { .. } | CfRule::ColorScale { .. } | CfRule::DataBar { .. } | CfRule::IconSet { .. } => {
|
||||
Some(StatsNeed::Numbers)
|
||||
}
|
||||
@@ -72,58 +70,15 @@ thread_local! {
|
||||
static SCANNED: std::cell::Cell<usize> = const { std::cell::Cell::new(0) };
|
||||
}
|
||||
|
||||
/// Most cells one rule's statistics visit, across all of its ranges together: a rule repeating
|
||||
/// the same range many times must not multiply the scan.
|
||||
const STATS_SCAN_LIMIT: usize = 2_000_000;
|
||||
|
||||
/// Steps through the spilled cells inside `r`, row-major: `Some(cell)` for each, and `None` for a
|
||||
/// jump past the spills left or right of the columns on a row (one step per row, however many
|
||||
/// there are), so unrelated spills between a narrow range's corners cost little.
|
||||
fn spill_steps(sh: &Sheet, r: RangeRef) -> impl Iterator<Item = Option<CellRef>> + '_ {
|
||||
let mut cur = Some(r.start);
|
||||
std::iter::from_fn(move || {
|
||||
let at = cur.filter(|at| *at <= r.end)?;
|
||||
let (&c, _) = sh.spill.range(at..=r.end).next()?;
|
||||
if c.col < r.start.col {
|
||||
cur = Some(CellRef::new(c.row, r.start.col));
|
||||
Some(None)
|
||||
} else if c.col > r.end.col {
|
||||
cur = c.row.checked_add(1).map(|row| CellRef::new(row, r.start.col));
|
||||
Some(None)
|
||||
} else {
|
||||
cur = c.col.checked_add(1).map(|col| CellRef::new(c.row, col));
|
||||
Some(Some(c))
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
fn numbers_in(sh: &Sheet, ranges: &[RangeRef], need: StatsNeed) -> (Vec<f64>, HashMap<String, usize>) {
|
||||
let mut nums = Vec::new();
|
||||
let mut counts = HashMap::new();
|
||||
let mut budget = STATS_SCAN_LIMIT;
|
||||
let mut seen: HashSet<RangeRef> = HashSet::new();
|
||||
let used = sh.used_range();
|
||||
for r in ranges {
|
||||
if budget == 0 {
|
||||
break;
|
||||
}
|
||||
let r = match used.and_then(|u| r.intersection(&u)) {
|
||||
let r = match sh.used_range().and_then(|u| r.intersection(&u)) {
|
||||
Some(x) => x,
|
||||
None => continue,
|
||||
};
|
||||
// The same range listed twice counts its cells once.
|
||||
if !seen.insert(r) {
|
||||
continue;
|
||||
}
|
||||
// Only stored and spilled cells hold values: blank cells of a sparse range cost nothing,
|
||||
// so a value far down a huge range is still counted. Every cell looked at is charged.
|
||||
let stored = sh.cells.iter_range(r).map(|(c, _)| (Some(c), true));
|
||||
let spilled = spill_steps(sh, r).map(|c| (c, false));
|
||||
for (c, is_stored) in stored.chain(spilled).take(budget) {
|
||||
budget -= 1;
|
||||
// a skip over spills outside the columns (charged, not read), or a spilled value
|
||||
// under a stored cell, which was read with it
|
||||
let Some(c) = c.filter(|c| is_stored || sh.cells.get(*c).is_none()) else { continue };
|
||||
for c in r.iter().take(2_000_000) {
|
||||
#[cfg(test)]
|
||||
SCANNED.with(|n| n.set(n.get() + 1));
|
||||
let v = sh.value(c);
|
||||
@@ -323,8 +278,7 @@ impl CfCache {
|
||||
_ => None,
|
||||
},
|
||||
CfRule::ColorScale { stops } => {
|
||||
// Only two- and three-colour scales paint: evaluate no thresholds for others.
|
||||
if let Some(n) = v.as_f64().filter(|_| matches!(stops.len(), 2 | 3)) {
|
||||
if let Some(n) = v.as_f64() {
|
||||
let pts: Vec<(f64, [u8; 3])> = stops
|
||||
.iter()
|
||||
.map(|(k, col)| (threshold(wb, sheet, c, st, k), col.resolve(&wb.theme).unwrap_or([255, 255, 255])))
|
||||
@@ -414,9 +368,6 @@ impl CfCache {
|
||||
/// Adds the parts `lower` sets that `style` leaves unset (differential styles: default parts
|
||||
/// are unset).
|
||||
fn layer_under(style: &mut Style, lower: &Style) {
|
||||
if style.align.reading_order.is_none() {
|
||||
style.align.reading_order = lower.align.reading_order;
|
||||
}
|
||||
if style.fill.pattern == PatternType::None {
|
||||
style.fill = lower.fill;
|
||||
}
|
||||
@@ -466,28 +417,6 @@ mod tests {
|
||||
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn reading_order_respects_conditional_rule_priority_and_explicit_context() {
|
||||
for (higher, expected) in [
|
||||
(json!({"font": {"bold": true}}), "RightToLeft"),
|
||||
(json!({"align": {"reading_order": "Context"}}), "Context"),
|
||||
(json!({"align": {"reading_order": "LeftToRight"}}), "LeftToRight"),
|
||||
] {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("cell.set", json!({"cell": "A1", "input": "مرحبا 123 ABC"})).unwrap();
|
||||
for style in [json!({"align": {"reading_order": "RightToLeft"}}), higher] {
|
||||
s.execute("home.conditionalFormat", json!({"range": "A1", "rule": {"type": "noBlanks", "style": style}})).unwrap();
|
||||
}
|
||||
let wb = &s.doc().unwrap().wb;
|
||||
let (reopened, _) = gridcraft_xlsx::read_xlsx(&gridcraft_xlsx::write_xlsx(wb).unwrap()).unwrap();
|
||||
for wb in [wb, &reopened] {
|
||||
let style = CfCache::new().look(wb, 0, CellRef::new(0, 0)).unwrap().style.unwrap();
|
||||
assert_eq!(serde_json::to_value(style.align).unwrap()["reading_order"], expected);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn every_matching_rule_applies() {
|
||||
let mut s = crate::Session::new();
|
||||
@@ -631,89 +560,6 @@ mod tests {
|
||||
assert_eq!(SCANNED.with(|n| n.get()), 10_000);
|
||||
}
|
||||
|
||||
/// A rule listing the same range many times scans it once, and never more than the shared
|
||||
/// limit however many ranges it holds.
|
||||
#[test]
|
||||
fn repeated_ranges_share_one_scan_budget() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[1], [2], [3]]})).unwrap();
|
||||
s.execute("range.setValues", json!({"range": "XFD1048576", "values": [[4]]})).unwrap();
|
||||
let d = s.doc().unwrap();
|
||||
let sh = d.wb.sheet(0).unwrap();
|
||||
let dense = RangeRef::parse("A1:XFD1048576").unwrap();
|
||||
let small = RangeRef::parse("A1:A3").unwrap();
|
||||
SCANNED.with(|n| n.set(0));
|
||||
let st = RuleStats::build(sh, &vec![small; 1000], StatsNeed::Numbers);
|
||||
assert_eq!(st.sorted, vec![1.0, 2.0, 3.0]);
|
||||
assert_eq!(SCANNED.with(|n| n.get()), 3);
|
||||
let shifted: Vec<RangeRef> = (0..50).map(|i| RangeRef::new(CellRef::new(i, 0), dense.end)).collect();
|
||||
SCANNED.with(|n| n.set(0));
|
||||
RuleStats::build(sh, &shifted, StatsNeed::Numbers);
|
||||
assert!(SCANNED.with(|n| n.get()) <= STATS_SCAN_LIMIT);
|
||||
}
|
||||
|
||||
/// Blank cells of a sparse range are not charged to the scan budget: a value at the end of two
|
||||
/// million-row columns still counts, so duplicates are found.
|
||||
#[test]
|
||||
fn sparse_ranges_keep_their_values_under_the_scan_budget() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[7]]})).unwrap();
|
||||
s.execute("range.setValues", json!({"range": "B1048576", "values": [[7]]})).unwrap();
|
||||
let d = s.doc().unwrap();
|
||||
let sh = d.wb.sheet(0).unwrap();
|
||||
let ranges = [RangeRef::parse("A1:A1048576").unwrap(), RangeRef::parse("B1:B1048576").unwrap()];
|
||||
SCANNED.with(|n| n.set(0));
|
||||
let st = RuleStats::build(sh, &ranges, StatsNeed::Counts);
|
||||
assert_eq!(st.counts.get("7"), Some(&2));
|
||||
assert_eq!(SCANNED.with(|n| n.get()), 2);
|
||||
}
|
||||
|
||||
/// Spills left and right of a narrow range, between its corners, are skipped a row at a time:
|
||||
/// they do not use the budget that the in-range spills need.
|
||||
#[test]
|
||||
fn spills_outside_a_narrow_range_do_not_use_its_scan_budget() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
let d = s.doc().unwrap();
|
||||
let mut sh = d.wb.sheet(0).unwrap().clone();
|
||||
for row in 0..1000 {
|
||||
for col in [0u32, 5, 6, 7] {
|
||||
sh.spill.insert(CellRef::new(row, col), Value::Number(1.0));
|
||||
}
|
||||
}
|
||||
let r = RangeRef::parse("F1:G1000").unwrap();
|
||||
let steps: Vec<_> = spill_steps(&sh, r).collect();
|
||||
assert_eq!(steps.iter().flatten().count(), 2000);
|
||||
// column A is skipped one step a row, never one per cell
|
||||
assert!(steps.len() <= 2000 + 1000, "{}", steps.len());
|
||||
let narrow = RangeRef::parse("C1:C1000").unwrap();
|
||||
assert_eq!(spill_steps(&sh, narrow).flatten().count(), 0);
|
||||
}
|
||||
|
||||
/// A colour scale with an unsupported number of stops paints nothing and evaluates none of them.
|
||||
#[test]
|
||||
fn colour_scales_with_other_stop_counts_paint_nothing() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[1], [5]]})).unwrap();
|
||||
let mut wb: Workbook = (*s.doc().unwrap().wb).clone();
|
||||
let stop = (CfValueKind::Formula("=1".into()), Color::default());
|
||||
let rule = |n: usize| CondFormat {
|
||||
ranges: vec![RangeRef::parse("A1:A2").unwrap()],
|
||||
rule: CfRule::ColorScale { stops: vec![stop.clone(); n] },
|
||||
priority: 1,
|
||||
stop_if_true: false,
|
||||
};
|
||||
wb.sheet_mut(0).unwrap().cond_formats.push(rule(4));
|
||||
let mut cache = CfCache::new();
|
||||
assert_eq!(cache.look(&wb, 0, CellRef::parse("A1").unwrap()).and_then(|l| l.scale_fill), None);
|
||||
wb.sheet_mut(0).unwrap().cond_formats = vec![rule(2)];
|
||||
let mut cache = CfCache::new();
|
||||
assert!(cache.look(&wb, 0, CellRef::parse("A1").unwrap()).and_then(|l| l.scale_fill).is_some());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn cell_value_operator_names() {
|
||||
let mut s = crate::Session::new();
|
||||
|
||||
@@ -10,10 +10,6 @@ use crate::io::esc;
|
||||
/// Widest or tallest chart written; larger anchors are clamped.
|
||||
const MAX_SIDE: f32 = 16_384.0;
|
||||
|
||||
/// Most bytes of SVG written for the charts of one web page together. A chart past what is
|
||||
/// left makes the export fail rather than grow without bound or be dropped without a word.
|
||||
pub(crate) const MAX_PAGE_SVG_BYTES: usize = 64 * 1024 * 1024;
|
||||
|
||||
struct HelveticaMeasure;
|
||||
|
||||
impl gridcraft_chart::Measure for HelveticaMeasure {
|
||||
@@ -22,40 +18,24 @@ impl gridcraft_chart::Measure for HelveticaMeasure {
|
||||
}
|
||||
}
|
||||
|
||||
/// `chart` of sheet `sheet` as an `<svg>` element placed by `style`, `None` for a degenerate size,
|
||||
/// or an error for an element longer than `budget` bytes.
|
||||
pub(crate) fn chart_svg(wb: &Workbook, sheet: usize, chart: &Chart, style: &str, budget: usize) -> crate::Result<Option<String>> {
|
||||
/// `chart` of sheet `sheet` as an `<svg>` element placed by `style`, or `None` for a degenerate size.
|
||||
pub(crate) fn chart_svg(wb: &Workbook, sheet: usize, chart: &Chart, style: &str) -> Option<String> {
|
||||
let (w, h) = (chart.anchor.width.min(MAX_SIDE), chart.anchor.height.min(MAX_SIDE));
|
||||
if !(w >= 1.0 && h >= 1.0) {
|
||||
return Ok(None);
|
||||
return None;
|
||||
}
|
||||
let data = gridcraft_chart::resolve(wb, sheet, chart);
|
||||
let prims = gridcraft_chart::render(chart, &data, w, h, &HelveticaMeasure);
|
||||
match prims_svg_within(&prims, w, h, style, budget) {
|
||||
Some(svg) => Ok(Some(svg)),
|
||||
None => Err(crate::EngineError::Other(format!(
|
||||
"the charts would make the web page over {} MiB. Save as .xlsx or PDF instead.",
|
||||
MAX_PAGE_SVG_BYTES / (1024 * 1024)
|
||||
))),
|
||||
}
|
||||
Some(prims_svg(&prims, w, h, style))
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
fn prims_svg(prims: &[Prim], w: f32, h: f32, style: &str) -> String {
|
||||
prims_svg_within(prims, w, h, style, usize::MAX).unwrap_or_default()
|
||||
}
|
||||
|
||||
/// The `<svg>` element, or `None` as soon as it grows past `budget` bytes.
|
||||
fn prims_svg_within(prims: &[Prim], w: f32, h: f32, style: &str, budget: usize) -> Option<String> {
|
||||
let mut out = String::new();
|
||||
let _ = write!(
|
||||
out,
|
||||
"<svg xmlns=\"http://www.w3.org/2000/svg\" width=\"{w}\" height=\"{h}\" viewBox=\"0 0 {w} {h}\" font-family=\"Helvetica,Arial,sans-serif\" style=\"{style}\">"
|
||||
);
|
||||
for p in prims {
|
||||
if out.len() > budget {
|
||||
return None;
|
||||
}
|
||||
match p {
|
||||
Prim::Rect { x, y, w, h, fill, stroke, radius } => {
|
||||
let _ = write!(out, "<rect x=\"{}\" y=\"{}\" width=\"{}\" height=\"{}\"", num(*x), num(*y), num(*w), num(*h));
|
||||
@@ -112,7 +92,7 @@ fn prims_svg_within(prims: &[Prim], w: f32, h: f32, style: &str, budget: usize)
|
||||
}
|
||||
}
|
||||
out.push_str("</svg>");
|
||||
(out.len() <= budget).then_some(out)
|
||||
out
|
||||
}
|
||||
|
||||
/// Fill and stroke attributes of an open shape element; the caller closes it.
|
||||
@@ -169,17 +149,6 @@ fn num(v: f32) -> String {
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn an_svg_past_its_byte_budget_is_left_out() {
|
||||
let rect = Prim::Rect { x: 1.0, y: 2.0, w: 3.0, h: 4.0, fill: Some([255, 0, 0, 255]), stroke: None, radius: 0.0 };
|
||||
let prims = vec![rect; 10_000];
|
||||
let full = prims_svg(&prims, 100.0, 50.0, "");
|
||||
assert!(full.len() > 100_000);
|
||||
assert_eq!(prims_svg_within(&prims, 100.0, 50.0, "", full.len()), Some(full.clone()));
|
||||
assert_eq!(prims_svg_within(&prims, 100.0, 50.0, "", full.len() - 1), None);
|
||||
assert_eq!(prims_svg_within(&prims, 100.0, 50.0, "", 1_000), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn every_primitive_becomes_an_svg_element() {
|
||||
let red = [255, 0, 0, 255];
|
||||
|
||||
+5
-150
@@ -366,10 +366,6 @@ pub struct Ctx<'a> {
|
||||
pub wb: Workbook,
|
||||
pub changed: Vec<Key>,
|
||||
pub structural: bool,
|
||||
/// Recalculate every formula without the other effects of `structural` (table headers are not
|
||||
/// resynchronized, array formulas are still checked): for edits that change only what a
|
||||
/// function reads, like sheet protection.
|
||||
pub recalc_all: bool,
|
||||
pub sel: &'a mut Selection,
|
||||
/// Rows whose automatic height should be recomputed (sheet, row).
|
||||
pub fit_rows: Vec<(usize, u32)>,
|
||||
@@ -404,14 +400,13 @@ pub(crate) fn commit<R>(d: &mut DocState, f: impl FnOnce(&mut Ctx) -> Result<R>)
|
||||
wb: (*d.wb).clone(),
|
||||
changed: Vec::new(),
|
||||
structural: false,
|
||||
recalc_all: false,
|
||||
sel: &mut sel,
|
||||
fit_rows: Vec::new(),
|
||||
protection_checked: false,
|
||||
edits: Vec::new(),
|
||||
};
|
||||
let r = f(&mut ctx)?;
|
||||
let Ctx { mut wb, mut changed, structural, recalc_all, mut fit_rows, protection_checked, edits, .. } = ctx;
|
||||
let Ctx { mut wb, mut changed, structural, mut fit_rows, protection_checked, edits, .. } = ctx;
|
||||
// Equations whose source changed (or that arrived without a layout, as from a paste) are
|
||||
// laid out here, so painting only ever reads the cache.
|
||||
wb.refresh_equations();
|
||||
@@ -440,7 +435,7 @@ pub(crate) fn commit<R>(d: &mut DocState, f: impl FnOnce(&mut Ctx) -> Result<R>)
|
||||
if let Some(t0) = t0 {
|
||||
eprintln!("commit: edit {:?}, {} changed", since(t0), changed.len());
|
||||
}
|
||||
if structural || recalc_all {
|
||||
if structural {
|
||||
d.calc.recalc_all(&mut wb);
|
||||
} else if !changed.is_empty() {
|
||||
d.calc.cells_changed(&mut wb, &changed);
|
||||
@@ -451,18 +446,8 @@ pub(crate) fn commit<R>(d: &mut DocState, f: impl FnOnce(&mut Ctx) -> Result<R>)
|
||||
}
|
||||
fit_rows.sort_unstable();
|
||||
fit_rows.dedup();
|
||||
let mut sweep: Option<(usize, MergeSweep)> = None;
|
||||
let mut left = MERGE_FIT_WORK;
|
||||
for (si, row) in fit_rows {
|
||||
if sweep.as_ref().is_none_or(|(at, _)| *at != si) {
|
||||
if let Some((_, old)) = &sweep {
|
||||
left = old.left();
|
||||
}
|
||||
sweep = Some((si, MergeSweep::new(left)));
|
||||
}
|
||||
if let Some((_, sw)) = sweep.as_mut() {
|
||||
fit_row_height(&mut wb, si, row, sw);
|
||||
}
|
||||
auto_row_height(&mut wb, si, row);
|
||||
}
|
||||
if let Some(t0) = t0 {
|
||||
eprintln!("commit: + rows {:?}", since(t0));
|
||||
@@ -736,94 +721,6 @@ fn wrapped_line_count(line: &str, width: f32, size: f32, bold: bool) -> f32 {
|
||||
/// Line breaks in a cell's text count only when the cell wraps, as in Excel: with Wrap Text off
|
||||
/// a multi-line value shows on one line and AutoFit gives a one-line row.
|
||||
pub fn auto_row_height(wb: &mut Workbook, si: usize, row: u32) {
|
||||
let mut sweep = MergeSweep::new(MERGE_FIT_WORK);
|
||||
fit_row_height(wb, si, row, &mut sweep);
|
||||
}
|
||||
|
||||
/// Most merge comparisons one auto-fit pass (a commit, or opening a workbook) makes. Past it the
|
||||
/// remaining rows keep the heights they have rather than stall on a sheet whose merges overlap
|
||||
/// the fitted rows many times over.
|
||||
pub(crate) const MERGE_FIT_WORK: usize = 50_000_000;
|
||||
|
||||
/// The merged areas of one sheet that touch a row, for rows visited in ascending order: the
|
||||
/// merges are sorted by first row once and swept, so fitting many rows never rescans every merge
|
||||
/// per row. All comparisons count against [`MERGE_FIT_WORK`].
|
||||
pub(crate) struct MergeSweep {
|
||||
/// Built on the first row that needs merge checks, so rows that return early never pay for it.
|
||||
by_start: Option<Vec<gridcraft_core::RangeRef>>,
|
||||
next: usize,
|
||||
active: Vec<gridcraft_core::RangeRef>,
|
||||
budget: usize,
|
||||
/// The first row was answered by a plain scan: a pass that fits one row never pays for a sort.
|
||||
scanned: bool,
|
||||
}
|
||||
|
||||
impl MergeSweep {
|
||||
/// A sweep for one sheet's merges with `budget` comparisons left for the pass (carry
|
||||
/// [`MergeSweep::left`] of the previous sheet's sweep into the next). Every call to
|
||||
/// [`MergeSweep::at`] must pass that same sheet.
|
||||
pub(crate) fn new(budget: usize) -> Self {
|
||||
Self { by_start: None, next: 0, active: Vec::new(), budget, scanned: false }
|
||||
}
|
||||
|
||||
/// Comparisons left for the rest of the pass.
|
||||
pub(crate) fn left(&self) -> usize {
|
||||
self.budget
|
||||
}
|
||||
|
||||
/// Takes `n` comparisons from the budget; `false` once it is spent.
|
||||
fn charge(&mut self, n: usize) -> bool {
|
||||
match self.budget.checked_sub(n) {
|
||||
Some(rest) => {
|
||||
self.budget = rest;
|
||||
true
|
||||
}
|
||||
None => {
|
||||
self.budget = 0;
|
||||
false
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// The merges covering `row`, or `None` when the budget is spent. Rows must not decrease
|
||||
/// between calls (skipping rows is fine).
|
||||
fn at(&mut self, sh: &gridcraft_model::Sheet, row: u32) -> Option<Vec<gridcraft_core::RangeRef>> {
|
||||
if self.by_start.is_none() && !self.scanned {
|
||||
self.scanned = true;
|
||||
if !self.charge(sh.merges.len()) {
|
||||
return None;
|
||||
}
|
||||
return Some(sh.merges.iter().copied().filter(|m| !m.is_single() && m.start.row <= row && row <= m.end.row).collect());
|
||||
}
|
||||
if self.by_start.is_none() {
|
||||
// Sorting is about n log2 n comparisons, charged before it starts.
|
||||
let n = sh.merges.len();
|
||||
if !self.charge(n.saturating_mul((usize::BITS - n.leading_zeros()) as usize)) {
|
||||
return None;
|
||||
}
|
||||
let mut v: Vec<gridcraft_core::RangeRef> = sh.merges.iter().copied().filter(|m| !m.is_single()).collect();
|
||||
v.sort_unstable_by_key(|m| m.start.row);
|
||||
self.by_start = Some(v);
|
||||
}
|
||||
let by_start = self.by_start.as_ref()?;
|
||||
while let Some(m) = by_start.get(self.next).filter(|m| m.start.row <= row) {
|
||||
self.active.push(*m);
|
||||
self.next += 1;
|
||||
}
|
||||
if !self.charge(self.active.len()) {
|
||||
return None;
|
||||
}
|
||||
self.active.retain(|m| m.end.row >= row);
|
||||
Some(self.active.clone())
|
||||
}
|
||||
}
|
||||
|
||||
/// [`auto_row_height`] with the sheet's merges taken from `sweep`.
|
||||
pub(crate) fn fit_row_height(wb: &mut Workbook, si: usize, row: u32, sweep: &mut MergeSweep) {
|
||||
// Out of budget: the row keeps the height it has.
|
||||
if sweep.left() == 0 {
|
||||
return;
|
||||
}
|
||||
let Some(sh) = wb.sheet(si) else { return };
|
||||
if sh.rows.get(&row).is_some_and(|i| i.custom) {
|
||||
return;
|
||||
@@ -853,14 +750,9 @@ pub(crate) fn fit_row_height(wb: &mut Workbook, si: usize, row: u32, sweep: &mut
|
||||
// picked out once, and per cell the candidates are re-checked (normally none or a couple;
|
||||
// a sheet with many merges and many saved columns would otherwise scan all of them per
|
||||
// cell, rows × merges × columns).
|
||||
// Rows are fitted in ascending order, so the merges touching this one come from a sweep over
|
||||
// the sheet's merges sorted once, not from a scan of all of them. Out of budget: the row keeps
|
||||
// the height it has.
|
||||
let Some(row_merges) = sweep.at(sh, row) else { return };
|
||||
let row_merges: Vec<gridcraft_core::RangeRef> =
|
||||
sh.merges.iter().copied().filter(|m| !m.is_single() && m.start.row <= row && row <= m.end.row).collect();
|
||||
for (c, cell) in sh.cells.row(row, 0, gridcraft_core::MAX_COLS - 1) {
|
||||
if !sweep.charge(row_merges.len()) {
|
||||
return;
|
||||
}
|
||||
// Merged cells never auto-fit their rows, like Excel (column autofit skips them too).
|
||||
if row_merges.iter().any(|m| m.contains(CellRef::new(row, c))) {
|
||||
continue;
|
||||
@@ -904,40 +796,3 @@ impl Ctx<'_> {
|
||||
self.changed.push((sheet, c));
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod merge_sweep_tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn the_budget_carries_from_one_sheets_sweep_to_the_next() {
|
||||
let mut sheet = gridcraft_model::Sheet::new("S");
|
||||
for r in ["A1:B2", "C1:D2", "E1:F2"] {
|
||||
sheet.merges.push(gridcraft_core::RangeRef::parse(r).unwrap());
|
||||
}
|
||||
let mut first = MergeSweep::new(5);
|
||||
assert!(first.at(&sheet, 0).is_some(), "the first row is a plain scan: three comparisons");
|
||||
assert!(first.at(&sheet, 1).is_none(), "the second row needs the sort: six comparisons, two are left");
|
||||
assert_eq!(first.left(), 0);
|
||||
// The next sheet starts with what is left: nothing, so it fits nothing.
|
||||
let mut second = MergeSweep::new(first.left());
|
||||
assert!(second.at(&sheet, 0).is_none());
|
||||
assert!(MergeSweep::new(MERGE_FIT_WORK).at(&sheet, 0).is_some());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn sorting_the_merges_is_charged() {
|
||||
let mut sheet = gridcraft_model::Sheet::new("S");
|
||||
for r in ["A1:B2", "C1:D2", "E1:F2", "G1:H2"] {
|
||||
sheet.merges.push(gridcraft_core::RangeRef::parse(r).unwrap());
|
||||
}
|
||||
// The first row costs a plain scan (4); from the second on the merges are sorted, about
|
||||
// 4 * 3 comparisons: 7 left is not enough, 12 plus the row's 4 is.
|
||||
let mut sweep = MergeSweep::new(11);
|
||||
assert!(sweep.at(&sheet, 0).is_some());
|
||||
assert!(sweep.at(&sheet, 1).is_none());
|
||||
assert_eq!(sweep.left(), 0);
|
||||
let mut sweep = MergeSweep::new(20);
|
||||
assert!(sweep.at(&sheet, 0).is_some() && sweep.at(&sheet, 1).is_some());
|
||||
}
|
||||
}
|
||||
|
||||
@@ -152,30 +152,24 @@ fn input_range(s: &Session, p: &Json, keys: &[&str]) -> Result<(usize, RangeRef)
|
||||
/// Splits a range into variables (columns, or rows when `by_rows`); with `labels` the first
|
||||
/// cell of each names it, else `default(i)` does.
|
||||
fn variables(s: &Session, si: usize, r: RangeRef, by_rows: bool, labels: bool, default: &dyn Fn(usize) -> String) -> Result<Vec<Var>> {
|
||||
let outer = if by_rows { r.height() } else { r.width() };
|
||||
let mut out = Vec::with_capacity(outer as usize);
|
||||
for i in 0..outer {
|
||||
out.push(variable(s, si, r, by_rows, labels, default, i)?);
|
||||
}
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
/// Variable `i` of [`variables`], read alone so a caller that may stop early needn't build them all.
|
||||
fn variable(s: &Session, si: usize, r: RangeRef, by_rows: bool, labels: bool, default: &dyn Fn(usize) -> String, i: u32) -> Result<Var> {
|
||||
let d = s.doc()?;
|
||||
let sh = d.wb.sheet(si).ok_or(EngineError::NoDocument)?;
|
||||
let inner = if by_rows { r.width() } else { r.height() };
|
||||
let cell = |j: u32| if by_rows { CellRef::new(r.start.row + i, r.start.col + j) } else { CellRef::new(r.start.row + j, r.start.col + i) };
|
||||
let mut vals: Vec<Value> = (0..inner).map(|j| sh.value(cell(j))).collect();
|
||||
let mut name = default(i as usize);
|
||||
if labels && !vals.is_empty() {
|
||||
let first = vals.remove(0);
|
||||
let t = first.display();
|
||||
if !t.trim().is_empty() {
|
||||
name = t;
|
||||
let (outer, inner) = if by_rows { (r.height(), r.width()) } else { (r.width(), r.height()) };
|
||||
let mut out = Vec::with_capacity(outer as usize);
|
||||
for i in 0..outer {
|
||||
let cell = |j: u32| if by_rows { CellRef::new(r.start.row + i, r.start.col + j) } else { CellRef::new(r.start.row + j, r.start.col + i) };
|
||||
let mut vals: Vec<Value> = (0..inner).map(|j| sh.value(cell(j))).collect();
|
||||
let mut name = default(i as usize);
|
||||
if labels && !vals.is_empty() {
|
||||
let first = vals.remove(0);
|
||||
let t = first.display();
|
||||
if !t.trim().is_empty() {
|
||||
name = t;
|
||||
}
|
||||
}
|
||||
out.push(Var { name, vals });
|
||||
}
|
||||
Ok(Var { name, vals })
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
/// The numbers of a variable; blanks are skipped when `skip_blank`, anything else that isn't a
|
||||
@@ -278,25 +272,11 @@ struct Grid {
|
||||
cells: BTreeMap<(u32, u32), Value>,
|
||||
/// Cells shown as percentages.
|
||||
pct: Vec<(u32, u32)>,
|
||||
/// Rows and columns spanned by `cells`.
|
||||
rows: u32,
|
||||
cols: u32,
|
||||
/// A cell was dropped because the report outgrew [`MAX_OUTPUT_CELLS`]: the report is refused.
|
||||
too_large: bool,
|
||||
}
|
||||
|
||||
impl Grid {
|
||||
/// Stores a value. Past [`MAX_OUTPUT_CELLS`] (the most a report writes) nothing more is kept,
|
||||
/// so a builder that would produce millions of entries stops allocating them, and
|
||||
/// [`write_report`] refuses the report.
|
||||
fn put(&mut self, r: u32, c: u32, v: Value) {
|
||||
if !v.is_empty() {
|
||||
let (rows, cols) = (self.rows.max(r.saturating_add(1)), self.cols.max(c.saturating_add(1)));
|
||||
if self.too_large || u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
|
||||
self.too_large = true;
|
||||
return;
|
||||
}
|
||||
(self.rows, self.cols) = (rows, cols);
|
||||
self.cells.insert((r, c), v);
|
||||
}
|
||||
}
|
||||
@@ -316,7 +296,9 @@ impl Grid {
|
||||
}
|
||||
/// (rows, cols) spanned.
|
||||
fn extent(&self) -> (u32, u32) {
|
||||
(self.rows, self.cols)
|
||||
let rows = self.cells.keys().map(|(r, _)| r + 1).max().unwrap_or(0);
|
||||
let cols = self.cells.keys().map(|(_, c)| c + 1).max().unwrap_or(0);
|
||||
(rows, cols)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -328,7 +310,7 @@ fn idx(i: usize) -> u32 {
|
||||
|
||||
fn descriptive(s: &Session, p: &Json) -> Result<Grid> {
|
||||
let (si, r) = input_range(s, p, &["range", "input"])?;
|
||||
let (by_rows, labels) = (by_rows(p), flag(p, "labels"));
|
||||
let vars = variables(s, si, r, by_rows(p), flag(p, "labels"), &|i| format!("Column{}", i + 1))?;
|
||||
let summary = p.get("summary").map(|_| flag(p, "summary")).unwrap_or(true);
|
||||
let conf = confidence(p, "confidence")?;
|
||||
let kl = num(p, "kthLargest").filter(|k| *k >= 1.0).map(|k| k as usize);
|
||||
@@ -337,14 +319,7 @@ fn descriptive(s: &Session, p: &Json) -> Result<Grid> {
|
||||
return Err(bad(CMD, "choose at least one of summary, confidence, kthLargest or kthSmallest"));
|
||||
}
|
||||
let mut g = Grid::default();
|
||||
// Variables are read one at a time: a million-row range grouped by rows is a million variables.
|
||||
for i in 0..(if by_rows { r.height() } else { r.width() }) {
|
||||
// The report is refused once the grid is too large: the rest would only build discarded labels.
|
||||
if g.too_large {
|
||||
break;
|
||||
}
|
||||
let v = variable(s, si, r, by_rows, labels, &|i| format!("Column{}", i + 1), i)?;
|
||||
let i = i as usize;
|
||||
for (i, v) in vars.iter().enumerate() {
|
||||
let x = numbers(&v.vals, true)?;
|
||||
let c = idx(i).saturating_mul(2);
|
||||
g.txt(0, c, &v.name);
|
||||
@@ -1774,7 +1749,7 @@ fn sampling(s: &Session, p: &Json, seed: Option<u64>) -> Result<Grid> {
|
||||
fn write_report(s: &mut Session, p: &Json, tool: &str, g: Grid) -> Result<Json> {
|
||||
let (rows, cols) = g.extent();
|
||||
let (rows, cols) = (rows.max(1), cols.max(1));
|
||||
if g.too_large || u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
|
||||
if u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
|
||||
return Err(bad(CMD, "the report is too large"));
|
||||
}
|
||||
let d = s.doc()?;
|
||||
@@ -1939,28 +1914,6 @@ mod tests {
|
||||
x
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_report_stops_growing_past_the_output_limit() {
|
||||
// Entries past the limit are not kept, and the report is refused.
|
||||
let mut g = super::Grid::default();
|
||||
g.num(1999, 1999, 1.0);
|
||||
assert!(!g.too_large);
|
||||
g.num(0, 2000, 1.0);
|
||||
assert!(g.too_large);
|
||||
assert_eq!(g.cells.len(), 1);
|
||||
assert_eq!(g.extent(), (2000, 2000));
|
||||
// Row-grouped statistics of a tall, almost empty column: one variable per row would make
|
||||
// millions of report entries; the report is refused instead of built.
|
||||
let mut s = s();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[1]]})).unwrap();
|
||||
s.execute("range.setValues", json!({"range": "A1000000", "values": [[2]]})).unwrap();
|
||||
let err = s
|
||||
.execute("data.analysisToolPak", json!({"tool": "descriptive", "range": "A1:A1000000", "groupedBy": "rows", "output": "C1"}))
|
||||
.expect_err("report too large");
|
||||
assert!(err.to_string().contains("too large"), "{err}");
|
||||
assert_eq!(v(&s, "C1"), Value::Empty);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn forecast_standard_errors_match_the_full_window_rescan() {
|
||||
// #1043: the running sum must give what re-adding every window gave, gaps included.
|
||||
|
||||
@@ -98,7 +98,7 @@ pub fn specs() -> Vec<CommandSpec> {
|
||||
validation
|
||||
),
|
||||
cmd!(query "data.validate", "Check Value Against Validation", [], None, "{cell?, input} → {ok, message?}", has_doc, validate_cmd),
|
||||
cmd!(noundo "data.circleInvalid", "Circle Invalid Data", ["Data", "Data Tools", "Data Validation"], None, "{} → {invalid: [cell], truncated: bool (true when the 100,000-cell scan allowance ended before every range was checked)}", has_doc, circle_invalid),
|
||||
cmd!(noundo "data.circleInvalid", "Circle Invalid Data", ["Data", "Data Tools", "Data Validation"], None, "{} → invalid cells", has_doc, circle_invalid),
|
||||
CommandSpec {
|
||||
shortcut_mac: Some("Cmd+Shift+K"),
|
||||
..cmd!(
|
||||
|
||||
@@ -184,8 +184,7 @@ pub(crate) fn grow_autofilter(sh: &mut Sheet, at: CellRef) {
|
||||
pub(crate) fn apply_filter(wb: &mut Workbook, sheet: usize) {
|
||||
let Some(sh) = wb.sheet(sheet) else { return };
|
||||
let Some(af) = sh.autofilter.clone() else { return };
|
||||
// An imported range can name rows past the grid; none of those exist to hide.
|
||||
let r = RangeRef::new(af.range.start, CellRef::new(af.range.end.row.min(gridcraft_core::MAX_ROWS - 1), af.range.end.col));
|
||||
let r = af.range;
|
||||
let mut hide = Vec::new();
|
||||
// Precompute per-criterion top10 threshold and average once, not per row.
|
||||
let mut col_stats: std::collections::HashMap<usize, (Option<f64>, f64)> = std::collections::HashMap::new();
|
||||
@@ -318,9 +317,5 @@ pub fn filter_values(wb: &Workbook, sheet: usize, col: u32) -> Vec<(String, bool
|
||||
let key = if blank { sort_key(&Value::Empty) } else { sort_key(&v) };
|
||||
seen.entry((key, t)).or_insert(on);
|
||||
}
|
||||
let mut values: Vec<_> = seen.into_iter().collect();
|
||||
values.sort_by(|((a, ta), _), ((b, tb), _)| {
|
||||
a.0.cmp(&b.0).then(a.1.cmp(&b.1)).then_with(|| gridcraft_core::compare_text(&a.2, &b.2)).then_with(|| ta.cmp(tb))
|
||||
});
|
||||
values.into_iter().map(|((_, t), on)| (t, on)).collect()
|
||||
seen.into_iter().map(|((_, t), on)| (t, on)).collect()
|
||||
}
|
||||
|
||||
@@ -53,17 +53,6 @@ fn subtotal_puts_summary_rows_below_their_detail() {
|
||||
assert_eq!(sh.value(CellRef::new(3, 0)).display(), "West");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn filter_list_orders_arabic_like_sort() {
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["H"], ["بج"], ["بُت"], ["\u{200F}بب"]]})).unwrap();
|
||||
s.execute("data.filter", json!({"range": "A1:A4"})).unwrap();
|
||||
let doc = s.doc().unwrap();
|
||||
let list: Vec<String> = filter_values(&doc.wb, 0, 0).into_iter().map(|(t, _)| t).collect();
|
||||
assert_eq!(list, ["\u{200F}بب", "بُت", "بج"]);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn filter_by_huge_numeric_column_is_rejected_not_overflowed() {
|
||||
let mut s = Session::new();
|
||||
|
||||
@@ -5,9 +5,6 @@ use super::{
|
||||
|
||||
// ---------------------------------------------------------------- tools
|
||||
|
||||
/// Most cells (rows times compared columns) one Remove Duplicates call reads.
|
||||
const MAX_COMPARED_CELLS: u64 = 50_000_000;
|
||||
|
||||
pub(super) fn remove_duplicates(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let (sheet, r) = data_target(s, p, "data.removeDuplicates")?;
|
||||
let d = s.doc()?;
|
||||
@@ -27,10 +24,6 @@ pub(super) fn remove_duplicates(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
if cols.is_empty() || cols.iter().any(|c| !(r.start.col..=r.end.col).contains(c)) {
|
||||
return Err(bad("data.removeDuplicates", "`columns` must name at least one column of the range"));
|
||||
}
|
||||
// Rows are compared one by one, so a sparse sheet's enclosing rectangle is bounded like a list.
|
||||
if r.count() > MAX_COMPARED_CELLS {
|
||||
return Err(bad("data.removeDuplicates", "the range is too large"));
|
||||
}
|
||||
let start = r.start.row + header as u32;
|
||||
let mut removed = 0;
|
||||
edit(s, |cx| {
|
||||
|
||||
@@ -337,17 +337,11 @@ pub(super) fn circle_invalid(s: &mut Session, _: &Json) -> Result<Json> {
|
||||
let sheet = d.wb.active_sheet();
|
||||
let Some(sh) = d.wb.sheet(sheet) else { return Err(EngineError::NoDocument) };
|
||||
let mut bad_cells = Vec::new();
|
||||
// One scan allowance for the whole command: duplicate or overlapping ranges don't multiply it.
|
||||
let mut budget = 100_000usize;
|
||||
// Cells were left unchecked because the allowance ran out: said in the result.
|
||||
let mut truncated = false;
|
||||
for dv in &sh.validations {
|
||||
for r in &dv.ranges {
|
||||
let r = sh.used_range().and_then(|u| r.intersection(&u));
|
||||
let Some(r) = r else { continue };
|
||||
truncated |= r.count() > budget as u64;
|
||||
for c in r.iter().take(budget) {
|
||||
budget -= 1;
|
||||
for c in r.iter().take(100_000) {
|
||||
let text = sh.input_text(c);
|
||||
if !text.is_empty() && check_validation(&d.wb, sheet, c, &text).is_some() {
|
||||
bad_cells.push(c.a1());
|
||||
@@ -355,5 +349,5 @@ pub(super) fn circle_invalid(s: &mut Session, _: &Json) -> Result<Json> {
|
||||
}
|
||||
}
|
||||
}
|
||||
Ok(json!({"invalid": bad_cells, "truncated": truncated}))
|
||||
Ok(json!({"invalid": bad_cells}))
|
||||
}
|
||||
|
||||
@@ -14,13 +14,8 @@ fn matches(hay: &str, what: &str, case: bool, whole: bool) -> bool {
|
||||
}
|
||||
|
||||
/// `*`/`?` wildcard match with `~` escapes.
|
||||
///
|
||||
/// The pattern runs as a bit-parallel automaton over the text: one pass, each character costing
|
||||
/// one step per 64 pattern positions. Backtracking to the last `*` after a mismatch rescans the
|
||||
/// text and costs text × pattern comparisons on a long, repetitive cell, for every cell a
|
||||
/// filter or Find visits.
|
||||
pub(crate) fn wildcard(text: &str, pat: &str) -> bool {
|
||||
// The pattern as (character, escaped) items, runs of `*` folded into one.
|
||||
let t: Vec<char> = text.chars().collect();
|
||||
let mut p: Vec<(char, bool)> = Vec::new();
|
||||
let mut it = pat.chars();
|
||||
while let Some(c) = it.next() {
|
||||
@@ -28,85 +23,32 @@ pub(crate) fn wildcard(text: &str, pat: &str) -> bool {
|
||||
if let Some(n) = it.next() {
|
||||
p.push((n, true));
|
||||
}
|
||||
} else if c != '*' || p.last() != Some(&('*', false)) {
|
||||
} else {
|
||||
p.push((c, false));
|
||||
}
|
||||
}
|
||||
let words = (p.len() + 1).div_ceil(64);
|
||||
let bit = |set: &mut [u64], i: usize| set[i / 64] |= 1 << (i % 64);
|
||||
// Positions holding a `*` and a `?`, and where each literal character sits in the pattern.
|
||||
let mut stars = vec![0u64; words];
|
||||
let mut any = vec![0u64; words];
|
||||
let mut at: std::collections::HashMap<char, Vec<usize>> = std::collections::HashMap::new();
|
||||
for (i, item) in p.iter().enumerate() {
|
||||
match item {
|
||||
('*', false) => bit(&mut stars, i),
|
||||
('?', false) => bit(&mut any, i),
|
||||
(c, _) => at.entry(*c).or_default().push(i),
|
||||
}
|
||||
}
|
||||
// A literal's bit mask is built when the text first reads it, and kept for at most
|
||||
// `MAX_CACHED_MASKS` literals: a long pattern of mostly distinct characters would otherwise
|
||||
// allocate a mask of the pattern's width for each one up front. When there are more literals
|
||||
// than that, the ones kept are those at the most pattern positions (the dearest to rebuild);
|
||||
// each of the rest has at most pattern length / `MAX_CACHED_MASKS` positions, so rebuilding
|
||||
// its mask in `scratch` costs less than one step of the scan.
|
||||
const MAX_CACHED_MASKS: usize = 256;
|
||||
let cacheable: Option<std::collections::HashSet<char>> = (at.len() > MAX_CACHED_MASKS).then(|| {
|
||||
let mut by_count: Vec<(&char, usize)> = at.iter().map(|(c, v)| (c, v.len())).collect();
|
||||
by_count.sort_unstable_by_key(|&(_, n)| std::cmp::Reverse(n));
|
||||
by_count.into_iter().take(MAX_CACHED_MASKS).map(|(c, _)| *c).collect()
|
||||
});
|
||||
let mut masks: std::collections::HashMap<char, Vec<u64>> = std::collections::HashMap::new();
|
||||
let mut scratch = vec![0u64; words];
|
||||
// A set bit i: the text read so far matches the first i pattern items. A `*` matches nothing
|
||||
// as well, so a state at a `*` is also at the item after it.
|
||||
let close = |set: &mut Vec<u64>| {
|
||||
let mut carry = 0u64;
|
||||
for w in 0..words {
|
||||
let s = set[w] & stars[w];
|
||||
let next = (s << 1) | carry;
|
||||
carry = s >> 63;
|
||||
set[w] |= next;
|
||||
}
|
||||
};
|
||||
let mut state = vec![0u64; words];
|
||||
bit(&mut state, 0);
|
||||
close(&mut state);
|
||||
let mut next = vec![0u64; words];
|
||||
for ch in text.chars() {
|
||||
let lit: Option<&[u64]> = match at.get(&ch) {
|
||||
None => None,
|
||||
Some(positions) if cacheable.as_ref().is_none_or(|k| k.contains(&ch)) => {
|
||||
let mask = masks.entry(ch).or_insert_with(|| {
|
||||
let mut m = vec![0u64; words];
|
||||
positions.iter().for_each(|&i| bit(&mut m, i));
|
||||
m
|
||||
});
|
||||
Some(mask)
|
||||
}
|
||||
Some(positions) => {
|
||||
scratch.fill(0);
|
||||
positions.iter().for_each(|&i| bit(&mut scratch, i));
|
||||
Some(&scratch)
|
||||
}
|
||||
};
|
||||
let mut carry = 0u64;
|
||||
let mut alive = 0u64;
|
||||
for w in 0..words {
|
||||
// Items that read `ch` advance; a `*` reads it and stays.
|
||||
let m = state[w] & (any[w] | lit.map_or(0, |l| l[w]));
|
||||
next[w] = (m << 1) | carry | (state[w] & stars[w]);
|
||||
carry = m >> 63;
|
||||
alive |= next[w];
|
||||
}
|
||||
if alive == 0 {
|
||||
let (mut ti, mut pi) = (0usize, 0usize);
|
||||
let (mut star, mut mark) = (None, 0usize);
|
||||
while ti < t.len() {
|
||||
if pi < p.len() && (p[pi] == ('?', false) || (p[pi].0 == t[ti] && p[pi] != ('*', false))) {
|
||||
ti += 1;
|
||||
pi += 1;
|
||||
} else if pi < p.len() && p[pi] == ('*', false) {
|
||||
star = Some(pi);
|
||||
mark = ti;
|
||||
pi += 1;
|
||||
} else if let Some(sp) = star {
|
||||
pi = sp + 1;
|
||||
mark += 1;
|
||||
ti = mark;
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
std::mem::swap(&mut state, &mut next);
|
||||
close(&mut state);
|
||||
}
|
||||
state[p.len() / 64] & (1 << (p.len() % 64)) != 0
|
||||
while pi < p.len() && p[pi] == ('*', false) {
|
||||
pi += 1;
|
||||
}
|
||||
pi == p.len()
|
||||
}
|
||||
|
||||
pub(super) fn find(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
@@ -170,19 +112,9 @@ pub(super) fn find(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
Ok(json!({"cell": c.a1(), "sheet": si, "value": t, "count": hits.len()}))
|
||||
}
|
||||
|
||||
/// Most characters a cell holds (Excel's limit).
|
||||
const REPLACE_CELL_MAX_CHARS: usize = 32_767;
|
||||
/// Longest text one replacement may give a cell, in bytes: the character limit, each up to four
|
||||
/// bytes. Every occurrence copies the replacement text, so a short cell full of matches and a
|
||||
/// long replacement would otherwise expand to gigabytes. The character limit is checked on what
|
||||
/// fits in this.
|
||||
const REPLACE_CELL_MAX_BYTES: usize = REPLACE_CELL_MAX_CHARS * 4;
|
||||
/// Most bytes of replaced text one Replace All builds across all its cells.
|
||||
const REPLACE_TOTAL_MAX_BYTES: usize = 256 * 1024 * 1024;
|
||||
|
||||
pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let what = str_param(p, "what").unwrap_or("").to_string();
|
||||
let with = str_param(p, "with").unwrap_or("");
|
||||
let with = str_param(p, "with").unwrap_or("").to_string();
|
||||
if what.is_empty() {
|
||||
return Err(bad("edit.replace", "missing `what`"));
|
||||
}
|
||||
@@ -196,14 +128,9 @@ pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let n = edit(s, |cx| {
|
||||
let sheets: Vec<usize> = if workbook { (0..cx.wb.sheets.len()).collect() } else { vec![cx.wb.active_sheet()] };
|
||||
let mut n = 0;
|
||||
let mut total = 0usize;
|
||||
for si in sheets {
|
||||
let cells: Vec<(CellRef, String)> =
|
||||
// A hidden formula's source is not searched: a count or error would reveal its text.
|
||||
cx.wb
|
||||
.sheet(si)
|
||||
.map(|sh| sh.cells.iter().filter(|(c, cell)| cell.formula.is_none() && cell.py.is_none() || !crate::display::formula_hidden(&cx.wb, sh, *c)).map(|(c, _)| (c, sh.input_text(c))).collect())
|
||||
.unwrap_or_default();
|
||||
cx.wb.sheet(si).map(|sh| sh.cells.iter().map(|(c, _)| (c, sh.input_text(c))).collect()).unwrap_or_default();
|
||||
for (c, text) in cells {
|
||||
if (!all && c != active) || selection.as_ref().is_some_and(|sel| !sel.contains(c)) {
|
||||
continue;
|
||||
@@ -212,20 +139,12 @@ pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
continue;
|
||||
}
|
||||
let new = if whole {
|
||||
(with.len() <= REPLACE_CELL_MAX_BYTES).then(|| with.to_string())
|
||||
with.clone()
|
||||
} else if case {
|
||||
replace_cs_within(&text, &what, with, REPLACE_CELL_MAX_BYTES)
|
||||
text.replace(&what, &with)
|
||||
} else {
|
||||
replace_ci_within(&text, &what, with, REPLACE_CELL_MAX_BYTES)
|
||||
replace_ci(&text, &what, &with)
|
||||
};
|
||||
let new = new.filter(|n| n.chars().count() <= REPLACE_CELL_MAX_CHARS);
|
||||
let Some(new) = new else {
|
||||
return Err(EngineError::Other(format!("The replacement makes the text of {} too long.", c.a1())));
|
||||
};
|
||||
total = total.saturating_add(new.len());
|
||||
if total > REPLACE_TOTAL_MAX_BYTES {
|
||||
return Err(EngineError::Other("The replacements make too much text. Replace in a smaller range.".into()));
|
||||
}
|
||||
// A Python cell's text is its code: the replacement edits the code and the cell
|
||||
// stays a Python cell. Code that would no longer fit in a PY formula is left as it was.
|
||||
let old_py = cx.wb.sheet(si).and_then(|sh| sh.cell(c)).and_then(|x| x.py.clone());
|
||||
@@ -280,37 +199,7 @@ fn folds_in_place(s: &str) -> bool {
|
||||
})
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
pub(super) fn replace_ci(text: &str, what: &str, with: &str) -> String {
|
||||
replace_ci_within(text, what, with, usize::MAX).unwrap_or_default()
|
||||
}
|
||||
|
||||
/// `text` with every `what` replaced by `with`, or `None` once the result passes `limit` bytes
|
||||
/// (checked before each copy, so nothing larger is allocated).
|
||||
pub(super) fn replace_cs_within(text: &str, what: &str, with: &str, limit: usize) -> Option<String> {
|
||||
if what.is_empty() {
|
||||
return (text.len() <= limit).then(|| text.to_string());
|
||||
}
|
||||
let mut out = String::new();
|
||||
let mut i = 0;
|
||||
for (pos, m) in text.match_indices(what) {
|
||||
if out.len() + (pos - i) + with.len() > limit {
|
||||
return None;
|
||||
}
|
||||
out.push_str(text.get(i..pos).unwrap_or(""));
|
||||
out.push_str(with);
|
||||
i = pos + m.len();
|
||||
}
|
||||
let rest = text.get(i..).unwrap_or("");
|
||||
if out.len() + rest.len() > limit {
|
||||
return None;
|
||||
}
|
||||
out.push_str(rest);
|
||||
Some(out)
|
||||
}
|
||||
|
||||
/// Case-insensitive [`replace_cs_within`].
|
||||
pub(super) fn replace_ci_within(text: &str, what: &str, with: &str, limit: usize) -> Option<String> {
|
||||
let lower = text.to_lowercase();
|
||||
let w = what.to_lowercase();
|
||||
// Slicing `text` at positions found in `lower` is only sound when folding kept every byte
|
||||
@@ -318,22 +207,15 @@ pub(super) fn replace_ci_within(text: &str, what: &str, with: &str, limit: usize
|
||||
// boundaries (U+212A KELVIN SIGN shrinks 3→1 bytes, U+0130 grows 2→3), where the tail used
|
||||
// to be dropped. Those fall back to the case-sensitive replace, as length changes already do.
|
||||
if lower.len() != text.len() || w.is_empty() || !folds_in_place(text) || !folds_in_place(what) {
|
||||
return replace_cs_within(text, what, with, limit);
|
||||
return text.replace(what, with);
|
||||
}
|
||||
let mut out = String::new();
|
||||
let mut i = 0;
|
||||
while let Some(pos) = lower.get(i..).and_then(|s| s.find(&w)) {
|
||||
if out.len() + pos + with.len() > limit {
|
||||
return None;
|
||||
}
|
||||
out.push_str(text.get(i..i + pos).unwrap_or(""));
|
||||
out.push_str(with);
|
||||
i += pos + w.len();
|
||||
}
|
||||
let rest = text.get(i..).unwrap_or("");
|
||||
if out.len() + rest.len() > limit {
|
||||
return None;
|
||||
}
|
||||
out.push_str(rest);
|
||||
Some(out)
|
||||
out.push_str(text.get(i..).unwrap_or(""));
|
||||
out
|
||||
}
|
||||
|
||||
@@ -17,13 +17,11 @@ pub(in super::super) fn go_to(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
let Some(reference) = str_param(p, "reference").or_else(|| str_param(p, "range")) else { return Err(bad("edit.goTo", "missing `reference`")) };
|
||||
let d = s.doc()?;
|
||||
let mut names = Resolver::new(&d.wb);
|
||||
let mut target: Option<usize> = None;
|
||||
let mut areas: Vec<RangeRef> = Vec::new();
|
||||
let mut named = false;
|
||||
for part in reference_parts(reference) {
|
||||
let mut found = Vec::new();
|
||||
let Some((sheet, _, by_name)) = names.target(d.wb.active_sheet(), part, 0, Some(&mut found)) else {
|
||||
let Some((sheet, found, by_name)) = go_to_target(&d.wb, d.wb.active_sheet(), part, 0) else {
|
||||
return Err(EngineError::Other("Reference isn't valid.".into()));
|
||||
};
|
||||
// One selection holds the areas of one sheet.
|
||||
@@ -79,114 +77,57 @@ pub(super) fn reference_parts(text: &str) -> impl Iterator<Item = &str> {
|
||||
})
|
||||
}
|
||||
|
||||
/// Resolves Go To references against one workbook. Defined names are found through an index
|
||||
/// built once, and whether a name refers to a reference is remembered: listing the names of a
|
||||
/// workbook full of names that alias other names would otherwise scan every name for every alias
|
||||
/// of every name.
|
||||
struct Resolver<'a> {
|
||||
wb: &'a gridcraft_model::Workbook,
|
||||
/// Indexes into `wb.names` by scope and folded name, in workbook order.
|
||||
by_name: std::collections::HashMap<(Option<usize>, String), Vec<usize>>,
|
||||
/// [`Resolver::name_areas`] without areas, by name index, sheet and depth: the target sheet
|
||||
/// and the number of areas.
|
||||
memo: std::collections::HashMap<(usize, usize, usize), Option<(usize, usize)>>,
|
||||
/// Where one part of a Go To reference points, seen from sheet `sheet`: the sheet, its areas,
|
||||
/// and whether a defined name or a table gave them. `Sheet2!Name` is the name scoped to Sheet2.
|
||||
/// A name counts when what it refers to is a reference; a constant or a formula is not one.
|
||||
fn go_to_target(wb: &gridcraft_model::Workbook, sheet: usize, part: &str, depth: usize) -> Option<(usize, Vec<RangeRef>, bool)> {
|
||||
if part.is_empty() || depth > GO_TO_NAME_DEPTH {
|
||||
return None;
|
||||
}
|
||||
let (prefix, body) = split_sheet(part);
|
||||
let on = match &prefix {
|
||||
Some(n) => wb.sheet_index(n)?,
|
||||
None => sheet,
|
||||
};
|
||||
if let Some(r) = RangeRef::parse(&body.replace('$', "")) {
|
||||
return Some((on, vec![r], false));
|
||||
}
|
||||
let name = match prefix {
|
||||
Some(_) => wb.names.iter().find(|n| n.scope == Some(on) && gridcraft_formula::same_name(&n.name, body)),
|
||||
None => wb.name(body, sheet),
|
||||
};
|
||||
if let Some(n) = name {
|
||||
return name_areas(wb, n, sheet, depth).map(|(t, areas)| (t, areas, true));
|
||||
}
|
||||
if prefix.is_none()
|
||||
&& let Some((si, ti)) = wb.table(body)
|
||||
{
|
||||
// As in Excel, a table's name stands for its data, without the header and totals rows.
|
||||
let t = wb.sheet(si)?.tables.get(ti)?;
|
||||
return Some((si, vec![t.data_range().unwrap_or(t.range)], true));
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
impl<'a> Resolver<'a> {
|
||||
fn new(wb: &'a gridcraft_model::Workbook) -> Self {
|
||||
let mut by_name: std::collections::HashMap<(Option<usize>, String), Vec<usize>> = std::collections::HashMap::new();
|
||||
for (i, n) in wb.names.iter().enumerate() {
|
||||
by_name.entry((n.scope, gridcraft_formula::fold_name(&n.name))).or_default().push(i);
|
||||
}
|
||||
Self { wb, by_name, memo: std::collections::HashMap::new() }
|
||||
}
|
||||
|
||||
/// The first name of scope `scope` that `same` accepts, as a scan of `wb.names` finds it.
|
||||
/// Names `same` accepts fold alike, so only that bucket is checked.
|
||||
fn find(&self, scope: Option<usize>, body: &str, same: impl Fn(&str, &str) -> bool) -> Option<usize> {
|
||||
self.by_name
|
||||
.get(&(scope, gridcraft_formula::fold_name(body)))?
|
||||
.iter()
|
||||
.copied()
|
||||
.find(|&i| self.wb.names.get(i).is_some_and(|n| same(&n.name, body)))
|
||||
}
|
||||
|
||||
/// Where one part of a Go To reference points, seen from sheet `sheet`: the sheet, how many
|
||||
/// areas (appended to `areas` when given), and whether a defined name or a table gave them.
|
||||
/// `Sheet2!Name` is the name scoped to Sheet2. A name counts when what it refers to is a
|
||||
/// reference; a constant or a formula is not one.
|
||||
fn target(&mut self, sheet: usize, part: &str, depth: usize, areas: Option<&mut Vec<RangeRef>>) -> Option<(usize, usize, bool)> {
|
||||
let wb = self.wb;
|
||||
if part.is_empty() || depth > GO_TO_NAME_DEPTH {
|
||||
/// The sheet and areas defined name `n` refers to, seen from sheet `sheet`, when that is a
|
||||
/// reference.
|
||||
fn name_areas(wb: &gridcraft_model::Workbook, n: &gridcraft_model::DefinedName, sheet: usize, depth: usize) -> Option<(usize, Vec<RangeRef>)> {
|
||||
let from = n.scope.unwrap_or(sheet);
|
||||
let formula = n.formula.trim().trim_start_matches('=');
|
||||
let mut target: Option<usize> = None;
|
||||
let mut areas = Vec::new();
|
||||
for p in reference_parts(formula) {
|
||||
let (s, found, _) = go_to_target(wb, from, p, depth + 1)?;
|
||||
if target.is_some_and(|t| t != s) {
|
||||
return None;
|
||||
}
|
||||
let (prefix, body) = split_sheet(part);
|
||||
let on = match &prefix {
|
||||
Some(n) => wb.sheet_index(n)?,
|
||||
None => sheet,
|
||||
};
|
||||
if let Some(r) = RangeRef::parse(&body.replace('$', "")) {
|
||||
if let Some(a) = areas {
|
||||
a.push(r);
|
||||
}
|
||||
return Some((on, 1, false));
|
||||
target = Some(s);
|
||||
areas.extend(found);
|
||||
if areas.len() > GO_TO_MAX_AREAS {
|
||||
return None;
|
||||
}
|
||||
// As `Workbook::name`: the sheet's own name, else the workbook's.
|
||||
let name = match prefix {
|
||||
Some(_) => self.find(Some(on), body, gridcraft_formula::same_name),
|
||||
None => self.find(Some(sheet), body, str::eq_ignore_ascii_case).or_else(|| self.find(None, body, str::eq_ignore_ascii_case)),
|
||||
};
|
||||
if let Some(i) = name {
|
||||
return self.name_areas(i, sheet, depth, areas).map(|(t, n)| (t, n, true));
|
||||
}
|
||||
if prefix.is_none()
|
||||
&& let Some((si, ti)) = wb.table(body)
|
||||
{
|
||||
// As in Excel, a table's name stands for its data, without the header and totals rows.
|
||||
let t = wb.sheet(si)?.tables.get(ti)?;
|
||||
if let Some(a) = areas {
|
||||
a.push(t.data_range().unwrap_or(t.range));
|
||||
}
|
||||
return Some((si, 1, true));
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
/// The sheet and number of areas defined name `wb.names[i]` refers to, seen from sheet
|
||||
/// `sheet`, when that is a reference; the areas are appended to `areas` when given.
|
||||
fn name_areas(&mut self, i: usize, sheet: usize, depth: usize, mut areas: Option<&mut Vec<RangeRef>>) -> Option<(usize, usize)> {
|
||||
let key = (i, sheet, depth);
|
||||
if areas.is_none()
|
||||
&& let Some(known) = self.memo.get(&key)
|
||||
{
|
||||
return *known;
|
||||
}
|
||||
let out = self.name_areas_uncached(i, sheet, depth, areas.as_deref_mut());
|
||||
if areas.is_none() {
|
||||
self.memo.insert(key, out);
|
||||
}
|
||||
out
|
||||
}
|
||||
|
||||
fn name_areas_uncached(&mut self, i: usize, sheet: usize, depth: usize, mut areas: Option<&mut Vec<RangeRef>>) -> Option<(usize, usize)> {
|
||||
let n = self.wb.names.get(i)?;
|
||||
let from = n.scope.unwrap_or(sheet);
|
||||
let formula = n.formula.trim().trim_start_matches('=');
|
||||
let mut target: Option<usize> = None;
|
||||
let mut count = 0usize;
|
||||
for p in reference_parts(formula) {
|
||||
let (s, found, _) = self.target(from, p, depth + 1, areas.as_deref_mut())?;
|
||||
if target.is_some_and(|t| t != s) {
|
||||
return None;
|
||||
}
|
||||
target = Some(s);
|
||||
count += found;
|
||||
if count > GO_TO_MAX_AREAS {
|
||||
return None;
|
||||
}
|
||||
}
|
||||
target.map(|t| (t, count))
|
||||
}
|
||||
target.map(|t| (t, areas))
|
||||
}
|
||||
|
||||
/// The names the Go To dialog lists, seen from sheet `sheet`: the named ranges of the workbook and
|
||||
@@ -196,13 +137,11 @@ impl<'a> Resolver<'a> {
|
||||
pub fn go_to_names(wb: &gridcraft_model::Workbook, sheet: usize) -> Vec<String> {
|
||||
let own: std::collections::HashSet<String> =
|
||||
wb.names.iter().filter(|n| n.scope == Some(sheet)).map(|n| gridcraft_formula::fold_name(&n.name)).collect();
|
||||
let mut names = Resolver::new(wb);
|
||||
let mut out: Vec<String> = wb
|
||||
.names
|
||||
.iter()
|
||||
.enumerate()
|
||||
.filter(|(i, n)| !n.hidden && names.name_areas(*i, sheet, 0, None).is_some())
|
||||
.filter_map(|(_, n)| match n.scope {
|
||||
.filter(|n| !n.hidden && name_areas(wb, n, sheet, 0).is_some())
|
||||
.filter_map(|n| match n.scope {
|
||||
None if own.contains(&gridcraft_formula::fold_name(&n.name)) => None,
|
||||
None => Some(n.name.clone()),
|
||||
Some(s) if s == sheet => Some(n.name.clone()),
|
||||
|
||||
@@ -42,28 +42,6 @@ fn go_to_lists_the_named_ranges_of_every_sheet() {
|
||||
assert_eq!(list["names"], json!(["Areas", "Local", "Price", "Sheet1!Own", "type"]), "{list}");
|
||||
}
|
||||
|
||||
/// Thousands of names that alias other names are listed with one lookup per alias, not a scan of
|
||||
/// every name for each; aliases resolve through the chain like a single name does.
|
||||
#[test]
|
||||
fn go_to_lists_many_aliases_and_follows_them() {
|
||||
let mut s = go_to_workbook();
|
||||
let wb = std::sync::Arc::make_mut(&mut s.doc_mut().unwrap().wb);
|
||||
let alias = |name: String, formula: String| gridcraft_model::DefinedName { name, scope: None, formula, comment: String::new(), hidden: false };
|
||||
wb.names.push(alias("Chain0".into(), "Areas".into()));
|
||||
for i in 1..20_000 {
|
||||
wb.names.push(alias(format!("Chain{i}"), format!("Chain{},price", i - 1)));
|
||||
}
|
||||
let list = s.execute("edit.goTo", json!({"list": true})).unwrap();
|
||||
let names: Vec<&str> = list["names"].as_array().unwrap().iter().filter_map(|n| n.as_str()).collect();
|
||||
// Chain0..Chain6 are within the name depth Go To follows; longer chains are not ranges.
|
||||
for i in 0..7 {
|
||||
assert!(names.contains(&format!("Chain{i}").as_str()), "Chain{i} in {names:?}");
|
||||
}
|
||||
assert!(!names.contains(&"Chain7"), "{names:?}");
|
||||
let out = s.execute("edit.goTo", json!({"reference": "chain1"})).unwrap();
|
||||
assert_eq!(out["selection"], "A1,C3:C4,B2:B5", "{out}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn go_to_a_name_scoped_to_another_sheet_goes_to_that_sheet() {
|
||||
let mut s = go_to_workbook();
|
||||
@@ -324,103 +302,6 @@ fn wildcards() {
|
||||
assert!(wildcard("", "*"));
|
||||
}
|
||||
|
||||
/// The backtracking matcher `wildcard` replaced, as the reference for its results.
|
||||
fn wildcard_backtracking(text: &str, pat: &str) -> bool {
|
||||
let t: Vec<char> = text.chars().collect();
|
||||
let mut p: Vec<(char, bool)> = Vec::new();
|
||||
let mut it = pat.chars();
|
||||
while let Some(c) = it.next() {
|
||||
if c == '~' {
|
||||
if let Some(n) = it.next() {
|
||||
p.push((n, true));
|
||||
}
|
||||
} else {
|
||||
p.push((c, false));
|
||||
}
|
||||
}
|
||||
let (mut ti, mut pi) = (0usize, 0usize);
|
||||
let (mut star, mut mark) = (None, 0usize);
|
||||
while ti < t.len() {
|
||||
if pi < p.len() && (p[pi] == ('?', false) || (p[pi].0 == t[ti] && p[pi] != ('*', false))) {
|
||||
ti += 1;
|
||||
pi += 1;
|
||||
} else if pi < p.len() && p[pi] == ('*', false) {
|
||||
star = Some(pi);
|
||||
mark = ti;
|
||||
pi += 1;
|
||||
} else if let Some(sp) = star {
|
||||
pi = sp + 1;
|
||||
mark += 1;
|
||||
ti = mark;
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
while pi < p.len() && p[pi] == ('*', false) {
|
||||
pi += 1;
|
||||
}
|
||||
pi == p.len()
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wildcard_agrees_with_backtracking() {
|
||||
// Every text up to 5 characters over a small alphabet against every pattern up to 4 items.
|
||||
let alphabet = ['a', 'b', '*', '?', '~'];
|
||||
let all = |n: usize, chars: &[char]| -> Vec<String> {
|
||||
let mut out = vec![String::new()];
|
||||
let mut level = vec![String::new()];
|
||||
for _ in 0..n {
|
||||
level = level.iter().flat_map(|s| chars.iter().map(move |c| format!("{s}{c}"))).collect();
|
||||
out.extend(level.iter().cloned());
|
||||
}
|
||||
out
|
||||
};
|
||||
let texts = all(5, &['a', 'b', '*']);
|
||||
for pat in all(4, &alphabet) {
|
||||
for text in &texts {
|
||||
assert_eq!(wildcard(text, &pat), wildcard_backtracking(text, &pat), "{text:?} ~ {pat:?}");
|
||||
}
|
||||
}
|
||||
// Patterns longer than one 64-bit word.
|
||||
let long = format!("*{}*b?", "a".repeat(70));
|
||||
for text in [format!("x{}yb1", "a".repeat(70)), format!("{}b", "a".repeat(71)), "a".repeat(200)] {
|
||||
assert_eq!(wildcard(&text, &long), wildcard_backtracking(&text, &long), "{text}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wildcard_with_many_distinct_literals_agrees_with_backtracking() {
|
||||
// 600 distinct characters in the pattern: more than the masks kept, so the rest are rebuilt.
|
||||
let chars: Vec<char> = (0x4E00u32..0x4E00 + 600).filter_map(char::from_u32).collect();
|
||||
let word: String = chars.iter().collect();
|
||||
let pat = format!("*{word}?{}", chars[3]);
|
||||
for text in [format!("x{word}yq{}", chars[3]), word.clone(), format!("{word}{word}"), chars.iter().rev().collect::<String>()] {
|
||||
assert_eq!(wildcard(&text, &pat), wildcard_backtracking(&text, &pat), "{} chars", text.chars().count());
|
||||
}
|
||||
assert!(wildcard(&format!("{word}zz{}", chars[3]).replace("zz", "z"), &pat));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wildcard_keeps_masks_of_frequent_literals_past_the_cache_limit() {
|
||||
// 300 distinct characters come before the repeated `a`: past the cached masks, `a` (5,000 positions)
|
||||
// still has to be read from a kept mask, or it is rebuilt for each of the 20,000 `a` in the text.
|
||||
let rare: String = (0x4E00u32..0x4E00 + 300).filter_map(char::from_u32).collect();
|
||||
let pat = format!("*{rare}{}b", "a".repeat(5_000));
|
||||
let text = "a".repeat(20_000);
|
||||
assert!(!wildcard(&text, &pat));
|
||||
assert!(wildcard(&format!("{rare}{}b", "a".repeat(5_000)), &pat));
|
||||
assert!(!wildcard(&format!("{rare}{}", "a".repeat(5_000)), &pat));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wildcard_on_long_repetitive_text_is_linear() {
|
||||
// The backtracking matcher compared about text × pattern characters here (10^8).
|
||||
let text = "a".repeat(20_000);
|
||||
let pat = format!("*{}b", "a".repeat(5_000));
|
||||
assert!(!wildcard(&text, &pat));
|
||||
assert!(wildcard(&format!("{text}b"), &pat));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn tsv() {
|
||||
assert_eq!(parse_tsv("a\tb\n1\t2\n"), vec![vec!["a", "b"], vec!["1", "2"]]);
|
||||
@@ -543,50 +424,6 @@ fn drag_and_drop_on_another_sheet_stays_on_that_sheet() {
|
||||
assert_eq!(wb.sheet(two).unwrap().value(CellRef::new(0, 1)), Value::Empty);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn replace_refuses_text_that_would_grow_past_a_cell() {
|
||||
// 10,000 matches × a 1,000-character replacement would make a 10 MB cell (and a few such
|
||||
// cells gigabytes); the replacement is refused before that text is built.
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
let long = "a".repeat(10_000);
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[long.clone()], ["abc"]]})).unwrap();
|
||||
for case in [false, true] {
|
||||
let err = s.execute("edit.replace", json!({"what": "a", "with": "x".repeat(1_000), "matchCase": case})).expect_err("too long");
|
||||
assert!(err.to_string().contains("too long"), "{err}");
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A1"})).unwrap()["value"], json!(long));
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A2"})).unwrap()["value"], json!("abc"));
|
||||
}
|
||||
// A replacement that stays within a cell's limit still happens.
|
||||
s.execute("edit.replace", json!({"what": "b", "with": "x".repeat(100)})).unwrap();
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A2"})).unwrap()["value"], json!(format!("a{}c", "x".repeat(100))));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn replace_refuses_a_result_past_the_character_limit_though_it_fits_the_byte_limit() {
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["z"]]})).unwrap();
|
||||
for whole in [false, true] {
|
||||
let err = s.execute("edit.replace", json!({"what": "z", "with": "x".repeat(32_768), "wholeCell": whole})).expect_err("32,768 characters");
|
||||
assert!(err.to_string().contains("too long"), "{err}");
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A1"})).unwrap()["value"], json!("z"));
|
||||
}
|
||||
s.execute("edit.replace", json!({"what": "z", "with": "x".repeat(32_767)})).unwrap();
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A1"})).unwrap()["value"], json!("x".repeat(32_767)));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn bounded_replace_stops_at_its_limit() {
|
||||
use super::find::{replace_ci_within, replace_cs_within};
|
||||
assert_eq!(replace_cs_within("abab", "a", "xy", 6).as_deref(), Some("xybxyb"));
|
||||
assert_eq!(replace_cs_within("abab", "a", "xy", 5), None);
|
||||
assert_eq!(replace_ci_within("AbAb", "a", "xy", 6).as_deref(), Some("xybxyb"));
|
||||
assert_eq!(replace_ci_within("AbAb", "a", "xy", 5), None);
|
||||
assert_eq!(replace_cs_within("abc", "z", "", 3).as_deref(), Some("abc"));
|
||||
assert_eq!(replace_cs_within("abc", "z", "", 2), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn replace_case_insensitive() {
|
||||
assert_eq!(replace_ci("Apple apple APPLE", "apple", "pear"), "pear pear pear");
|
||||
|
||||
@@ -42,15 +42,6 @@ pub(super) fn table_resize(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
})
|
||||
}
|
||||
|
||||
/// Most cells (rows × columns, blanks included) Get Data imports. The rows are built densely
|
||||
/// before they are written, so a few sparse records or one far-away CSV cell must not expand
|
||||
/// into a rectangle of billions of slots.
|
||||
const MAX_IMPORT_CELLS: u64 = 10_000_000;
|
||||
|
||||
fn too_many_cells() -> EngineError {
|
||||
EngineError::Other(format!("The data is too large to import ({MAX_IMPORT_CELLS} cells at most)."))
|
||||
}
|
||||
|
||||
pub(super) fn from_csv(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let bytes = if let Some(t) = str_param(p, "text") {
|
||||
t.as_bytes().to_vec()
|
||||
@@ -66,27 +57,15 @@ pub(super) fn from_csv(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
if text.trim_start().starts_with('[') {
|
||||
let arr: Vec<Json> = serde_json::from_str(&text).map_err(|e| bad("data.getData", e.to_string()))?;
|
||||
let mut keys: Vec<String> = Vec::new();
|
||||
let mut seen: std::collections::HashSet<&str> = std::collections::HashSet::new();
|
||||
for o in &arr {
|
||||
if let Some(m) = o.as_object() {
|
||||
for k in m.keys() {
|
||||
if !seen.contains(k.as_str()) {
|
||||
// Stop collecting at the column limit.
|
||||
if keys.len() >= gridcraft_core::MAX_COLS as usize {
|
||||
return Err(EngineError::Other("The data has more fields than a sheet has columns.".into()));
|
||||
}
|
||||
seen.insert(k.as_str());
|
||||
if !keys.contains(k) {
|
||||
keys.push(k.clone());
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
if arr.len() as u64 + 1 > u64::from(gridcraft_core::MAX_ROWS) {
|
||||
return Err(EngineError::Other("The data has more rows than a sheet has.".into()));
|
||||
}
|
||||
if (arr.len() as u64 + 1).saturating_mul(keys.len().max(1) as u64) > MAX_IMPORT_CELLS {
|
||||
return Err(too_many_cells());
|
||||
}
|
||||
let mut out = vec![keys.iter().map(|k| json!(k)).collect::<Vec<_>>()];
|
||||
for o in &arr {
|
||||
out.push(keys.iter().map(|k| o.get(k).cloned().unwrap_or(Json::Null)).collect());
|
||||
@@ -96,17 +75,10 @@ pub(super) fn from_csv(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let delim = str_param(p, "delimiter")
|
||||
.and_then(|d| d.chars().next())
|
||||
.unwrap_or(if text.lines().next().is_some_and(|l| l.contains('\t')) { '\t' } else { ',' });
|
||||
let opts = gridcraft_xlsx::CsvOptions { delimiter: delim as u8, max_used_cells: Some(MAX_IMPORT_CELLS), ..Default::default() };
|
||||
// The cell limit is enforced while parsing, before an oversized file is stored.
|
||||
let wb = gridcraft_xlsx::read_csv(&bytes, &opts).map_err(|e| match e {
|
||||
gridcraft_xlsx::IoError::TooLarge(_) => too_many_cells(),
|
||||
e => EngineError::Other(e.to_string()),
|
||||
})?;
|
||||
let opts = gridcraft_xlsx::CsvOptions { delimiter: delim as u8, ..Default::default() };
|
||||
let wb = gridcraft_xlsx::read_csv(&bytes, &opts).map_err(|e| EngineError::Other(e.to_string()))?;
|
||||
let sh = wb.sheet(0).ok_or(EngineError::NoDocument)?;
|
||||
let Some(u) = sh.used_range() else { return Err(EngineError::Other("The file is empty.".into())) };
|
||||
if u.count() > MAX_IMPORT_CELLS {
|
||||
return Err(too_many_cells());
|
||||
}
|
||||
(u.start.row..=u.end.row)
|
||||
.map(|r| {
|
||||
(u.start.col..=u.end.col)
|
||||
@@ -157,35 +129,3 @@ pub(super) fn queries(s: &mut Session, _: &Json) -> Result<Json> {
|
||||
.collect(),
|
||||
))
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use serde_json::json;
|
||||
|
||||
use crate::Session;
|
||||
|
||||
fn s() -> Session {
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
s
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn sparse_imports_are_refused_before_their_rows_are_built() {
|
||||
let mut s = s();
|
||||
let sheets = s.doc().unwrap().wb.sheets.len();
|
||||
// 5,000 records with one field each: 5,001 × 5,000 slots once every record gets every field.
|
||||
let records: Vec<_> = (0..5_000).map(|i| json!({ format!("k{i}"): i })).collect();
|
||||
let text = serde_json::to_string(&records).unwrap();
|
||||
let err = s.execute("data.getData", json!({"text": text})).expect_err("too many cells");
|
||||
assert!(err.to_string().contains("too large"), "{err}");
|
||||
// A CSV whose last line reaches 16,001 columns, 1,000 lines down.
|
||||
let text = format!("1{}{}x", "\n".repeat(999), ",".repeat(16_000));
|
||||
let err = s.execute("data.fromTextCsv", json!({"text": text})).expect_err("too many cells");
|
||||
assert!(err.to_string().contains("too large"), "{err}");
|
||||
assert_eq!(s.doc().unwrap().wb.sheets.len(), sheets);
|
||||
// Small imports still work.
|
||||
let out = s.execute("data.getData", json!({"text": r#"[{"a": 1}, {"b": 2}]"#})).unwrap();
|
||||
assert_eq!((out["rows"].as_u64(), out["cols"].as_u64()), (Some(3), Some(2)), "{out}");
|
||||
}
|
||||
}
|
||||
|
||||
@@ -36,28 +36,13 @@ fn max_periods(history: usize) -> u32 {
|
||||
u32::try_from(MAX_WORK / h).unwrap_or(u32::MAX).clamp(1, MAX_PERIODS)
|
||||
}
|
||||
|
||||
/// Most forecast rows when the model is fit to the first `fit` of `history` rows, which span
|
||||
/// `slots` timeline slots once their gaps are filled. With a Forecast Start inside the history,
|
||||
/// every row after it is forecast too (a hindcast row holds the same ETS formulas, each refitting
|
||||
/// the model over the `slots` slots), so those rows (counted from the raw rows) count against
|
||||
/// [`MAX_WORK`] before any forecast row does. `None` when the hindcast alone exceeds it.
|
||||
fn max_periods_from(history: usize, fit: usize, slots: usize) -> Option<u32> {
|
||||
if fit >= history {
|
||||
return Some(max_periods(history));
|
||||
}
|
||||
let f = u64::try_from(slots.max(fit)).unwrap_or(u64::MAX).max(1);
|
||||
let hindcast = u64::try_from(history - fit).unwrap_or(u64::MAX);
|
||||
let rest = (MAX_WORK / f).checked_sub(hindcast).filter(|r| *r >= 1)?;
|
||||
Some(u32::try_from(rest).unwrap_or(u32::MAX).min(MAX_PERIODS))
|
||||
}
|
||||
|
||||
pub fn specs() -> Vec<CommandSpec> {
|
||||
vec![cmd!(
|
||||
"data.forecastSheet",
|
||||
"Forecast Sheet",
|
||||
["Data", "Forecast"],
|
||||
None,
|
||||
"{timeline?: \"A2:A25\", values?: \"B2:B25\" (default: the first two columns of `range`, the selection or the current region; a text first cell is the header), end?: timeline value or date text | periods?: n (default: a third of the history, at least 1; ≤10,000 and history × periods ≤ 500,000, or with a forecastStart inside the history (hindcast rows are forecast too) (the greater of the fitted rows and fitted slots × (periods + rows from forecastStart on, that row included)) ≤ 500,000; history ≤ 10,000 rows), forecastStart?: a timeline value or date text on the history, before its last point (hindcast: the model is fit to the points before it, the forecast columns start there and run over the actuals; default and the last point: forecast after the history), confidenceInterval?: bool (default true), confidence?: 0.95 or 95, seasonality?: \"auto\" | 0 (none) | 2–8,760, fillMissing?: interpolate|zeros, aggregate?: average|count|counta|max|median|min|sum, statistics?: bool, chart?: line|column} → {sheet, table, chart, history, periods}. Inserts a sheet before the active one: timeline, values, Forecast (FORECAST.ETS), Lower/Upper Confidence Bound (FORECAST.ETS.CONFINT), optional FORECAST.ETS.STAT table, and a chart.",
|
||||
"{timeline?: \"A2:A25\", values?: \"B2:B25\" (default: the first two columns of `range`, the selection or the current region; a text first cell is the header), end?: timeline value or date text | periods?: n (default: a third of the history, at least 1; ≤10,000 and history × periods ≤ 500,000; history ≤ 10,000 rows), forecastStart?: a timeline value or date text on the history, before its last point (hindcast: the model is fit to the points before it, the forecast columns start there and run over the actuals; default and the last point: forecast after the history), confidenceInterval?: bool (default true), confidence?: 0.95 or 95, seasonality?: \"auto\" | 0 (none) | 2–8,760, fillMissing?: interpolate|zeros, aggregate?: average|count|counta|max|median|min|sum, statistics?: bool, chart?: line|column} → {sheet, table, chart, history, periods}. Inserts a sheet before the active one: timeline, values, Forecast (FORECAST.ETS), Lower/Upper Confidence Bound (FORECAST.ETS.CONFINT), optional FORECAST.ETS.STAT table, and a chart.",
|
||||
has_doc,
|
||||
forecast_sheet
|
||||
)]
|
||||
@@ -272,23 +257,13 @@ fn forecast_sheet(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
}
|
||||
};
|
||||
// The slots the model is fit over: the fit rows' span of the timeline, gaps included.
|
||||
let slots = match (fit < rows.len(), times.get(fit.wrapping_sub(1)), times.first()) {
|
||||
(true, Some(last_fit), Some(first)) => match (timeline.steps_after_last(*last_fit), timeline.steps_after_last(*first)) {
|
||||
(Some(b), Some(a)) => ((b - a).round() + 1.0).clamp(0.0, 1e9) as usize,
|
||||
_ => fit,
|
||||
},
|
||||
_ => fit,
|
||||
};
|
||||
let cap = max_periods_from(rows.len(), fit, slots).ok_or_else(|| {
|
||||
EngineError::Other("The forecast start is too early for this much history: choose a later forecast start or less history.".into())
|
||||
})?;
|
||||
let cap = max_periods(rows.len());
|
||||
let periods = match end {
|
||||
Some(end) => steps_to(&timeline, end, cap)?,
|
||||
None => match number(p, "periods")? {
|
||||
Some(n) if n >= 1.0 && n.ceil() <= f64::from(cap) => n.ceil() as u32,
|
||||
Some(_) => return Err(bad(CMD, format!("`periods` is from 1 to {cap} for {} rows of history", rows.len()))),
|
||||
None => default_periods(rows.len()).min(cap),
|
||||
None => default_periods(rows.len()),
|
||||
},
|
||||
};
|
||||
let future: Vec<f64> =
|
||||
@@ -557,51 +532,6 @@ mod tests {
|
||||
assert_eq!(s.doc().unwrap().wb.sheets.len(), sheets - 1);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn hindcast_rows_count_against_the_work_limit() {
|
||||
use super::{MAX_WORK, max_periods, max_periods_from};
|
||||
assert_eq!(max_periods_from(1_300, 1_300, 1_300), Some(max_periods(1_300)));
|
||||
// 2,000 rows fit to the first 1,899: 101 hindcast rows, then 162 forecast rows.
|
||||
assert_eq!(max_periods_from(2_000, 1_899, 1_899), Some(162));
|
||||
// Fit to the first 1,000: the 1,000 hindcast rows alone are twice the limit.
|
||||
assert_eq!(max_periods_from(2_000, 1_000, 1_000), None);
|
||||
// 800 rows fit to the first 400, which span 571 slots (gaps filled): each formula refits
|
||||
// 571 slots, so 475 forecast rows after the 400 hindcast rows, not 850.
|
||||
assert_eq!(max_periods_from(800, 400, 571), Some(475));
|
||||
assert!((475 + 400) * 571 <= MAX_WORK as usize);
|
||||
for (h, f) in [(2_000usize, 1_899usize), (10_000, 9_990), (100, 2), (60, 50)] {
|
||||
if let Some(p) = max_periods_from(h, f, f) {
|
||||
assert!((u64::from(p) + (h - f) as u64) * f as u64 <= MAX_WORK, "{h} {f}");
|
||||
}
|
||||
}
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
let mut rows = vec![json!(["t", "y"])];
|
||||
rows.extend((1..=2_000).map(|t| json!([t, (t % 7) as f64 + t as f64 * 0.5])));
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": rows})).unwrap();
|
||||
let sheets = s.doc().unwrap().wb.sheets.len();
|
||||
let err = s.execute("data.forecastSheet", json!({"range": "A1:B2001", "forecastStart": 1_001})).expect_err("hindcast too long");
|
||||
assert!(err.to_string().contains("forecast start is too early"), "{err}");
|
||||
let err = s.execute("data.forecastSheet", json!({"range": "A1:B2001", "forecastStart": 1_900, "periods": 163})).expect_err("past the cap");
|
||||
assert!(err.to_string().contains("from 1 to 162"), "{err}");
|
||||
assert_eq!(s.doc().unwrap().wb.sheets.len(), sheets);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_hindcast_cap_follows_the_slots_the_fit_spans_not_its_rows() {
|
||||
// 800 rows, 20% of the slots missing: the first 400 rows span 500 slots, so each ETS
|
||||
// formula refits 500 slots and the cap is 1,000 - 400 = 600 (it was 850 by rows).
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
let times: Vec<u32> = (1u32..).filter(|t| !matches!(t % 10, 3 | 6)).take(800).collect();
|
||||
let mut rows = vec![json!(["t", "y"])];
|
||||
rows.extend(times.iter().map(|t| json!([t, (t % 7) as f64 + f64::from(*t) * 0.5])));
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": rows})).unwrap();
|
||||
let start = times[400];
|
||||
let err = s.execute("data.forecastSheet", json!({"range": "A1:B801", "forecastStart": start, "periods": 601})).expect_err("past the cap");
|
||||
assert!(err.to_string().contains("from 1 to 600"), "{err}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn history_and_periods_are_capped() {
|
||||
use super::{MAX_HISTORY, MAX_PERIODS, MAX_WORK, default_periods, max_periods};
|
||||
|
||||
@@ -85,7 +85,7 @@ pub fn specs() -> Vec<CommandSpec> {
|
||||
"Format Cells…",
|
||||
["Home", "Cells", "Format"],
|
||||
Some("Cmd+1"),
|
||||
"{range?, style: {font?, fill?, borders?, align?: {reading_order?: Context|LeftToRight|RightToLeft|null, ...}, numFmt?, protection?}} (partial Style JSON merged in; null clears reading_order)",
|
||||
"{range?, style: {font?, fill?, borders?, align?, numFmt?, protection?}} (partial Style JSON merged in)",
|
||||
has_doc,
|
||||
format_cells
|
||||
),
|
||||
@@ -870,67 +870,6 @@ mod tests {
|
||||
assert_eq!(s.execute("cell.get", json!({"cell": "A1"})).unwrap()["style"]["num_fmt"], "0.00");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_vowel_and_direction_marks_add_no_width() {
|
||||
let plain = approx_text_width("محمد يدرس في المدرسة", 11.0, false);
|
||||
assert_eq!(approx_text_width("مُحَمَّد يَدْرُسُ فِي الْمَدْرَسَةِ", 11.0, false), plain);
|
||||
assert_eq!(approx_text_width("\u{200F}محمد يدرس في المدرسة\u{200C}", 11.0, false), plain);
|
||||
assert_eq!(approx_text_width("ب\u{0897}\u{0898}\u{089F}", 11.0, false), approx_text_width("ب", 11.0, false));
|
||||
assert!(approx_text_width("\u{08E2}", 11.0, false) > 0.0, "the disputed end of ayah is a visible sign");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_text_fits_a_column_about_as_wide_as_latin_text_of_the_same_length() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["مُحَمَّد يَدْرُسُ فِي الْمَدْرَسَةِ", "Hello there in school"]]})).unwrap();
|
||||
s.execute("home.autofitColumnWidth", json!({"cols": "A:B"})).unwrap();
|
||||
let doc = s.doc().unwrap();
|
||||
let width = |c: u32| doc.wb.active().and_then(|sh| sh.cols.get(&c)).and_then(|e| e.size).unwrap();
|
||||
let ratio = width(0) / width(1);
|
||||
assert!((0.75..=1.15).contains(&ratio), "Arabic {} vs Latin {}", width(0), width(1));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn autofit_of_many_separate_columns_fits_only_the_selected_ones() {
|
||||
let mut s = crate::Session::new();
|
||||
s.new_workbook();
|
||||
let long = "a much longer text than the others";
|
||||
// Columns A, C, E, G, I, K are selected (six runs), the ones between hold long text too.
|
||||
let row: Vec<_> = (0..11).map(|c| if c % 2 == 0 { json!("x") } else { json!(long) }).collect();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [row]})).unwrap();
|
||||
s.execute("home.autofitColumnWidth", json!({"cols": "A:A,C:C,E:E,G:G,I:I,K:K"})).unwrap();
|
||||
let doc = s.doc().unwrap();
|
||||
let size = |c: u32| doc.wb.active().and_then(|sh| sh.cols.get(&c)).and_then(|e| e.size);
|
||||
assert!([0, 2, 4, 6, 8, 10].iter().all(|&c| size(c).is_some()));
|
||||
assert!([1, 3, 5, 7, 9].iter().all(|&c| size(c).is_none()));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_word_ligatures_are_wider_than_letters() {
|
||||
let em = |t: &str| approx_text_width(t, 72.0, false) / 96.0;
|
||||
assert!(em("\u{FDFD}") > 8.0);
|
||||
assert!(em("\u{FDFA}") > 1.0);
|
||||
let honorifics = ['\u{FD40}', '\u{FD4B}', '\u{FD4C}', '\u{FD4F}', '\u{FDC8}', '\u{FDCF}', '\u{FDFE}', '\u{FDFF}'];
|
||||
for honorific in honorifics.into_iter().chain(['\u{FBC3}', '\u{FBD2}', '\u{FD90}', '\u{FD91}']) {
|
||||
assert!(em(&honorific.to_string()) > 1.0, "U+{:04X}", u32::from(honorific));
|
||||
}
|
||||
assert!(em("ب") < 0.6);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn arabic_digits_are_as_wide_as_ascii_digits_and_punctuation_as_a_letter() {
|
||||
let width = |t: &str| approx_text_width(t, 11.0, false);
|
||||
assert_eq!(width("١٢٣"), width("123"));
|
||||
assert_eq!(width("۴۵۶"), width("456"));
|
||||
assert_eq!(width("،؟"), width("بب"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn cjk_text_keeps_a_full_em_per_character() {
|
||||
assert_eq!(approx_text_width("中文", 12.0, false), 2.0 * 12.0 * 96.0 / 72.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn width_units() {
|
||||
assert_eq!(chars_to_points(8.43), 64.0);
|
||||
|
||||
@@ -136,55 +136,12 @@ pub fn approx_text_width(text: &str, size: f32, bold: bool) -> f32 {
|
||||
c if c.is_ascii_uppercase() => 0.65,
|
||||
c if c.is_ascii_digit() => 0.56,
|
||||
c if c.is_ascii() => 0.52,
|
||||
c if is_zero_width(c) => 0.0,
|
||||
c => arabic_em(c).unwrap_or(1.0),
|
||||
_ => 1.0,
|
||||
};
|
||||
}
|
||||
w * size * if bold { 1.07 } else { 1.0 } * 96.0 / 72.0
|
||||
}
|
||||
|
||||
/// Arabic vowel and Quranic marks, which sit on the previous letter, and invisible format
|
||||
/// characters such as direction marks and zero-width joiners.
|
||||
fn is_zero_width(c: char) -> bool {
|
||||
matches!(
|
||||
c,
|
||||
'\u{0610}'..='\u{061A}'
|
||||
| '\u{061C}'
|
||||
| '\u{064B}'..='\u{065F}'
|
||||
| '\u{0670}'
|
||||
| '\u{06D6}'..='\u{06DC}'
|
||||
| '\u{06DF}'..='\u{06E4}'
|
||||
| '\u{06E7}'..='\u{06E8}'
|
||||
| '\u{06EA}'..='\u{06ED}'
|
||||
| '\u{0897}'..='\u{089F}'
|
||||
| '\u{08CA}'..='\u{08E1}'
|
||||
| '\u{08E3}'..='\u{08FF}'
|
||||
| '\u{200B}'..='\u{200F}'
|
||||
| '\u{202A}'..='\u{202E}'
|
||||
| '\u{2060}'..='\u{2069}'
|
||||
| '\u{FE20}'..='\u{FE2F}'
|
||||
| '\u{FEFF}'
|
||||
)
|
||||
}
|
||||
|
||||
/// Width in ems of an Arabic-script character, measured from shaped Noto Sans Arabic, Noto Naskh
|
||||
/// Arabic and DejaVu Sans: joined letters average about half an em, word ligatures such as ﷺ and the
|
||||
/// rial sign about 1.2 em, the honorific ligatures (U+FD40–FD4B) about 1.6 em, and the bismillah
|
||||
/// ligature about 11 em. The later honorific ligatures (U+FBC3–FBD2, U+FD4C–FD4F, U+FD90–FD91,
|
||||
/// U+FDC8–FDCF) are not measured and get the 1.6 em of their siblings rather than the letter width.
|
||||
/// Arabic-Indic digits are as wide as ASCII digits (0.53–0.57 em in Noto Sans Arabic, Arial, Tahoma
|
||||
/// and Times New Roman); Arabic punctuation (0.2–0.5 em in the same fonts) keeps the letter width.
|
||||
fn arabic_em(c: char) -> Option<f32> {
|
||||
match c {
|
||||
'\u{FDFD}' => Some(11.0),
|
||||
'\u{FBC3}'..='\u{FBD2}' | '\u{FD40}'..='\u{FD4F}' | '\u{FD90}'..='\u{FD91}' | '\u{FDC8}'..='\u{FDCF}' => Some(1.6),
|
||||
'\u{FDF0}'..='\u{FDFC}' | '\u{FDFE}'..='\u{FDFF}' => Some(1.2),
|
||||
'\u{0660}'..='\u{0669}' | '\u{06F0}'..='\u{06F9}' => Some(0.56),
|
||||
'\u{0600}'..='\u{06FF}' | '\u{0750}'..='\u{077F}' | '\u{0870}'..='\u{08FF}' | '\u{FB50}'..='\u{FDFF}' | '\u{FE70}'..='\u{FEFC}' => Some(0.48),
|
||||
_ => None,
|
||||
}
|
||||
}
|
||||
|
||||
/// Distinct texts AutoFit measures exactly; past this many, `Session::text_width` is asked for a
|
||||
/// quick measure so a huge column doesn't lay out every value.
|
||||
const EXACT_TEXT_WIDTHS: usize = 2_000;
|
||||
@@ -194,35 +151,16 @@ pub(super) fn autofit_cols(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let measure = s.text_width.clone();
|
||||
edit(s, |cx| {
|
||||
let Some(sh) = cx.wb.sheet(si) else { return Ok(Json::Null) };
|
||||
let mut widths = Vec::new();
|
||||
let mut measured: std::collections::HashMap<StyleId, std::collections::HashMap<String, f32>> = Default::default();
|
||||
let mut distinct = 0usize;
|
||||
// Selected columns, then one pass over the populated cells (the store is row-oriented, so a
|
||||
// scan per column would revisit every row for each selected column, empty or not).
|
||||
let cols: std::collections::BTreeSet<u32> = lines.iter().flat_map(|&(a, b)| a..=b.min(a + 16383)).collect();
|
||||
let mut maxes: std::collections::BTreeMap<u32, f32> = Default::default();
|
||||
if let Some(u) = sh.used_range() {
|
||||
// One pass per run of adjacent selected columns, so columns between distant selections
|
||||
// (A and XFD) are never visited.
|
||||
let mut runs: Vec<(u32, u32)> = Vec::new();
|
||||
for &c in &cols {
|
||||
match runs.last_mut() {
|
||||
Some((_, end)) if *end + 1 == c => *end = c,
|
||||
_ => runs.push((c, c)),
|
||||
}
|
||||
}
|
||||
// Every run is a pass over all the stored rows: many runs (alternating columns) are one
|
||||
// pass over their whole span instead, the unselected columns skipped.
|
||||
if runs.len() > 4
|
||||
&& let (Some(&(first, _)), Some(&(_, last))) = (runs.first(), runs.last())
|
||||
{
|
||||
runs = vec![(first, last)];
|
||||
}
|
||||
for (first, last) in runs {
|
||||
let span = RangeRef::new(CellRef::new(u.start.row, first), CellRef::new(u.end.row, last));
|
||||
for (cell, _) in sh.cells.iter_range(span) {
|
||||
if !cols.contains(&cell.col) {
|
||||
continue;
|
||||
}
|
||||
let used = sh.used_range();
|
||||
for (a, b) in lines {
|
||||
for c in a..=b.min(a + 16383) {
|
||||
let Some(u) = used else { continue };
|
||||
let col = RangeRef::new(CellRef::new(u.start.row, c), CellRef::new(u.end.row, c));
|
||||
let mut max: f32 = 0.0;
|
||||
for (cell, _) in sh.cells.iter_range(col) {
|
||||
if sh.merge_at(cell).is_some_and(|m| !m.is_single()) {
|
||||
continue;
|
||||
}
|
||||
@@ -245,14 +183,14 @@ pub(super) fn autofit_cols(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
None => approx_text_width(line, st.font.size, st.font.bold),
|
||||
};
|
||||
let w = w + 6.0 + st.align.indent as f32 * 9.0 + sh.filter_button_space(cell);
|
||||
let max = maxes.entry(cell.col).or_insert(0.0);
|
||||
*max = max.max(w);
|
||||
max = max.max(w + 6.0 + st.align.indent as f32 * 9.0 + sh.filter_button_space(cell));
|
||||
}
|
||||
}
|
||||
if max > 0.0 {
|
||||
widths.push((c, max.clamp(8.0, 1785.0)));
|
||||
}
|
||||
}
|
||||
}
|
||||
let widths: Vec<(u32, f32)> = maxes.into_iter().filter(|&(_, m)| m > 0.0).map(|(c, m)| (c, m.clamp(8.0, 1785.0))).collect();
|
||||
let sh = cx.sheet_mut(si)?;
|
||||
for (c, w) in widths {
|
||||
let e = sh.cols.entry(c).or_default();
|
||||
|
||||
@@ -209,41 +209,6 @@ fn show_values_as() {
|
||||
assert_eq!(text(&s, "B3"), "Biggest sale");
|
||||
}
|
||||
|
||||
/// Running Total over nested rows adds up each child's value across the top-level items, in the
|
||||
/// order they appear: the same value as summing the plain values of the earlier items' rows.
|
||||
#[test]
|
||||
fn running_total_over_nested_rows_adds_up_across_the_top_level_items() {
|
||||
let mut s = sales();
|
||||
s.execute("insert.pivotTable", json!({"rows": ["Region", "Product"], "values": ["Sales"]})).unwrap();
|
||||
let r = pt(&s).last_range.unwrap();
|
||||
let rows: Vec<u32> = (r.start.row + 1..r.end.row).collect();
|
||||
let read = |s: &Session| -> Vec<(String, f64)> {
|
||||
rows.iter().map(|row| (text(s, &CellRef::new(*row, 0).a1()), val(s, &CellRef::new(*row, 1).a1()).as_f64().unwrap_or(0.0))).collect()
|
||||
};
|
||||
let before = read(&s);
|
||||
s.execute("pivot.valueSettings", json!({"index": 0, "showAs": "runningTotal"})).unwrap();
|
||||
let after = read(&s);
|
||||
// Region rows carry the report's top-level items (not indented); the others are products.
|
||||
let indent = |s: &Session, row: u32| {
|
||||
let d = s.doc().unwrap();
|
||||
let sh = d.wb.active().unwrap();
|
||||
d.wb.styles.get(sh.style_id(CellRef::new(row, 0))).align.indent
|
||||
};
|
||||
let (mut regions, mut products) = (0.0, std::collections::HashMap::<String, f64>::new());
|
||||
for (i, row) in rows.iter().enumerate() {
|
||||
let name = &before[i].0;
|
||||
let total = if indent(&s, *row) == 0 {
|
||||
regions += before[i].1;
|
||||
regions
|
||||
} else {
|
||||
let t = products.entry(name.clone()).or_insert(0.0);
|
||||
*t += before[i].1;
|
||||
*t
|
||||
};
|
||||
assert!((after[i].1 - total).abs() < 1e-9, "row {name}: {} vs {total}", after[i].1);
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn compact_vs_tabular_shapes() {
|
||||
let mut s = sales();
|
||||
|
||||
@@ -32,69 +32,8 @@ use header_footer::draw_header_footer;
|
||||
pub use header_footer::{
|
||||
HF_FIELDS, HfContext, HfSection, codes_to_fields, expand_header_footer, fields_to_codes, hf_date_time, join_header_footer, split_header_footer,
|
||||
};
|
||||
type DecodedPicture = (Vec<u8>, u32, u32);
|
||||
|
||||
/// What one PDF render shares across the pages an object spans.
|
||||
#[derive(Default)]
|
||||
struct PicCache {
|
||||
/// Decoded pictures (bytes, width, height) keyed by (sheet, picture); `None` = undecodable.
|
||||
decoded: std::collections::HashMap<(usize, usize), Option<DecodedPicture>>,
|
||||
/// Laid-out charts (with their [`chart_cost`]) keyed by (sheet, chart), rendered once however
|
||||
/// many pages they touch. Only charts that can still be drawn are kept.
|
||||
charts: std::collections::HashMap<(usize, usize), (Vec<gridcraft_chart::Prim>, usize)>,
|
||||
/// Charts too big for what was left of the budget: never drawn, so not rendered again.
|
||||
rejected_charts: std::collections::HashSet<(usize, usize)>,
|
||||
/// Bytes of picture pixels embedded so far, and chart cost units emitted so far. Each page an
|
||||
/// object touches gets its own copy in the PDF, so a workbook-controlled anchor over thousands
|
||||
/// of pages would multiply them without bound.
|
||||
image_bytes: usize,
|
||||
chart_prims: usize,
|
||||
}
|
||||
|
||||
impl PicCache {
|
||||
/// The chart's primitives for one more page, or `None` once they no longer fit
|
||||
/// [`MAX_CHART_PRIMS`]. `render` runs at most once per chart; a chart that doesn't fit is
|
||||
/// dropped at once and remembered, so its vectors are never retained.
|
||||
fn chart(&mut self, key: (usize, usize), render: impl FnOnce() -> Vec<gridcraft_chart::Prim>) -> Option<&[gridcraft_chart::Prim]> {
|
||||
if self.rejected_charts.contains(&key) {
|
||||
return None;
|
||||
}
|
||||
let (prims, cost) = match self.charts.remove(&key) {
|
||||
Some(cached) => cached,
|
||||
None => {
|
||||
let prims = render();
|
||||
let cost = chart_cost(&prims);
|
||||
(prims, cost)
|
||||
}
|
||||
};
|
||||
if self.chart_prims.saturating_add(cost) > MAX_CHART_PRIMS {
|
||||
self.rejected_charts.insert(key);
|
||||
return None;
|
||||
}
|
||||
self.chart_prims += cost;
|
||||
Some(&self.charts.entry(key).or_insert((prims, cost)).0)
|
||||
}
|
||||
}
|
||||
|
||||
/// What a chart's drawing costs in the PDF: one unit per primitive plus one per path point (a
|
||||
/// smoothed line holds thousands of points, each written out) and per point of a wedge's outline.
|
||||
fn chart_cost(prims: &[gridcraft_chart::Prim]) -> usize {
|
||||
use gridcraft_chart::Prim;
|
||||
prims
|
||||
.iter()
|
||||
.map(|p| match p {
|
||||
Prim::Line { pts, .. } | Prim::Polygon { pts, .. } => pts.len().saturating_add(1),
|
||||
// At most 400 steps along each of the outer and inner arcs (see `draw_prims`).
|
||||
Prim::Wedge { .. } => 803,
|
||||
_ => 1,
|
||||
})
|
||||
.fold(0, usize::saturating_add)
|
||||
}
|
||||
|
||||
/// Most picture pixel bytes (RGBA, before PDF packing) one render embeds across all pages.
|
||||
const MAX_EMBEDDED_IMAGE_BYTES: usize = 512 * 1024 * 1024;
|
||||
/// Most chart cost units (primitives and path points, see [`chart_cost`]) one render draws across all pages.
|
||||
const MAX_CHART_PRIMS: usize = 4_000_000;
|
||||
/// Decoded pictures (bytes, width, height) keyed by (sheet, picture); `None` = undecodable.
|
||||
type PicCache = std::collections::HashMap<(usize, usize), Option<(Vec<u8>, u32, u32)>>;
|
||||
|
||||
#[cfg(test)]
|
||||
use layout::bands;
|
||||
@@ -163,7 +102,7 @@ pub fn render_pdf(wb: &Workbook, layouts: &[SheetLayout], file: &str, path: &str
|
||||
let mut printed_blocks = Vec::new();
|
||||
// Decoded pictures are reused across the pages their anchor covers: a floating picture
|
||||
// spanning 40 pages was decoded (and pushed into the PDF) 40 times (#2389).
|
||||
let mut pics = PicCache::default();
|
||||
let mut pics = PicCache::new();
|
||||
for lay in layouts {
|
||||
let Some(sh) = wb.sheet(lay.sheet) else { continue };
|
||||
let mut cf = if sh.cond_formats.is_empty() { None } else { Some(crate::cf::CfCache::new()) };
|
||||
|
||||
@@ -228,82 +228,6 @@ fn draw_border(pg: &mut Page, b: &BorderLine, wb: &Workbook, start: Point, end:
|
||||
pg.stroke_line_dashed(Point::new(x0, y0), Point::new(x1, y1), lw, color, &border_dash(b.style, k));
|
||||
}
|
||||
|
||||
/// Merged areas active at one row of a downward sweep, kept disjoint: by first column, their last
|
||||
/// column and last row. Adding, dropping and looking up a merge costs log(active merges).
|
||||
#[derive(Default)]
|
||||
struct MergeSweep {
|
||||
cols: std::collections::BTreeMap<u32, (u32, u32)>,
|
||||
/// (last row, first column) of the active merges, soonest ending first.
|
||||
ends: std::collections::BinaryHeap<std::cmp::Reverse<(u32, u32)>>,
|
||||
}
|
||||
|
||||
impl MergeSweep {
|
||||
/// Moves down to `row` (rows never decrease), dropping the merges that end above it.
|
||||
fn advance(&mut self, row: u32) {
|
||||
while let Some(std::cmp::Reverse((end, col))) = self.ends.peek().copied() {
|
||||
if end >= row {
|
||||
break;
|
||||
}
|
||||
self.ends.pop();
|
||||
self.cols.remove(&col);
|
||||
}
|
||||
}
|
||||
|
||||
/// Does an active merge meet columns `c0..=c1`? The intervals are disjoint, so only the last
|
||||
/// one starting at or before `c1` can.
|
||||
fn meets(&self, c0: u32, c1: u32) -> bool {
|
||||
self.cols.range(..=c1).next_back().is_some_and(|(_, (end, _))| *end >= c0)
|
||||
}
|
||||
|
||||
/// Adds merge `m`, at row `row`, unless it overlaps an active one.
|
||||
fn add(&mut self, m: RangeRef, row: u32) -> bool {
|
||||
self.advance(row);
|
||||
if self.meets(m.start.col, m.end.col) {
|
||||
return false;
|
||||
}
|
||||
self.cols.insert(m.start.col, (m.end.col, m.end.row));
|
||||
self.ends.push(std::cmp::Reverse((m.end.row, m.start.col)));
|
||||
true
|
||||
}
|
||||
}
|
||||
|
||||
/// The merged areas that meet `block`, ordered by the row where they enter it, then by column,
|
||||
/// leaving out any that overlaps one kept before it (only a damaged file has overlapping merges;
|
||||
/// Excel drops them when it repairs one). One sweep down the rows instead of a scan of every
|
||||
/// merge per cell.
|
||||
fn block_merges(sh: &Sheet, block: RangeRef) -> Vec<RangeRef> {
|
||||
let top = |m: &RangeRef| m.start.row.max(block.start.row);
|
||||
let mut merges: Vec<RangeRef> = sh.merges.iter().filter(|m| m.intersects(&block)).copied().collect();
|
||||
merges.sort_by_key(|m| (top(m), m.start.col));
|
||||
let mut sweep = MergeSweep::default();
|
||||
merges.retain(|m| sweep.add(*m, top(m)));
|
||||
merges
|
||||
}
|
||||
|
||||
/// Answers "is this cell inside a merge?" for cells asked in row-major order, from the merges of
|
||||
/// [`block_merges`].
|
||||
struct Covered<'a> {
|
||||
merges: &'a [RangeRef],
|
||||
top: u32,
|
||||
next: usize,
|
||||
sweep: MergeSweep,
|
||||
}
|
||||
|
||||
impl<'a> Covered<'a> {
|
||||
fn new(merges: &'a [RangeRef], block: RangeRef) -> Self {
|
||||
Self { merges, top: block.start.row, next: 0, sweep: MergeSweep::default() }
|
||||
}
|
||||
|
||||
fn covers(&mut self, c: CellRef) -> bool {
|
||||
while let Some(m) = self.merges.get(self.next).filter(|m| m.start.row.max(self.top) <= c.row) {
|
||||
self.sweep.add(*m, m.start.row.max(self.top));
|
||||
self.next += 1;
|
||||
}
|
||||
self.sweep.advance(c.row);
|
||||
self.sweep.meets(c.col, c.col)
|
||||
}
|
||||
}
|
||||
|
||||
/// Draws the cells of a block: fills, gridlines, data bars, borders and text.
|
||||
pub(super) fn draw_block(pg: &mut Page, b: &Block, gridlines: bool, cf: &mut Option<crate::cf::CfCache>) {
|
||||
let (bx, by) = (b.xs.first().copied().unwrap_or(0.0), b.ys.first().copied().unwrap_or(0.0));
|
||||
@@ -313,18 +237,17 @@ pub(super) fn draw_block(pg: &mut Page, b: &Block, gridlines: bool, cf: &mut Opt
|
||||
}
|
||||
let block = b.range();
|
||||
let sh = b.sh;
|
||||
let merges = block_merges(sh, block);
|
||||
let by_start: std::collections::HashMap<CellRef, RangeRef> = merges.iter().map(|m| (m.start, *m)).collect();
|
||||
// Cells worth visiting, in row-major order (which `Covered` needs).
|
||||
let mut covered = Covered::new(&merges, block);
|
||||
let merges: Vec<RangeRef> = sh.merges.iter().filter(|m| m.intersects(&block)).copied().collect();
|
||||
let covered = |c: CellRef| merges.iter().any(|m| m.contains(c));
|
||||
// Cells worth visiting.
|
||||
let cells: Vec<CellRef> = if block.count() <= DENSE_BLOCK {
|
||||
block.iter().filter(|c| !covered.covers(*c)).collect()
|
||||
block.iter().filter(|c| !covered(*c)).collect()
|
||||
} else {
|
||||
let mut set: Vec<CellRef> = sh.cells.iter_range(block).map(|(c, _)| c).collect();
|
||||
set.extend(sh.spill.range(block.start..=block.end).map(|(c, _)| *c).filter(|c| block.contains(*c)));
|
||||
set.sort_unstable();
|
||||
set.dedup();
|
||||
set.retain(|c| !covered.covers(*c));
|
||||
set.retain(|c| !covered(*c));
|
||||
set
|
||||
};
|
||||
pg.save();
|
||||
@@ -354,14 +277,14 @@ pub(super) fn draw_block(pg: &mut Page, b: &Block, gridlines: bool, cf: &mut Opt
|
||||
.filter(|(c, l)| {
|
||||
l.fill.is_some()
|
||||
|| l.cf.is_some()
|
||||
|| by_start.contains_key(c)
|
||||
|| merges.iter().any(|m| m.start == *c)
|
||||
|| !sh.value(*c).is_empty()
|
||||
|| l.style.borders != Borders::default()
|
||||
|| l.style.num_fmt.as_str() == "checkbox"
|
||||
})
|
||||
.collect();
|
||||
let rect_of = |c: CellRef| -> (f32, f32, f32, f32, bool) {
|
||||
match by_start.get(&c) {
|
||||
match merges.iter().find(|m| m.start == c) {
|
||||
Some(m) => {
|
||||
let (x, y, w, h) = b.rect(*m);
|
||||
(x, y, w, h, true)
|
||||
@@ -399,7 +322,7 @@ pub(super) fn draw_block(pg: &mut Page, b: &Block, gridlines: bool, cf: &mut Opt
|
||||
continue;
|
||||
}
|
||||
let bd = &l.style.borders;
|
||||
let (right, bottom) = match by_start.get(c).filter(|_| merged) {
|
||||
let (right, bottom) = match merges.iter().find(|m| m.start == *c).filter(|_| merged) {
|
||||
Some(m) => (
|
||||
b.wb.styles.get(sh.style_id(CellRef::new(m.start.row, m.end.col))).borders.right,
|
||||
b.wb.styles.get(sh.style_id(CellRef::new(m.end.row, m.start.col))).borders.bottom,
|
||||
@@ -484,12 +407,11 @@ pub(super) struct PrintedBlock {
|
||||
pub(super) fn draw_block_links(pg: &mut Page, b: &Block, page: usize, pending: &mut Vec<PendingPdfLink>) {
|
||||
let range = b.range();
|
||||
let (Some(&bx), Some(&by), Some(&right), Some(&bottom)) = (b.xs.first(), b.ys.first(), b.xs.last(), b.ys.last()) else { return };
|
||||
let by_start: std::collections::HashMap<CellRef, RangeRef> = block_merges(b.sh, range).into_iter().map(|m| (m.start, m)).collect();
|
||||
for (cell, link) in b.sh.hyperlinks.range(range.start..=range.end) {
|
||||
if !range.contains(*cell) {
|
||||
continue;
|
||||
}
|
||||
let cell_range = by_start.get(cell).copied().unwrap_or(RangeRef::cell(*cell));
|
||||
let cell_range = b.sh.merges.iter().find(|m| m.start == *cell).copied().unwrap_or(RangeRef::cell(*cell));
|
||||
let (x, y, w, h) = b.rect(cell_range);
|
||||
let (x0, y0, x1, y1) = (x.max(bx), y.max(by), (x + w).min(right), (y + h).min(bottom));
|
||||
if x1 <= x0 || y1 <= y0 {
|
||||
@@ -581,44 +503,3 @@ pub(super) fn make_block<'a>(
|
||||
}
|
||||
Block { wb, si, sh, rows, cols, xs, ys, k, scale: lay.scale }
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod merge_tests {
|
||||
use super::{CellRef, Covered, RangeRef, Sheet, block_merges};
|
||||
|
||||
fn r(a: &str) -> RangeRef {
|
||||
RangeRef::parse(a).unwrap()
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn block_merges_drop_overlaps_and_cover_like_a_scan() {
|
||||
let mut sh = Sheet::new("S");
|
||||
sh.merges = vec![r("B2:C3"), r("E1:F8"), r("B2:B5"), r("A10:D10"), r("H3:H4"), r("Z1:Z2"), r("C3:D4"), r("A1:A1048576")];
|
||||
let block = r("B1:H12");
|
||||
let merges = block_merges(&sh, block);
|
||||
// B2:B5 and C3:D4 overlap B2:C3, which comes first; Z1:Z2 is outside the block.
|
||||
assert_eq!(merges, vec![r("E1:F8"), r("B2:C3"), r("H3:H4"), r("A10:D10")]);
|
||||
let mut covered = Covered::new(&merges, block);
|
||||
for c in block.iter() {
|
||||
assert_eq!(covered.covers(c), merges.iter().any(|m| m.contains(c)), "{}", c.a1());
|
||||
}
|
||||
// Cells asked sparsely, still in order.
|
||||
let mut covered = Covered::new(&merges, block);
|
||||
for (a, want) in [("C1", false), ("F5", true), ("G5", false), ("D10", true), ("H12", false)] {
|
||||
assert_eq!(covered.covers(CellRef::parse(a).unwrap()), want, "{a}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn many_stacked_merges_are_swept_once() {
|
||||
// 50,000 merges over the same cells: one is kept, and the block is checked against it
|
||||
// without comparing every cell with every merge.
|
||||
let mut sh = Sheet::new("S");
|
||||
sh.merges = (0..50_000u32).map(|i| RangeRef::new(CellRef::new(0, 0), CellRef::new(199, 1 + i % 50))).collect();
|
||||
let block = r("A1:AZ200");
|
||||
let merges = block_merges(&sh, block);
|
||||
assert_eq!(merges, vec![r("A1:B200")]);
|
||||
let mut covered = Covered::new(&merges, block);
|
||||
assert_eq!(block.iter().filter(|c| covered.covers(*c)).count(), 400);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -224,12 +224,6 @@ pub fn layout_sheet(wb: &Workbook, si: usize, opts: &PrintOptions, range_overrid
|
||||
Some(r) => trim_full(sh, r),
|
||||
None => print_range(sh).unwrap_or(RangeRef::cell(CellRef::new(0, 0))),
|
||||
};
|
||||
// A print area from a file or JSON can name coordinates past the grid; pagination walks every
|
||||
// line of the area (zero-height rows never fill a band), so it is held to the grid first.
|
||||
let area = RangeRef::new(
|
||||
CellRef::new(area.start.row.min(MAX_ROWS - 1), area.start.col.min(MAX_COLS - 1)),
|
||||
CellRef::new(area.end.row.min(MAX_ROWS - 1), area.end.col.min(MAX_COLS - 1)),
|
||||
);
|
||||
let title_rows = title_span(st.title_rows, MAX_ROWS);
|
||||
let title_cols = title_span(st.title_cols, MAX_COLS);
|
||||
let digits = format!("{}", area.end.row + 1).len() as f32;
|
||||
@@ -385,19 +379,3 @@ pub(super) fn segments(titles: Option<(u32, u32)>, band: (u32, u32)) -> Vec<(u32
|
||||
pub(super) fn seg_size(sh: &Sheet, seg: (u32, u32), rows: bool) -> f32 {
|
||||
if rows { (seg.0..=seg.1).map(|r| sh.row_height(r)).sum() } else { (seg.0..=seg.1).map(|c| sh.col_width(c)).sum() }
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn a_print_area_past_the_grid_is_held_to_the_grid() {
|
||||
let mut wb = Workbook::new();
|
||||
let sh = wb.sheet_mut(0).unwrap();
|
||||
sh.default_row_height = 0.0;
|
||||
sh.print.print_area = Some(RangeRef { start: CellRef::new(0, 0), end: CellRef::new(u32::MAX, u32::MAX) });
|
||||
let lay = layout_sheet(&wb, 0, &PrintOptions::default(), None).unwrap();
|
||||
assert_eq!((lay.area.end.row, lay.area.end.col), (MAX_ROWS - 1, MAX_COLS - 1));
|
||||
assert!(lay.pages.len() <= MAX_PAGES);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -88,14 +88,9 @@ pub(super) fn draw_objects(pg: &mut Page, wb: &Workbook, si: usize, sh: &Sheet,
|
||||
let (x, y, w, h) = place(r);
|
||||
// About 150 dpi is plenty for print.
|
||||
let (mw, mh) = (((w / 72.0) * 150.0).ceil().clamp(1.0, 4000.0) as u32, ((h / 72.0) * 150.0).ceil().clamp(1.0, 4000.0) as u32);
|
||||
let decoded = pics.decoded.entry((si, imi)).or_insert_with(|| decode_image(&im.data, mw, mh));
|
||||
let decoded = pics.entry((si, imi)).or_insert_with(|| decode_image(&im.data, mw, mh));
|
||||
if let Some((px, pw, ph)) = decoded.as_ref() {
|
||||
// Every page the picture touches embeds its own copy: past the budget it is left out.
|
||||
let bytes = px.len();
|
||||
if pics.image_bytes.saturating_add(bytes) <= super::MAX_EMBEDDED_IMAGE_BYTES {
|
||||
pics.image_bytes += bytes;
|
||||
pg.image_rgba(Rect::new(x, y, w, h), px, [*pw, *ph]);
|
||||
}
|
||||
pg.image_rgba(Rect::new(x, y, w, h), px, [*pw, *ph]);
|
||||
}
|
||||
}
|
||||
for shp in &sh.shapes {
|
||||
@@ -125,20 +120,15 @@ pub(super) fn draw_objects(pg: &mut Page, wb: &Workbook, si: usize, sh: &Sheet,
|
||||
}
|
||||
draw_shape(pg, wb, shp, b, (x, y, w, h));
|
||||
}
|
||||
for (chi, ch) in sh.charts.iter().enumerate() {
|
||||
for ch in &sh.charts {
|
||||
let r = anchor_rect(sh, &ch.anchor);
|
||||
if !visible(r) {
|
||||
continue;
|
||||
}
|
||||
let (x, y, _, _) = place(r);
|
||||
// As for pictures, each page gets its own copy of the drawing commands.
|
||||
let prims = pics.chart((si, chi), || {
|
||||
let data = gridcraft_chart::resolve(wb, si, ch);
|
||||
gridcraft_chart::render(ch, &data, r.2 as f32, r.3 as f32, &PdfMeasure)
|
||||
});
|
||||
if let Some(prims) = prims {
|
||||
draw_prims(pg, prims, Xf { ox: x, oy: y, k: b.k });
|
||||
}
|
||||
let data = gridcraft_chart::resolve(wb, si, ch);
|
||||
let prims = gridcraft_chart::render(ch, &data, r.2 as f32, r.3 as f32, &PdfMeasure);
|
||||
draw_prims(pg, &prims, Xf { ox: x, oy: y, k: b.k });
|
||||
}
|
||||
pg.restore();
|
||||
}
|
||||
|
||||
@@ -938,27 +938,3 @@ fn equations_print_with_the_symbol_font() {
|
||||
// A glyph no standard font has is a [?] box.
|
||||
assert!(contains(&pdf, "(?) Tj"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn chart_budget_counts_path_points_and_drops_charts_that_do_not_fit() {
|
||||
use gridcraft_chart::Prim;
|
||||
let line = |n: usize| Prim::Line { pts: vec![[0.0, 0.0]; n], color: [0, 0, 0, 255], width: 1.0, dash: false };
|
||||
// A smoothed line is its points, not one unit.
|
||||
assert_eq!(chart_cost(&[line(4000), Prim::Circle { cx: 0.0, cy: 0.0, r: 1.0, fill: None, stroke: None }]), 4002);
|
||||
// Room for exactly two copies of a 1,000-point chart.
|
||||
let mut pics = PicCache { chart_prims: MAX_CHART_PRIMS - 2002, ..Default::default() };
|
||||
let mut renders = 0;
|
||||
for expect in [true, true, false, false] {
|
||||
let got = pics.chart((0, 0), || {
|
||||
renders += 1;
|
||||
vec![line(1000)]
|
||||
});
|
||||
assert_eq!(got.is_some(), expect);
|
||||
}
|
||||
assert_eq!(renders, 1, "a rejected chart is not rendered again");
|
||||
assert!(pics.charts.is_empty(), "and its primitives are not kept");
|
||||
// A chart too big for what is left is dropped without being cached at all.
|
||||
let mut pics = PicCache { chart_prims: MAX_CHART_PRIMS - 10, ..Default::default() };
|
||||
assert!(pics.chart((0, 1), || vec![line(1000)]).is_none());
|
||||
assert!(pics.charts.is_empty() && pics.rejected_charts.contains(&(0, 1)));
|
||||
}
|
||||
|
||||
@@ -163,8 +163,6 @@ fn protect_sheet(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
let what = d.wb.sheet(sheet).map(|sh| sh.name.clone()).unwrap_or_default();
|
||||
let prot = SheetProtection { password: str_param(p, "password").filter(|x| !x.is_empty()).map(|pw| new_password_hash(d, &what, pw)), ..prot };
|
||||
edit(s, |cx| {
|
||||
// FORMULATEXT reads a sheet's protection: its results are recalculated.
|
||||
cx.recalc_all = true;
|
||||
cx.sheet_mut(sheet)?.protection = Some(prot.clone());
|
||||
Ok(Json::Null)
|
||||
})
|
||||
@@ -184,10 +182,7 @@ fn unprotect_sheet(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
}
|
||||
edit(s, |cx| {
|
||||
// Only removing protection changes what FORMULATEXT reads; a no-op unprotect skips the full recalculation.
|
||||
if cx.sheet_mut(sheet)?.protection.take().is_some() {
|
||||
cx.recalc_all = true;
|
||||
}
|
||||
cx.sheet_mut(sheet)?.protection = None;
|
||||
Ok(Json::Null)
|
||||
})
|
||||
}
|
||||
@@ -409,21 +404,6 @@ mod tests {
|
||||
assert_eq!(at(&s), (0, "A1".to_string()));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn protecting_and_unprotecting_leave_table_headers_alone() {
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["a", "b"], [1, 2]]})).unwrap();
|
||||
s.execute("insert.table", json!({"range": "A1:B2", "header": true})).unwrap();
|
||||
// An imported header can hold a number: only a structural edit would turn it back into text.
|
||||
let d = s.doc_mut().unwrap();
|
||||
std::sync::Arc::make_mut(&mut d.wb).sheet_mut(0).unwrap().set_value(gridcraft_core::CellRef::new(0, 0), gridcraft_core::Value::Number(7.0));
|
||||
s.execute("review.protectSheet", json!({})).unwrap();
|
||||
s.execute("review.unprotectSheet", json!({})).unwrap();
|
||||
let v = s.doc().unwrap().wb.sheet(0).unwrap().value(gridcraft_core::CellRef::new(0, 0));
|
||||
assert_eq!(v, gridcraft_core::Value::Number(7.0));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn protection_that_forbids_selecting_locked_cells_is_enforced_and_kept() {
|
||||
let mut s = book();
|
||||
|
||||
@@ -52,19 +52,14 @@ pub(super) fn abs_ref(c: CellRef) -> String {
|
||||
format!("${}${}", gridcraft_core::col_to_letters(c.col), u64::from(c.row) + 1)
|
||||
}
|
||||
|
||||
/// How many cells to the left and above a cell its name is searched for.
|
||||
const NAME_REACH: u32 = 1_000;
|
||||
|
||||
/// A cell's name in the reports: the nearest text to its left and the nearest text above it.
|
||||
pub(super) fn cell_name(sh: &Sheet, c: CellRef) -> String {
|
||||
let text = |c: CellRef| match sh.value(c) {
|
||||
Value::Text(t) if !t.trim().is_empty() => Some(t.to_string()),
|
||||
_ => None,
|
||||
};
|
||||
// Labels are next to their cells: looking further than this on a sparse sheet only costs time
|
||||
// (a thousand cells near the last row would each probe a million empty coordinates).
|
||||
let left = (c.col.saturating_sub(NAME_REACH)..c.col).rev().find_map(|col| text(CellRef::new(c.row, col)));
|
||||
let top = (c.row.saturating_sub(NAME_REACH)..c.row).rev().find_map(|row| text(CellRef::new(row, c.col)));
|
||||
let left = (0..c.col).rev().find_map(|col| text(CellRef::new(c.row, col)));
|
||||
let top = (0..c.row).rev().find_map(|row| text(CellRef::new(row, c.col)));
|
||||
[left, top].into_iter().flatten().collect::<Vec<_>>().join(" ")
|
||||
}
|
||||
|
||||
@@ -443,17 +438,3 @@ pub(super) fn report(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn a_name_is_searched_only_a_bounded_distance() {
|
||||
let mut sh = Sheet::new("S".to_string());
|
||||
sh.set_cell(CellRef::new(0, 0), gridcraft_model::Cell::value(Value::text("Far")));
|
||||
sh.set_cell(CellRef::new(5, 4), gridcraft_model::Cell::value(Value::text("Near")));
|
||||
assert_eq!(cell_name(&sh, CellRef::new(5, 5)), "Near");
|
||||
assert_eq!(cell_name(&sh, CellRef::new(0, NAME_REACH + 1)), "");
|
||||
}
|
||||
}
|
||||
|
||||
@@ -83,9 +83,6 @@ pub(crate) fn criterion_matches(v: &Value, text: &str, crit: &Value) -> bool {
|
||||
}
|
||||
}
|
||||
|
||||
/// Most (list row, criteria row) pairs one Advanced Filter tests.
|
||||
const MAX_FILTER_TESTS: u64 = 50_000_000;
|
||||
|
||||
pub(super) fn advanced_filter(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
const CMD: &str = "data.advancedFilter";
|
||||
let d = s.doc()?;
|
||||
@@ -136,10 +133,6 @@ pub(super) fn advanced_filter(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
rows.push(conds);
|
||||
}
|
||||
}
|
||||
// Each list row is tried against every criteria row, so bound their product too.
|
||||
if u64::from(list.height()) * rows.len() as u64 > MAX_FILTER_TESTS {
|
||||
return Err(bad(CMD, "the list and criteria ranges together are too large"));
|
||||
}
|
||||
let matches_row = |row: u32| -> bool {
|
||||
if rows.is_empty() {
|
||||
return true;
|
||||
|
||||
@@ -61,24 +61,6 @@ enum Key {
|
||||
|
||||
const MAX_CONSOLIDATE_CELLS: u64 = 5_000_000;
|
||||
|
||||
/// The cells a consolidated table writes: every row key meets every column key, plus a label for
|
||||
/// each key on the sides that have labels.
|
||||
fn output_cells(rows: usize, cols: usize, top: bool, left: bool) -> u64 {
|
||||
let (r, c) = (rows as u64, cols as u64);
|
||||
r.saturating_mul(c).saturating_add(if top { c } else { 0 }).saturating_add(if left { r } else { 0 })
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
#[test]
|
||||
fn the_output_limit_counts_label_cells() {
|
||||
// A million row keys and four column keys are 4 million cells, 5 million with the row labels.
|
||||
assert_eq!(output_cells(1_000_000, 4, false, false), 4_000_000);
|
||||
assert_eq!(output_cells(1_000_000, 4, false, true), 5_000_000);
|
||||
assert!(output_cells(1_000_000, 5, false, true) > MAX_CONSOLIDATE_CELLS);
|
||||
assert_eq!(output_cells(3, 2, true, true), 6 + 2 + 3);
|
||||
assert_eq!(output_cells(usize::MAX, usize::MAX, true, true), u64::MAX);
|
||||
}
|
||||
|
||||
pub(super) fn consolidate(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
const CMD: &str = "data.consolidate";
|
||||
let func = str_param(p, "function").unwrap_or("sum").to_string();
|
||||
@@ -107,20 +89,20 @@ pub(super) fn consolidate(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
}
|
||||
None => (d.wb.active_sheet(), d.selection.active),
|
||||
};
|
||||
// Keys in first-seen order, each with its index: a lookup is one probe, not a scan.
|
||||
let mut row_keys: Vec<Key> = Vec::new();
|
||||
let mut col_keys: Vec<Key> = Vec::new();
|
||||
let mut row_index: HashMap<Key, usize> = HashMap::new();
|
||||
let mut col_index: HashMap<Key, usize> = HashMap::new();
|
||||
let mut row_label: HashMap<Key, String> = HashMap::new();
|
||||
let mut col_label: HashMap<Key, String> = HashMap::new();
|
||||
let mut aggs: HashMap<(usize, usize), Agg> = HashMap::new();
|
||||
let mut total: u64 = 0;
|
||||
let index_of = |keys: &mut Vec<Key>, index: &mut HashMap<Key, usize>, k: Key| -> usize {
|
||||
*index.entry(k).or_insert_with_key(|k| {
|
||||
keys.push(k.clone());
|
||||
keys.len() - 1
|
||||
})
|
||||
let index_of = |keys: &mut Vec<Key>, k: Key| -> usize {
|
||||
match keys.iter().position(|x| *x == k) {
|
||||
Some(i) => i,
|
||||
None => {
|
||||
keys.push(k);
|
||||
keys.len() - 1
|
||||
}
|
||||
}
|
||||
};
|
||||
for t in &srcs {
|
||||
let (si, r) = parse_range(&d.wb, t, CMD)?;
|
||||
@@ -149,7 +131,7 @@ pub(super) fn consolidate(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
} else {
|
||||
Key::Pos(col - c0)
|
||||
};
|
||||
cols.push((col, index_of(&mut col_keys, &mut col_index, k)));
|
||||
cols.push((col, index_of(&mut col_keys, k)));
|
||||
}
|
||||
for row in r0..=r.end.row {
|
||||
let k = if left {
|
||||
@@ -163,7 +145,7 @@ pub(super) fn consolidate(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
} else {
|
||||
Key::Pos(row - r0)
|
||||
};
|
||||
let ri = index_of(&mut row_keys, &mut row_index, k);
|
||||
let ri = index_of(&mut row_keys, k);
|
||||
for (col, ci) in &cols {
|
||||
aggs.entry((ri, *ci)).or_default().add(&sh.value(CellRef::new(row, *col)));
|
||||
}
|
||||
@@ -175,9 +157,6 @@ pub(super) fn consolidate(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
{
|
||||
return Err(EngineError::Other("The consolidated data doesn't fit below and right of the destination.".into()));
|
||||
}
|
||||
if output_cells(row_keys.len(), col_keys.len(), top, left) > MAX_CONSOLIDATE_CELLS {
|
||||
return Err(bad(CMD, "the consolidated table would be too large"));
|
||||
}
|
||||
let mut cells: Vec<(CellRef, Value)> = Vec::new();
|
||||
if top {
|
||||
for (ci, k) in col_keys.iter().enumerate() {
|
||||
|
||||
@@ -189,9 +189,6 @@ pub(super) fn scenario_delete(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
Ok(json!({"deleted": removed.name}))
|
||||
}
|
||||
|
||||
/// Most changing-cell-by-scenario cells one Scenario Summary writes.
|
||||
const MAX_SUMMARY_CELLS: usize = 1_000_000;
|
||||
|
||||
pub(super) fn scenario_summary(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
const CMD: &str = "data.scenarioSummary";
|
||||
let d = s.doc()?;
|
||||
@@ -214,24 +211,17 @@ pub(super) fn scenario_summary(s: &mut Session, p: &Json) -> Result<Json> {
|
||||
// Changing cells of every scenario, first-seen order; per scenario its values.
|
||||
let mut changing: Vec<CellRef> = Vec::new();
|
||||
let mut per: Vec<HashMap<CellRef, Value>> = Vec::new();
|
||||
let mut seen: std::collections::HashSet<CellRef> = Default::default();
|
||||
for sc in &scenarios {
|
||||
let (_, cells) = cell_list(&d.wb, &sc.changing, CMD, MAX_SCENARIO_CELLS)?;
|
||||
let mut m = HashMap::new();
|
||||
for (c, v) in cells.iter().zip(sc.values.iter()) {
|
||||
if seen.insert(*c) {
|
||||
if !changing.contains(c) {
|
||||
changing.push(*c);
|
||||
}
|
||||
m.insert(*c, json_value(v, &d.wb));
|
||||
}
|
||||
per.push(m);
|
||||
}
|
||||
// The summary has a cell for every changing cell in every scenario, filled from the current
|
||||
// value where a scenario doesn't set it, so small disjoint scenarios multiply.
|
||||
// Each result cell is a row of the summary too, with a cell per scenario.
|
||||
if changing.len().saturating_add(results.len()).saturating_mul(per.len()) > MAX_SUMMARY_CELLS {
|
||||
return Err(bad(CMD, "the scenarios together change too many cells for a summary"));
|
||||
}
|
||||
let current: Vec<Value> = changing.iter().map(|c| sh.value(*c)).collect();
|
||||
let current_results: Vec<Value> = results.iter().map(|c| sh.value(*c)).collect();
|
||||
let mut trial = Trial::new(d);
|
||||
|
||||
@@ -297,47 +297,3 @@ fn consolidate_by_position_and_label() {
|
||||
assert!(s.execute("data.consolidate", json!({"sources": []})).is_err());
|
||||
assert!(s.execute("data.consolidate", json!({"sources": ["A1"], "function": "median"})).is_err());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn advanced_filter_refuses_a_list_times_criteria_product_over_the_work_limit() {
|
||||
let mut s = s();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["Name"], ["A"]]})).unwrap();
|
||||
s.execute("range.setValues", json!({"range": "H1", "values": [["Name"], ["A"]]})).unwrap();
|
||||
// A sparse million-row list against 9,999 blank criteria rows would be ~10 billion tests.
|
||||
assert!(s.execute("data.advancedFilter", json!({"range": "A2:A1048576", "criteria": "H1:H10000"})).is_err());
|
||||
}
|
||||
|
||||
/// A hundred thousand positional rows take one key lookup each; a tall source and a wide one
|
||||
/// whose keys would meet in a table of 160 million cells are refused before it is built.
|
||||
#[test]
|
||||
fn consolidate_keys_are_indexed_and_the_output_is_bounded() {
|
||||
let mut s = s();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[1], [2]]})).unwrap();
|
||||
let r = s.execute("data.consolidate", json!({"sources": ["A1:A100000"], "destination": "C1"})).unwrap();
|
||||
assert_eq!(r["rows"], 100_000);
|
||||
assert_eq!(n(&s, "Sheet1!C2"), 2.0);
|
||||
let r = s.execute("data.consolidate", json!({"sources": ["A1:A10000", "A1:XFC1"], "destination": "A1"}));
|
||||
assert!(r.unwrap_err().to_string().contains("too large"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn scenario_summary_refuses_a_matrix_over_its_cell_limit() {
|
||||
let mut s = s();
|
||||
// 200 scenarios of 32 disjoint cells: 6,400 changing cells x 200 scenarios = 1.28 million cells.
|
||||
for i in 0..200 {
|
||||
let (a, b) = (i * 32 + 1, i * 32 + 32);
|
||||
s.execute("data.scenarioAdd", json!({"name": format!("S{i}"), "changing": format!("A{a}:A{b}")})).unwrap();
|
||||
}
|
||||
assert!(s.execute("data.scenarioSummary", json!({})).is_err());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn scenario_summary_counts_result_rows_in_its_cell_limit() {
|
||||
let mut s = s();
|
||||
// 900 scenarios of one distinct cell each are 810,000 cells; 255 result rows take it past the limit.
|
||||
for i in 0..900 {
|
||||
s.execute("data.scenarioAdd", json!({"name": format!("S{i}"), "changing": format!("A{}", i + 1)})).unwrap();
|
||||
}
|
||||
let err = s.execute("data.scenarioSummary", json!({"resultCells": "C1:C255"})).unwrap_err().to_string();
|
||||
assert!(err.contains("too many cells"), "{err}");
|
||||
}
|
||||
|
||||
@@ -8,13 +8,8 @@ use gridcraft_model::{Sheet, Workbook};
|
||||
use gridcraft_numfmt::{Formatted, NumberFormat, format_value};
|
||||
|
||||
/// Parsed number formats are cached by code.
|
||||
pub(crate) fn parsed(code: &str) -> NumberFormat {
|
||||
fn parsed(code: &str) -> NumberFormat {
|
||||
static CACHE: std::sync::LazyLock<Mutex<HashMap<String, NumberFormat>>> = std::sync::LazyLock::new(|| Mutex::new(HashMap::new()));
|
||||
// A code of more bytes than 4 × the character limit can't be within it: it shows as General
|
||||
// without being hashed or scanned again for every cell that shares it.
|
||||
if code.len() > 4 * gridcraft_numfmt::MAX_FORMAT_CODE_CHARS {
|
||||
return NumberFormat::parse("General");
|
||||
}
|
||||
let lock = || CACHE.lock().unwrap_or_else(std::sync::PoisonError::into_inner);
|
||||
if let Some(f) = lock().get(code) {
|
||||
return f.clone();
|
||||
@@ -29,12 +24,7 @@ pub(crate) fn parsed(code: &str) -> NumberFormat {
|
||||
f
|
||||
}
|
||||
|
||||
/// The parsed format of `code`; one longer than the character limit is a General format that keeps
|
||||
/// the code it was made from. The engine itself formats through `parsed`, which doesn't copy it.
|
||||
pub fn number_format(code: &str) -> NumberFormat {
|
||||
if code.len() > 4 * gridcraft_numfmt::MAX_FORMAT_CODE_CHARS {
|
||||
return NumberFormat::parse(code);
|
||||
}
|
||||
parsed(code)
|
||||
}
|
||||
|
||||
@@ -238,31 +228,3 @@ pub fn stats(sheet: &Sheet, ranges: &[RangeRef]) -> Stats {
|
||||
}
|
||||
s
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
/// A format code far past Excel's 255 characters shows its values as General, for every cell
|
||||
/// that shares it, without being hashed into the cache each time.
|
||||
#[test]
|
||||
fn an_overlong_code_formats_as_general() {
|
||||
let wb = Workbook::new();
|
||||
let code = format!("0.00{}", "\"x\"".repeat(400_000));
|
||||
for _ in 0..1_000 {
|
||||
assert_eq!(format(&Value::Number(1.5), &code, &wb).text, "1.5");
|
||||
}
|
||||
assert_eq!(format(&Value::Number(1.5), "0.00", &wb).text, "1.50");
|
||||
}
|
||||
|
||||
/// The public `number_format` keeps the code of an overlong format; it is General, but not the
|
||||
/// same format as a real `General`.
|
||||
#[test]
|
||||
fn number_format_keeps_an_overlong_code() {
|
||||
let code = format!("0.00{}", "\"x\"".repeat(2_000));
|
||||
let f = number_format(&code);
|
||||
assert!(f.is_general());
|
||||
assert_eq!(f.code(), code);
|
||||
assert_ne!(f, number_format("General"));
|
||||
}
|
||||
}
|
||||
|
||||
+39
-172
@@ -3,19 +3,12 @@
|
||||
use std::fmt::Write as _;
|
||||
|
||||
use gridcraft_core::{CellRef, RangeRef};
|
||||
use gridcraft_model::{LineIndex, Sheet, Workbook};
|
||||
use gridcraft_model::{Sheet, Workbook};
|
||||
|
||||
use crate::{EngineError, Result};
|
||||
|
||||
mod sylk;
|
||||
|
||||
/// The function library's answers to what the XLSX and XLS readers and the XLSX writer need to know about functions
|
||||
/// (which names are built in, which parameters take a single value).
|
||||
pub const XLSX_FUNCTIONS: gridcraft_xlsx::FunctionInfo = gridcraft_xlsx::FunctionInfo {
|
||||
is_builtin: |name| gridcraft_functions::lookup(name).is_some(),
|
||||
is_scalar_arg: |name, i| gridcraft_functions::lookup(name).is_none_or(|s| s.is_scalar_arg(i)),
|
||||
};
|
||||
|
||||
/// File formats GridCraft reads or writes.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
|
||||
pub enum FileKind {
|
||||
@@ -100,18 +93,14 @@ pub fn open_bytes_with(name: &str, bytes: &[u8], password: Option<&str>, csv_as_
|
||||
/// line-height estimate can differ from Excel's, and shrinking files that already look
|
||||
/// right would be a visible regression. Custom-height rows are left alone entirely.
|
||||
fn fit_imported_rows(wb: &mut Workbook) {
|
||||
// The whole pass shares one budget of merge comparisons: later sheets keep their heights once it is spent.
|
||||
let mut left = crate::cmd::MERGE_FIT_WORK;
|
||||
for si in 0..wb.sheets.len() {
|
||||
let Some(sh) = wb.sheet(si) else { continue };
|
||||
let mut rows: Vec<u32> = sh.cells.iter().map(|(c, _)| c.row).collect();
|
||||
rows.sort_unstable();
|
||||
rows.dedup();
|
||||
// One sweep over the sheet's merges for all of its rows, in ascending order.
|
||||
let mut sweep = crate::cmd::MergeSweep::new(left);
|
||||
for row in rows {
|
||||
let before = wb.sheet(si).and_then(|sh| sh.rows.get(&row)).and_then(|i| i.size);
|
||||
crate::cmd::fit_row_height(wb, si, row, &mut sweep);
|
||||
crate::cmd::auto_row_height(wb, si, row);
|
||||
let Some(before) = before else { continue };
|
||||
let after = wb.sheet(si).and_then(|sh| sh.rows.get(&row)).and_then(|i| i.size);
|
||||
if after.is_none_or(|a| before > a + 0.5)
|
||||
@@ -120,7 +109,6 @@ fn fit_imported_rows(wb: &mut Workbook) {
|
||||
sh.rows.entry(row).or_default().size = Some(before);
|
||||
}
|
||||
}
|
||||
left = sweep.left();
|
||||
}
|
||||
}
|
||||
|
||||
@@ -276,8 +264,7 @@ fn read_plain(name: &str, bytes: &[u8], csv_as_text: bool) -> Result<(Workbook,
|
||||
let kind = FileKind::from_path(name);
|
||||
let sniffed = gridcraft_xlsx::sniff(bytes);
|
||||
if sniffed == gridcraft_xlsx::Format::Xls {
|
||||
let (wb, report) =
|
||||
gridcraft_xlsx::read_xls_with(bytes, XLSX_FUNCTIONS).map_err(|e| EngineError::Other(format!("We can't open '{name}': {e}")))?;
|
||||
let (wb, report) = gridcraft_xlsx::read_xls(bytes).map_err(|e| EngineError::Other(format!("We can't open '{name}': {e}")))?;
|
||||
return Ok((wb, report.warnings));
|
||||
}
|
||||
if sniffed == gridcraft_xlsx::Format::LegacyBinary {
|
||||
@@ -302,13 +289,12 @@ fn read_plain(name: &str, bytes: &[u8], csv_as_text: bool) -> Result<(Workbook,
|
||||
// Text named `.xlsx` is read as text.
|
||||
if sniffed == gridcraft_xlsx::Format::Xlsx || (kind == Some(FileKind::Xlsx) && sniffed != gridcraft_xlsx::Format::Csv) {
|
||||
// Google Sheets' export wrapper is read as the formula inside when GridCraft calculates it (#2871).
|
||||
let (wb, report) = gridcraft_xlsx::read_xlsx_with(bytes, (XLSX_FUNCTIONS, &gridcraft_calc::is_known_function as &dyn Fn(&str) -> bool))
|
||||
.map_err(|e| {
|
||||
EngineError::Other(match (&e, gridcraft_xlsx::oversized_part(bytes)) {
|
||||
(gridcraft_xlsx::IoError::TooLarge(_), Some((part, size))) => part_too_large_message(name, &part, size),
|
||||
_ => format!("We can't open '{name}': {e}"),
|
||||
})
|
||||
})?;
|
||||
let (wb, report) = gridcraft_xlsx::read_xlsx_with(bytes, &gridcraft_calc::is_known_function).map_err(|e| {
|
||||
EngineError::Other(match (&e, gridcraft_xlsx::oversized_part(bytes)) {
|
||||
(gridcraft_xlsx::IoError::TooLarge(_), Some((part, size))) => part_too_large_message(name, &part, size),
|
||||
_ => format!("We can't open '{name}': {e}"),
|
||||
})
|
||||
})?;
|
||||
return Ok((wb, report.warnings));
|
||||
}
|
||||
if let Some(read) = read_html_table(name, kind, bytes, csv_as_text) {
|
||||
@@ -516,7 +502,7 @@ pub fn save_bytes_as(wb: &Workbook, path: &str, kind: FileKind) -> Result<Vec<u8
|
||||
None => named,
|
||||
};
|
||||
let book = std::path::Path::new(path).file_name().and_then(std::ffi::OsStr::to_str).unwrap_or_default();
|
||||
gridcraft_xlsx::write_xlsx_named_with(named.as_ref().unwrap_or(wb), gridcraft_xlsx::XlsxKind::from_path(path), book, XLSX_FUNCTIONS)
|
||||
gridcraft_xlsx::write_xlsx_named(named.as_ref().unwrap_or(wb), gridcraft_xlsx::XlsxKind::from_path(path), book)
|
||||
.map_err(|e| EngineError::Other(e.to_string()))
|
||||
}
|
||||
FileKind::Xls => {
|
||||
@@ -581,7 +567,7 @@ pub fn csv_values_bytes(wb: &Workbook) -> Result<Vec<u8>> {
|
||||
let Some(sh) = wb.sheet(wb.active_sheet()) else { return Ok(Vec::new()) };
|
||||
gridcraft_xlsx::write_csv(sh, wb.csv_delimiter.unwrap_or(b','), |c| match sh.value(c) {
|
||||
gridcraft_core::Value::Number(n) if n.is_finite() => {
|
||||
let shape = crate::display::parsed(wb.styles.get(sh.style_id(c)).num_fmt.as_str()).date_shape(n);
|
||||
let shape = crate::display::number_format(wb.styles.get(sh.style_id(c)).num_fmt.as_str()).date_shape(n);
|
||||
match shape.zip(gridcraft_core::date::datetime_from_serial(wb.date_system, n)) {
|
||||
Some(((date, time), d)) => iso(&d, date, time),
|
||||
None => n.to_string(),
|
||||
@@ -611,24 +597,6 @@ pub(crate) fn esc(s: &str) -> String {
|
||||
s.replace('&', "&").replace('<', "<").replace('>', ">").replace('"', """)
|
||||
}
|
||||
|
||||
/// [`esc`], or `None` once the escaped text would pass `limit` bytes (nothing larger is built).
|
||||
fn esc_within(s: &str, limit: usize) -> Option<String> {
|
||||
let mut out = String::with_capacity(s.len().min(limit));
|
||||
for c in s.chars() {
|
||||
match c {
|
||||
'&' => out.push_str("&"),
|
||||
'<' => out.push_str("<"),
|
||||
'>' => out.push_str(">"),
|
||||
'"' => out.push_str("""),
|
||||
c => out.push(c),
|
||||
}
|
||||
if out.len() > limit {
|
||||
return None;
|
||||
}
|
||||
}
|
||||
Some(out)
|
||||
}
|
||||
|
||||
fn html_cell_css(wb: &Workbook, sh: &Sheet, c: CellRef, dir: CellDir) -> String {
|
||||
let st = wb.styles.get(sh.style_id(c));
|
||||
let mut css = String::new();
|
||||
@@ -704,30 +672,6 @@ fn table_dir(sh: &Sheet) -> &'static str {
|
||||
const CLIPBOARD_HTML_MAX_BYTES: usize = 4 * 1024 * 1024;
|
||||
const CLIPBOARD_HTML_MAX_CELLS: u64 = 10_000;
|
||||
|
||||
/// Most cells the merge index holds: one entry per merged cell inside the exported range.
|
||||
const MERGE_INDEX_MAX_ENTRIES: usize = 1_000_000;
|
||||
|
||||
/// The merge covering each cell of `rect`, as `Sheet::merge_at` would find it (the first in the
|
||||
/// sheet's list), built once so a table lookup is one probe instead of a scan of every merge.
|
||||
/// `None` when overlapping merges would cost more than a few visits per cell of `rect`, or merged
|
||||
/// cells fill more than [`MERGE_INDEX_MAX_ENTRIES`]: a hostile file's merge list can't multiply the
|
||||
/// work or the memory of a bounded export.
|
||||
fn merge_index(sh: &Sheet, rect: RangeRef) -> Option<std::collections::HashMap<CellRef, RangeRef>> {
|
||||
let mut map = std::collections::HashMap::new();
|
||||
let mut left = rect.count().saturating_mul(4);
|
||||
for m in &sh.merges {
|
||||
let Some(part) = m.intersection(&rect) else { continue };
|
||||
left = left.checked_sub(part.count())?;
|
||||
for c in part.iter() {
|
||||
map.entry(c).or_insert(*m);
|
||||
}
|
||||
if map.len() > MERGE_INDEX_MAX_ENTRIES {
|
||||
return None;
|
||||
}
|
||||
}
|
||||
Some(map)
|
||||
}
|
||||
|
||||
/// An HTML fragment for the same range as `plain_text`, previously produced by
|
||||
/// `display::range_text`. Large copies keep plain text only; never return a partial table.
|
||||
pub(crate) fn range_html(wb: &Workbook, sheet: usize, r: RangeRef, plain_text: &str) -> Option<String> {
|
||||
@@ -736,7 +680,6 @@ pub(crate) fn range_html(wb: &Workbook, sheet: usize, r: RangeRef, plain_text: &
|
||||
return None;
|
||||
}
|
||||
let sh = wb.sheet(sheet)?;
|
||||
let merges = merge_index(sh, r)?;
|
||||
let mut out = ClipboardHtml(String::new());
|
||||
write!(out, "<table{} style=\"border-collapse:collapse;font-family:Calibri,Carlito,Arial,sans-serif;font-size:11pt\">", table_dir(sh)).ok()?;
|
||||
for row in r.start.row..=r.end.row {
|
||||
@@ -746,7 +689,7 @@ pub(crate) fn range_html(wb: &Workbook, sheet: usize, r: RangeRef, plain_text: &
|
||||
out.write_str("<tr>").ok()?;
|
||||
for col in r.start.col..=r.end.col {
|
||||
let c = CellRef::new(row, col);
|
||||
let span = if let Some(merged) = merges.get(&c).and_then(|m| m.intersection(&r)) {
|
||||
let span = if let Some(merged) = sh.merges.iter().find(|m| m.contains(c)).and_then(|m| m.intersection(&r)) {
|
||||
let first_row = (merged.start.row..=merged.end.row).find(|&row| !sh.is_row_hidden(row))?;
|
||||
if c != CellRef::new(first_row, merged.start.col) {
|
||||
continue;
|
||||
@@ -809,14 +752,6 @@ const HTML_EXPORT_MAX_ROWS: u32 = 100_001;
|
||||
/// tall, wide one) would otherwise be gigabytes of `<td></td>` before finishing (#2387).
|
||||
const HTML_EXPORT_MAX_CELLS: u64 = 4_000_000;
|
||||
|
||||
/// Most bytes of one web page. Every cell may hold long text (shared across many cells in the
|
||||
/// file), so the cell count alone doesn't bound the page; past this the export fails instead.
|
||||
const HTML_EXPORT_MAX_BYTES: usize = 256 * 1024 * 1024;
|
||||
|
||||
fn html_too_big() -> EngineError {
|
||||
EngineError::Other(format!("the web page would be over {} MiB. Save as .xlsx or CSV instead.", HTML_EXPORT_MAX_BYTES / (1024 * 1024)))
|
||||
}
|
||||
|
||||
/// A sheet as a standalone HTML table (Save as Web Page). A sheet with pictures gets the sheet's
|
||||
/// own column widths and row heights so each picture and chart (inline SVG) can sit at its anchor.
|
||||
pub fn to_html(wb: &Workbook, sheet: usize) -> Result<String> {
|
||||
@@ -859,24 +794,6 @@ pub fn to_html(wb: &Workbook, sheet: usize) -> Result<String> {
|
||||
r.width()
|
||||
)));
|
||||
}
|
||||
let merges = merge_index(sh, r)
|
||||
.ok_or_else(|| EngineError::Other("the sheet's merged cells overlap too much, or cover too many cells, for a web page.".into()))?;
|
||||
// Visible rows and columns of the table before each line, so a merge's span is counted
|
||||
// in constant time whatever the file says its size is.
|
||||
let visible_before = |hidden: &dyn Fn(u32) -> bool, from: u32, to: u32| {
|
||||
let mut v = Vec::with_capacity((to - from) as usize + 2);
|
||||
v.push(0u32);
|
||||
for line in from..=to {
|
||||
v.push(v.last().copied().unwrap_or(0) + u32::from(!hidden(line)));
|
||||
}
|
||||
v
|
||||
};
|
||||
let rows_before = visible_before(&|y| sh.is_row_hidden(y), r.start.row, r.end.row);
|
||||
let cols_before = visible_before(&|x| sh.is_col_hidden(x), r.start.col, r.end.col);
|
||||
let visible = |before: &[u32], from: u32, a: u32, b: u32| {
|
||||
let at = |line: u32| before.get((line - from) as usize).copied().unwrap_or(0);
|
||||
at(b + 1) - at(a)
|
||||
};
|
||||
if fixed {
|
||||
out.push_str("<colgroup>");
|
||||
for col in (r.start.col..=r.end.col).filter(|&c| !sh.is_col_hidden(c)) {
|
||||
@@ -895,31 +812,27 @@ pub fn to_html(wb: &Workbook, sheet: usize) -> Result<String> {
|
||||
}
|
||||
for col in r.start.col..=r.end.col {
|
||||
let c = CellRef::new(row, col);
|
||||
let merge = merges.get(&c);
|
||||
if merge.is_some_and(|m| m.start != c) || (fixed && sh.is_col_hidden(col)) {
|
||||
if sh.merges.iter().any(|m| m.contains(c) && m.start != c) || (fixed && sh.is_col_hidden(col)) {
|
||||
continue;
|
||||
}
|
||||
let text = crate::display::cell_text(wb, sh, c);
|
||||
let dir = CellDir::of(sh, c, &text);
|
||||
let css = html_cell_css(wb, sh, c, dir);
|
||||
let span = merge
|
||||
let span = sh
|
||||
.merges
|
||||
.iter()
|
||||
.find(|m| m.start == c)
|
||||
.map(|m| {
|
||||
if fixed {
|
||||
// Spans count the table's visible lines: the merge is clipped to the table.
|
||||
let rows = visible(&rows_before, r.start.row, m.start.row, m.end.row.min(r.end.row));
|
||||
let cols = visible(&cols_before, r.start.col, m.start.col, m.end.col.min(r.end.col));
|
||||
let rows = (m.start.row..=m.end.row).filter(|&y| !sh.is_row_hidden(y)).count();
|
||||
let cols = (m.start.col..=m.end.col).filter(|&x| !sh.is_col_hidden(x)).count();
|
||||
format!(" rowspan=\"{rows}\" colspan=\"{cols}\"")
|
||||
} else {
|
||||
format!(" rowspan=\"{}\" colspan=\"{}\"", m.height(), m.width())
|
||||
}
|
||||
})
|
||||
.unwrap_or_default();
|
||||
// The text is escaped only as far as the page still has room for it.
|
||||
let Some(escaped) = esc_within(&text, HTML_EXPORT_MAX_BYTES.saturating_sub(out.len())) else { return Err(html_too_big()) };
|
||||
let _ = write!(out, "<td style=\"{css}\"{}{span}>{escaped}</td>", dir.attr());
|
||||
if out.len() > HTML_EXPORT_MAX_BYTES {
|
||||
return Err(html_too_big());
|
||||
}
|
||||
let _ = write!(out, "<td style=\"{css}\"{}{span}>{}</td>", dir.attr(), esc(&text));
|
||||
}
|
||||
out.push_str("</tr>");
|
||||
}
|
||||
@@ -927,48 +840,34 @@ pub fn to_html(wb: &Workbook, sheet: usize) -> Result<String> {
|
||||
out.push_str("</table>");
|
||||
if fixed {
|
||||
// Sheet units are CSS pixels; the table starts at the used range, not at A1.
|
||||
// Indexed positions: each anchor would otherwise sum every custom row and column before it.
|
||||
let ri = LineIndex::new(&sh.rows, sh.default_row_height, gridcraft_core::MAX_ROWS);
|
||||
let ci = LineIndex::new(&sh.cols, sh.default_col_width, gridcraft_core::MAX_COLS);
|
||||
let (left, top) = used.map_or((0.0, 0.0), |r| (ci.start(r.start.col), ri.start(r.start.row)));
|
||||
let (left, top) = used.map_or((0.0, 0.0), |r| (sh.col_left(r.start.col), sh.row_top(r.start.row)));
|
||||
let place = |a: &gridcraft_model::Anchor| {
|
||||
format!(
|
||||
"position:absolute;left:{}px;top:{}px;width:{}px;height:{}px",
|
||||
ci.start(a.cell.col) - left + f64::from(a.dx),
|
||||
ri.start(a.cell.row) - top + f64::from(a.dy),
|
||||
sh.col_left(a.cell.col) - left + f64::from(a.dx),
|
||||
sh.row_top(a.cell.row) - top + f64::from(a.dy),
|
||||
a.width.max(0.0),
|
||||
a.height.max(0.0)
|
||||
)
|
||||
};
|
||||
for im in &sh.images {
|
||||
// The base64 text is a third longer than the picture: refused before it is built.
|
||||
if out.len().saturating_add(im.data.len().div_ceil(3).saturating_mul(4)) > HTML_EXPORT_MAX_BYTES {
|
||||
return Err(html_too_big());
|
||||
}
|
||||
// The attributes are escaped only as far as the page has room for them.
|
||||
let room = HTML_EXPORT_MAX_BYTES.saturating_sub(out.len());
|
||||
let (Some(mime), Some(alt)) = (esc_within(&im.mime, room), esc_within(&im.alt, room)) else { return Err(html_too_big()) };
|
||||
let _ = write!(out, "<img src=\"data:{mime};base64,{}\" alt=\"{alt}\" style=\"{}\">", base64_encode(&im.data), place(&im.anchor));
|
||||
if out.len() > HTML_EXPORT_MAX_BYTES {
|
||||
return Err(html_too_big());
|
||||
}
|
||||
let _ = write!(
|
||||
out,
|
||||
"<img src=\"data:{};base64,{}\" alt=\"{}\" style=\"{}\">",
|
||||
esc(&im.mime),
|
||||
base64_encode(&im.data),
|
||||
esc(&im.alt),
|
||||
place(&im.anchor)
|
||||
);
|
||||
}
|
||||
let mut svg_budget = crate::chart_svg::MAX_PAGE_SVG_BYTES;
|
||||
for ch in &sh.charts {
|
||||
if let Some(svg) = crate::chart_svg::chart_svg(wb, sheet, ch, &place(&ch.anchor), svg_budget)? {
|
||||
svg_budget -= svg.len();
|
||||
if let Some(svg) = crate::chart_svg::chart_svg(wb, sheet, ch, &place(&ch.anchor)) {
|
||||
out.push_str(&svg);
|
||||
if out.len() > HTML_EXPORT_MAX_BYTES {
|
||||
return Err(html_too_big());
|
||||
}
|
||||
}
|
||||
}
|
||||
out.push_str("</div>");
|
||||
}
|
||||
out.push_str("</body></html>");
|
||||
if out.len() > HTML_EXPORT_MAX_BYTES {
|
||||
return Err(html_too_big());
|
||||
}
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
@@ -980,12 +879,12 @@ thread_local! {
|
||||
/// While one is alive, [`read_file`] and [`write_file`] on this thread refuse every path: a
|
||||
/// stored script's steps run inside one (#2178), so no command reaches the file system from a
|
||||
/// script whatever its parameters are called. Dropping it (also on unwind) lifts the refusal.
|
||||
pub struct ScriptSandbox(std::marker::PhantomData<*const ()>);
|
||||
pub(crate) struct ScriptSandbox(());
|
||||
|
||||
impl ScriptSandbox {
|
||||
pub fn enter() -> Self {
|
||||
pub(crate) fn enter() -> Self {
|
||||
let _ = SCRIPT_SANDBOX.try_with(|c| c.set(c.get().saturating_add(1)));
|
||||
ScriptSandbox(std::marker::PhantomData)
|
||||
ScriptSandbox(())
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1013,12 +912,6 @@ const MAX_READ_BYTES: u64 = 2 * 1024 * 1024 * 1024;
|
||||
pub fn read_file(path: &str) -> Result<Vec<u8>> {
|
||||
use std::io::Read;
|
||||
script_sandbox_check(path)?;
|
||||
// Checked before opening: opening a FIFO nobody writes to blocks the calling (GUI) thread.
|
||||
// shortcut: a FIFO swapped in between this check and the open still blocks; close that with
|
||||
// O_NONBLOCK when the crate takes a libc dependency.
|
||||
if !std::fs::metadata(path).map_err(|e| EngineError::Other(format!("{path}: {e}")))?.is_file() {
|
||||
return Err(EngineError::Other(format!("{path}: only regular files can be opened")));
|
||||
}
|
||||
let f = std::fs::File::open(path).map_err(|e| EngineError::Other(format!("{path}: {e}")))?;
|
||||
// A symlink to an endless device reports length 0 but streams forever, so metadata
|
||||
// alone never bounds the read: require a real file and cap the bytes actually taken.
|
||||
@@ -1050,37 +943,18 @@ pub fn write_file(path: &str, bytes: &[u8]) -> Result<()> {
|
||||
// link's target; and it inherits the destination's access restrictions, so saving a
|
||||
// private workbook doesn't widen who can read it.
|
||||
let tmp = format!("{path}.sctmp");
|
||||
let prior_meta = std::fs::metadata(path).ok();
|
||||
let prior = prior_meta.as_ref().map(std::fs::Metadata::permissions);
|
||||
let prior = std::fs::metadata(path).ok().map(|m| m.permissions());
|
||||
let mut opts = std::fs::OpenOptions::new();
|
||||
opts.write(true).create_new(true);
|
||||
// Over an existing file the replacement starts owner-only, whatever the umask: its group is
|
||||
// still the saving user's primary one until it is changed below, and a member of that group
|
||||
// must not be able to open it (and keep the descriptor) before then.
|
||||
#[cfg(unix)]
|
||||
if prior.is_some() {
|
||||
use std::os::unix::fs::OpenOptionsExt;
|
||||
opts.mode(0o600);
|
||||
if let Some(p) = &prior {
|
||||
use std::os::unix::fs::{OpenOptionsExt, PermissionsExt};
|
||||
opts.mode(p.mode());
|
||||
}
|
||||
let mut f = opts.open(&tmp).map_err(|e| {
|
||||
let _ = std::fs::remove_file(&tmp);
|
||||
EngineError::Other(format!("{path}: {e}"))
|
||||
})?;
|
||||
// The replacement belongs to the saving user's primary group: restore the old file's group and
|
||||
// then its mode. When the group can't be restored its access is dropped rather than handed to a
|
||||
// different group; failing to read that back is a failed save.
|
||||
#[cfg(unix)]
|
||||
if let Some(m) = &prior_meta {
|
||||
use std::os::unix::fs::{MetadataExt, PermissionsExt};
|
||||
let _ = std::os::unix::fs::fchown(&f, None, Some(m.gid()));
|
||||
let group_kept = f.metadata().map(|now| now.gid() == m.gid()).map_err(|e| {
|
||||
let _ = std::fs::remove_file(&tmp);
|
||||
EngineError::Other(format!("{path}: {e}"))
|
||||
})?;
|
||||
// Best effort: a file system that can't change modes keeps the owner-only one. Without the
|
||||
// group, the old owner and other bits are kept and the group, setuid and setgid bits dropped.
|
||||
let _ = f.set_permissions(std::fs::Permissions::from_mode(restored_mode(m.mode(), group_kept)));
|
||||
}
|
||||
#[cfg(not(unix))]
|
||||
if let Some(p) = &prior {
|
||||
let _ = f.set_permissions(p.clone());
|
||||
@@ -1097,13 +971,6 @@ pub fn write_file(path: &str, bytes: &[u8]) -> Result<()> {
|
||||
})
|
||||
}
|
||||
|
||||
/// The mode a saved file gets from the file it replaces (`old`): all of it when its group was
|
||||
/// restored; without the group, the owner and other bits only.
|
||||
#[cfg(unix)]
|
||||
fn restored_mode(old: u32, group_kept: bool) -> u32 {
|
||||
if group_kept { old & 0o7777 } else { old & 0o707 }
|
||||
}
|
||||
|
||||
#[cfg(target_arch = "wasm32")]
|
||||
pub fn write_file(path: &str, _bytes: &[u8]) -> Result<()> {
|
||||
script_sandbox_check(path)?;
|
||||
|
||||
@@ -51,14 +51,6 @@ fn save_keeps_the_destination_access_restrictions() {
|
||||
write_file(dest.to_str().unwrap(), b"v2").unwrap();
|
||||
let mode = std::fs::metadata(&dest).unwrap().permissions().mode() & 0o777;
|
||||
assert_eq!(mode, 0o600);
|
||||
// The mode is set outright from the old file's, not through the umask: group and other access
|
||||
// it had stay as they were.
|
||||
for old in [0o640, 0o664] {
|
||||
std::fs::set_permissions(&dest, std::fs::Permissions::from_mode(old)).unwrap();
|
||||
write_file(dest.to_str().unwrap(), b"v3").unwrap();
|
||||
assert_eq!(std::fs::metadata(&dest).unwrap().permissions().mode() & 0o777, old);
|
||||
}
|
||||
assert!(!dir.join("private.xlsx.sctmp").exists());
|
||||
// And a brand-new file is created with the restrictive, not permissive, default.
|
||||
let fresh = dir.join("new.xlsx");
|
||||
write_file(fresh.to_str().unwrap(), b"v").unwrap();
|
||||
@@ -67,17 +59,6 @@ fn save_keeps_the_destination_access_restrictions() {
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
}
|
||||
|
||||
/// Without the old group, a saved file keeps the old owner and other bits and nothing for a group.
|
||||
#[cfg(not(target_arch = "wasm32"))]
|
||||
#[cfg(unix)]
|
||||
#[test]
|
||||
fn restored_mode_drops_the_group_and_special_bits_when_the_group_is_lost() {
|
||||
assert_eq!(super::restored_mode(0o100_664, true), 0o664);
|
||||
assert_eq!(super::restored_mode(0o100_664, false), 0o604);
|
||||
assert_eq!(super::restored_mode(0o102_444, false), 0o404, "an original read-only file stays read-only, setgid is dropped");
|
||||
assert_eq!(super::restored_mode(0o104_755, false), 0o705, "setuid is dropped");
|
||||
}
|
||||
|
||||
#[cfg(not(target_arch = "wasm32"))]
|
||||
#[test]
|
||||
fn open_refuses_directories_and_unknowable_lengths() {
|
||||
@@ -370,19 +351,6 @@ fn script_sandbox_refuses_file_access_until_dropped() {
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
}
|
||||
|
||||
/// The guard's drop lowers the counter of the thread it runs on, so it must not move between threads:
|
||||
/// naming the trait's method is ambiguous (a compile error) if `ScriptSandbox` is `Send`.
|
||||
#[test]
|
||||
fn script_sandbox_is_not_send() {
|
||||
trait AmbiguousIfSend<A> {
|
||||
fn check() {}
|
||||
}
|
||||
impl<T: ?Sized> AmbiguousIfSend<()> for T {}
|
||||
struct IsSend;
|
||||
impl<T: ?Sized + Send> AmbiguousIfSend<IsSend> for T {}
|
||||
<ScriptSandbox as AmbiguousIfSend<_>>::check();
|
||||
}
|
||||
|
||||
/// A minimal compound (CFB/OLE2) file — FAT in sector 0, directory in sector 1 — holding a
|
||||
/// root entry and the given stream names.
|
||||
fn cfb(streams: &[&str]) -> Vec<u8> {
|
||||
@@ -793,91 +761,3 @@ fn sylk_survives_huge_relative_offsets_and_width_floods() {
|
||||
let (wb, _) = open_bytes("widths.slk", &flood).unwrap();
|
||||
assert_eq!(wb.sheet(0).unwrap().value(CellRef::parse("A1").unwrap()), gridcraft_core::Value::Number(1.0));
|
||||
}
|
||||
|
||||
/// A FIFO with no writer would block an open forever: it is refused before it is opened.
|
||||
#[cfg(unix)]
|
||||
#[test]
|
||||
fn reading_a_fifo_is_refused_without_blocking() {
|
||||
let dir = std::env::temp_dir().join(format!("gridcraft-fifo-{}", std::process::id()));
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
std::fs::create_dir_all(&dir).unwrap();
|
||||
let fifo = dir.join("pipe.xlsx");
|
||||
if !std::process::Command::new("mkfifo").arg(&fifo).status().is_ok_and(|s| s.success()) {
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
return; // no mkfifo here
|
||||
}
|
||||
let err = read_file(fifo.to_str().unwrap()).unwrap_err().to_string();
|
||||
assert!(err.contains("only regular files"), "{err}");
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
}
|
||||
|
||||
/// Merges far from what is copied or exported, in a hostile number, are indexed once instead of
|
||||
/// being scanned for every cell; a merge running past the table is clipped to it.
|
||||
#[test]
|
||||
fn html_merge_lookups_and_spans_stay_bounded() {
|
||||
let mut wb = book_with_picture("A1", (0, 0));
|
||||
let sh = wb.sheet_mut(0).unwrap();
|
||||
for i in 0..200_000u32 {
|
||||
sh.merges.push(RangeRef::new(CellRef::new(500_000 + i * 2, 5), CellRef::new(500_000 + i * 2, 6)));
|
||||
}
|
||||
sh.merges.push(RangeRef::new(CellRef::new(0, 0), CellRef::new(gridcraft_core::MAX_ROWS - 1, 1)));
|
||||
sh.set_value(CellRef::new(2, 1), gridcraft_core::Value::Number(2.0));
|
||||
let html = copied(&wb, "A1:C50");
|
||||
assert!(html.contains("rowspan=\"50\" colspan=\"2\""), "{html}");
|
||||
let page = to_html(&wb, 0).unwrap();
|
||||
// The table is A1:B3 (the used range): the whole-height merge spans its three rows only.
|
||||
assert!(page.contains("rowspan=\"3\" colspan=\"2\""), "{page}");
|
||||
// Merges stacked over one another past a few per cell are refused, not walked.
|
||||
let sh = wb.sheet_mut(0).unwrap();
|
||||
sh.merges = vec![RangeRef::new(CellRef::new(0, 0), CellRef::new(2, 1)); 100];
|
||||
assert!(to_html(&wb, 0).is_err());
|
||||
assert!(range_html(&wb, 0, RangeRef::parse("A1:B3").unwrap(), "").is_none());
|
||||
}
|
||||
|
||||
/// Thousands of pictures over thousands of custom rows: each is placed through a row index.
|
||||
#[test]
|
||||
fn html_export_places_many_pictures_over_many_custom_rows() {
|
||||
let mut wb = book_with_picture("A1", (0, 0));
|
||||
let sh = wb.sheet_mut(0).unwrap();
|
||||
for row in 0..50_000 {
|
||||
sh.rows.insert(row, gridcraft_model::LineInfo { size: Some(30.0), ..Default::default() });
|
||||
}
|
||||
let image = sh.images[0].clone();
|
||||
for i in 0..5_000u32 {
|
||||
let mut im = image.clone();
|
||||
im.anchor.cell = CellRef::new(40_000 + i, 0);
|
||||
sh.images.push(im);
|
||||
}
|
||||
let page = to_html(&wb, 0).unwrap();
|
||||
assert!(page.contains(&format!("top:{}px", 30 * 44_999)), "last picture placed by the index");
|
||||
}
|
||||
|
||||
/// One long text shown by many cells: the page fails at its byte limit instead of growing on.
|
||||
#[test]
|
||||
fn html_export_refuses_a_page_past_its_byte_limit() {
|
||||
let mut wb = Workbook::new();
|
||||
let sh = wb.sheet_mut(0).unwrap();
|
||||
let text = gridcraft_core::Value::text("x".repeat(32_000));
|
||||
for row in 0..10_000 {
|
||||
sh.set_value(CellRef::new(row, 0), text.clone());
|
||||
}
|
||||
let err = to_html(&wb, 0).unwrap_err();
|
||||
assert!(err.to_string().contains("MiB"), "{err}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn escaping_stops_at_the_room_left() {
|
||||
assert_eq!(esc_within("a&b<c>\"d", 100).as_deref(), Some(esc("a&b<c>\"d").as_str()));
|
||||
assert_eq!(esc_within("&&&", 14), None, "three `&` are 15 bytes");
|
||||
assert_eq!(esc_within("&&&", 15).as_deref(), Some("&&&"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn the_merge_index_holds_a_bounded_number_of_cells() {
|
||||
let mut sh = Sheet::new("S");
|
||||
let rect = RangeRef::new(CellRef::new(0, 0), CellRef::new(1_100_000, 0));
|
||||
sh.merges.push(RangeRef::new(CellRef::new(0, 0), CellRef::new(999_999, 0)));
|
||||
assert_eq!(merge_index(&sh, rect).map(|m| m.len()), Some(1_000_000));
|
||||
sh.merges = vec![RangeRef::new(CellRef::new(0, 0), CellRef::new(1_000_000, 0))];
|
||||
assert!(merge_index(&sh, rect).is_none());
|
||||
}
|
||||
|
||||
@@ -237,22 +237,6 @@ const JOURNAL_MAX_ENTRIES: usize = 10_000;
|
||||
pub(crate) const JOURNAL_MAX_BYTES: usize = 64 * 1024 * 1024;
|
||||
const UNDO_BYTES_LIMIT: usize = 128 * 1024 * 1024;
|
||||
|
||||
/// Comment and hyperlink text of a sheet: a changed sheet's snapshot keeps its own copy of these
|
||||
/// strings (an edit deep-copies them), so they count against the history budget with the cells.
|
||||
fn sheet_text_bytes(sheet: &gridcraft_model::Sheet) -> usize {
|
||||
let comments = sheet.comments.values().fold(0usize, |sum, c| {
|
||||
// The vectors' own storage counts too: an empty reply still takes an element.
|
||||
let replies = c.replies.iter().fold(0usize, |n, (a, t)| n.saturating_add(a.len()).saturating_add(t.len()));
|
||||
let replies = replies.saturating_add(c.replies.capacity().saturating_mul(std::mem::size_of::<(String, String)>()));
|
||||
let dates = c.dates.iter().fold(0usize, |n, d| n.saturating_add(d.len()));
|
||||
let dates = dates.saturating_add(c.dates.capacity().saturating_mul(std::mem::size_of::<String>()));
|
||||
sum.saturating_add(std::mem::size_of::<gridcraft_model::Comment>() + c.author.len() + c.text.len() + replies + dates)
|
||||
});
|
||||
sheet.hyperlinks.values().fold(comments, |sum, h| {
|
||||
sum.saturating_add(std::mem::size_of::<gridcraft_model::Hyperlink>() + h.target.len() + h.tooltip.as_ref().map_or(0, String::len))
|
||||
})
|
||||
}
|
||||
|
||||
fn history_retained_bytes(older: &Workbook, newer: &Workbook) -> usize {
|
||||
let mut bytes = std::mem::size_of::<Workbook>();
|
||||
for (index, sheet) in older.sheets.iter().enumerate() {
|
||||
@@ -264,7 +248,6 @@ fn history_retained_bytes(older: &Workbook, newer: &Workbook) -> usize {
|
||||
let empty = gridcraft_model::CellStore::new();
|
||||
let next = newer.sheets.get(index).map_or(&empty, |next| &next.cells);
|
||||
bytes = bytes.saturating_add(sheet.cells.retained_bytes_against(next));
|
||||
bytes = bytes.saturating_add(sheet_text_bytes(sheet));
|
||||
}
|
||||
// Snapshots share the macro list until a change to it; a snapshot with its own list keeps it.
|
||||
if !older.macros.shares(&newer.macros) {
|
||||
|
||||
@@ -440,7 +440,7 @@ pub fn infer_input(wb: &Workbook, sheet: usize, r: RangeRef) -> NamedInput {
|
||||
Some(match v {
|
||||
Value::Empty => return None,
|
||||
Value::Text(t) if t.is_empty() => return None,
|
||||
Value::Number(_) if crate::display::parsed(wb.styles.get(sh.style_id(c)).num_fmt.as_str()).is_date() => ColType::Date,
|
||||
Value::Number(_) if crate::display::number_format(wb.styles.get(sh.style_id(c)).num_fmt.as_str()).is_date() => ColType::Date,
|
||||
Value::Number(_) => ColType::Number,
|
||||
Value::Text(_) => ColType::Text,
|
||||
Value::Bool(_) => ColType::Bool,
|
||||
|
||||
@@ -335,7 +335,7 @@ fn check_input(wb: &Workbook, input: &gridcraft_model::NamedInput, place: Place)
|
||||
gridcraft_model::ColType::Number => v.is_number(),
|
||||
gridcraft_model::ColType::Text => v.is_text(),
|
||||
gridcraft_model::ColType::Bool => matches!(v, Value::Bool(_)),
|
||||
gridcraft_model::ColType::Date => v.is_number() && crate::display::parsed(&date()).is_date(),
|
||||
gridcraft_model::ColType::Date => v.is_number() && crate::display::number_format(&date()).is_date(),
|
||||
};
|
||||
if !ok {
|
||||
wrong.push((c, holds(&v, ty)));
|
||||
@@ -1281,18 +1281,9 @@ fn content_differs(b: Option<&Cell>, a: Option<&Cell>) -> bool {
|
||||
/// Objects added, changed and removed between two versions of a sheet.
|
||||
fn objects(b: &Sheet, a: &Sheet) -> (usize, usize, usize) {
|
||||
fn count<T: PartialEq>(b: &[T], a: &[T], id: impl Fn(&T) -> u32) -> (usize, usize, usize) {
|
||||
// Objects grouped by id, so each lookup is a map probe instead of a scan of the list.
|
||||
fn group<'a, T>(v: &'a [T], id: &impl Fn(&T) -> u32) -> std::collections::HashMap<u32, Vec<&'a T>> {
|
||||
let mut m: std::collections::HashMap<u32, Vec<&T>> = std::collections::HashMap::new();
|
||||
for x in v {
|
||||
m.entry(id(x)).or_default().push(x);
|
||||
}
|
||||
m
|
||||
}
|
||||
let (gb, ga) = (group(b, &id), group(a, &id));
|
||||
let added = a.iter().filter(|x| !gb.contains_key(&id(x))).count();
|
||||
let removed = b.iter().filter(|x| !ga.contains_key(&id(x))).count();
|
||||
let changed = b.iter().filter(|x| ga.get(&id(x)).is_some_and(|ys| ys.iter().any(|y| *y != *x))).count();
|
||||
let added = a.iter().filter(|x| !b.iter().any(|y| id(y) == id(x))).count();
|
||||
let removed = b.iter().filter(|x| !a.iter().any(|y| id(y) == id(x))).count();
|
||||
let changed = b.iter().filter(|x| a.iter().any(|y| id(y) == id(x) && y != *x)).count();
|
||||
(added, changed, removed)
|
||||
}
|
||||
let parts = [
|
||||
|
||||
@@ -182,7 +182,7 @@ pub fn read_source(wb: &Workbook, text: &str, default_sheet: usize) -> Result<So
|
||||
for (n, a) in numeric.iter_mut().zip(any.iter()) {
|
||||
*n = *n && *a;
|
||||
}
|
||||
let dates = numeric.iter().zip(formats.iter()).map(|(n, f)| *n && crate::display::parsed(f).is_date()).collect();
|
||||
let dates = numeric.iter().zip(formats.iter()).map(|(n, f)| *n && crate::display::number_format(f).is_date()).collect();
|
||||
Ok(Source { sheet: si, range: r, headers, formats, numeric, dates, rows })
|
||||
}
|
||||
|
||||
|
||||
@@ -57,10 +57,6 @@ fn rows_ref(sheet: &str, origin: CellRef, col: u32, rows: &[u32]) -> Option<Stri
|
||||
/// Most category rows a PivotChart plots.
|
||||
const MAX_CHART_ROWS: usize = 10_000;
|
||||
|
||||
/// Most series a PivotChart gets: the chart renderer reads at most 255 series, and each one owns
|
||||
/// reference text as long as its category rows, copied into every linked chart.
|
||||
const MAX_CHART_SERIES: usize = 255;
|
||||
|
||||
/// A PivotChart's series for a report written at `origin`: categories are the innermost row
|
||||
/// items, one series per item column (column items × value fields); totals are left out.
|
||||
pub fn chart_series(sheet: &str, origin: CellRef, out: &Output) -> Vec<Series> {
|
||||
@@ -68,7 +64,6 @@ pub fn chart_series(sheet: &str, origin: CellRef, out: &Output) -> Vec<Series> {
|
||||
let categories = out.label_col.and_then(|c| rows_ref(sheet, origin, c, &rows));
|
||||
out.chart_cols
|
||||
.iter()
|
||||
.take(MAX_CHART_SERIES)
|
||||
.filter_map(|(c, name)| {
|
||||
Some(Series {
|
||||
name: Some(format!("\"{}\"", name.replace('"', "\"\""))),
|
||||
@@ -111,22 +106,3 @@ pub fn clear_output(wb: &mut Workbook, sheet: usize, idx: usize) {
|
||||
p.last_range = None;
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn a_pivot_chart_gets_at_most_the_series_the_renderer_reads() {
|
||||
let out = Output {
|
||||
grid: vec![],
|
||||
row_kinds: vec![],
|
||||
body_offset: 0,
|
||||
width: 1000,
|
||||
chart_rows: vec![1, 2, 3],
|
||||
chart_cols: (1..1000).map(|c| (c, format!("S{c}"))).collect(),
|
||||
label_col: Some(0),
|
||||
};
|
||||
assert_eq!(chart_series("Sheet1", CellRef::new(0, 0), &out).len(), MAX_CHART_SERIES);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,7 +1,5 @@
|
||||
use super::items::{Acc, FieldItems, build_tree, collect_siblings, is_collapsed, item_label};
|
||||
use super::output::{
|
||||
AggMap, CellKind, ColKind, ColSpec, Ctx, Grid, Line, LineKind, Output, RowKind, ValSrc, caption, data_cols_exceed_sheet, flatten_cols,
|
||||
};
|
||||
use super::output::{AggMap, CellKind, ColKind, ColSpec, Ctx, Grid, Line, LineKind, Output, RowKind, ValSrc, data_cols_exceed_sheet, flatten_cols};
|
||||
use super::{
|
||||
CalcSet, CellError, CellRef, HashMap, HashSet, MAX_AXIS_FIELDS, MAX_COLS, MAX_VALUE_FIELDS, PivotDateGroup, PivotField, PivotFilter, PivotFunc,
|
||||
PivotLayout, PivotShowAs, PivotTable, PivotValue, Source, Value, Workbook, read_source,
|
||||
@@ -96,10 +94,6 @@ struct AggregateData {
|
||||
col_leaves: HashSet<Vec<u32>>,
|
||||
}
|
||||
|
||||
/// Most accumulators (groups times value columns) one report may hold: twice what the 5 million
|
||||
/// group limit allowed at one value field.
|
||||
const MAX_ACCUMULATORS: usize = 16_000_000;
|
||||
|
||||
fn aggregate_record(
|
||||
input: &AggregateInput<'_>,
|
||||
data: &mut AggregateData,
|
||||
@@ -158,9 +152,7 @@ fn aggregate_record(
|
||||
if !input.cols.is_empty() && data_cols_exceed_sheet(data.col_leaves.len(), input.pt.grand_totals_rows, input.vals.len()) {
|
||||
return Err("The PivotTable report will not fit on the sheet.".into());
|
||||
}
|
||||
// Each entry holds one accumulator per value field and calculated-field source column, so the
|
||||
// width counts too: a handful of groups over hundreds of fields is large as well.
|
||||
if data.aggs.len() > 5_000_000 || data.aggs.len().saturating_mul(input.vals.len() + input.extra.len()) > MAX_ACCUMULATORS {
|
||||
if data.aggs.len() > 5_000_000 {
|
||||
return Err("The PivotTable report is too large.".into());
|
||||
}
|
||||
Ok(())
|
||||
@@ -267,8 +259,8 @@ fn write_headers(g: &mut Grid, input: HeaderInput<'_>) -> Result<(), String> {
|
||||
let lv = spec.path.len().saturating_sub(1);
|
||||
let label = spec.path.last().and_then(|it| col_items.get(lv).map(|fi| fi.label(*it))).unwrap_or_default();
|
||||
let text = match vi {
|
||||
Some(vi) if k > 1 => caption(&label, &val_name(*vi)),
|
||||
_ => caption(&label, "Total"),
|
||||
Some(vi) if k > 1 => format!("{label} {}", val_name(*vi)),
|
||||
_ => format!("{label} Total"),
|
||||
};
|
||||
g.set(first_level + lv, c, Value::text(text), CellKind::ColLabel, 0, None);
|
||||
}
|
||||
@@ -326,7 +318,6 @@ struct ShowContext<'a> {
|
||||
value_index: usize,
|
||||
level0: &'a [u32],
|
||||
siblings: &'a HashMap<Vec<u32>, Vec<u32>>,
|
||||
cache: &'a ShowCache,
|
||||
}
|
||||
|
||||
fn number(value: &Option<Value>) -> Option<f64> {
|
||||
@@ -343,89 +334,42 @@ fn ratio(a: Value, b: Option<Value>) -> Value {
|
||||
}
|
||||
}
|
||||
|
||||
/// What Running Total and Rank need from other cells, computed once per sibling set instead of
|
||||
/// once per displayed value (which rescanned the whole set every time).
|
||||
type ShowKey = (Vec<u32>, Vec<u32>, usize);
|
||||
/// Cumulative values by top-level item, and the total of all of them.
|
||||
type RunningTotals = (HashMap<u32, f64>, f64);
|
||||
/// Top-level items by (rest of a row path, column path).
|
||||
type ItemsByRest = HashMap<(Vec<u32>, Vec<u32>), Vec<u32>>;
|
||||
|
||||
#[derive(Default)]
|
||||
struct ShowCache {
|
||||
/// Ranked values' sibling numbers, sorted ascending without repeats, by (parent, column, value).
|
||||
ranks: std::cell::RefCell<HashMap<ShowKey, Vec<f64>>>,
|
||||
/// Running totals by top-level item, with the total of all of them, by (rest of path, column, value).
|
||||
running: std::cell::RefCell<HashMap<ShowKey, RunningTotals>>,
|
||||
/// The top-level items that have a row path ending in `rest` (with column path `col`), in
|
||||
/// top-level order, by (`rest`, `col`): built once from the report's cells, so a running total
|
||||
/// visits the items that can hold a number, not every top-level item per distinct `rest`.
|
||||
items_by_rest: std::cell::RefCell<Option<ItemsByRest>>,
|
||||
}
|
||||
|
||||
fn running_total(raw: Value, lookup: &ValueLookup<'_>, context: &ShowContext<'_>) -> Value {
|
||||
let Some((first, rest)) = context.row_path.split_first() else { return raw };
|
||||
let Some((first, _)) = context.row_path.split_first() else { return raw };
|
||||
if !matches!(raw, Value::Number(_)) {
|
||||
return raw;
|
||||
}
|
||||
let mut index = context.cache.items_by_rest.borrow_mut();
|
||||
let items_by_rest = index.get_or_insert_with(|| {
|
||||
let position: HashMap<u32, usize> = context.level0.iter().enumerate().rev().map(|(at, item)| (*item, at)).collect();
|
||||
let mut map = ItemsByRest::new();
|
||||
for (row, col) in lookup.aggs.keys() {
|
||||
if let Some((item, rest)) = row.split_first()
|
||||
&& position.contains_key(item)
|
||||
{
|
||||
map.entry((rest.to_vec(), col.clone())).or_default().push(*item);
|
||||
}
|
||||
let mut total = 0.0;
|
||||
for item in context.level0 {
|
||||
let mut path = context.row_path.to_vec();
|
||||
if let Some(first_item) = path.first_mut() {
|
||||
*first_item = *item;
|
||||
}
|
||||
for items in map.values_mut() {
|
||||
items.sort_by_key(|item| position.get(item).copied().unwrap_or(usize::MAX));
|
||||
total += number(&lookup.get(&path, context.col_path, context.value_index)).unwrap_or(0.0);
|
||||
if item == first {
|
||||
break;
|
||||
}
|
||||
map
|
||||
});
|
||||
let mut cache = context.cache.running.borrow_mut();
|
||||
let (by_item, all) = cache.entry((rest.to_vec(), context.col_path.to_vec(), context.value_index)).or_insert_with(|| {
|
||||
let mut total = 0.0;
|
||||
let mut by_item = HashMap::new();
|
||||
for item in items_by_rest.get(&(rest.to_vec(), context.col_path.to_vec())).into_iter().flatten() {
|
||||
let mut path = Vec::with_capacity(rest.len() + 1);
|
||||
path.push(*item);
|
||||
path.extend_from_slice(rest);
|
||||
// An item with no number here adds nothing and is not stored (the shown cell always has one).
|
||||
let Some(value) = number(&lookup.get(&path, context.col_path, context.value_index)) else { continue };
|
||||
total += value;
|
||||
by_item.entry(*item).or_insert(total);
|
||||
}
|
||||
(by_item, total)
|
||||
});
|
||||
Value::number(by_item.get(first).copied().unwrap_or(*all))
|
||||
}
|
||||
Value::number(total)
|
||||
}
|
||||
|
||||
fn ranked_value(raw: Value, lookup: &ValueLookup<'_>, context: &ShowContext<'_>) -> Value {
|
||||
let Some((_, parent)) = context.row_path.split_last() else { return Value::Empty };
|
||||
let Some(me) = raw.as_f64() else { return raw };
|
||||
let mut cache = context.cache.ranks.borrow_mut();
|
||||
let sorted = cache.entry((parent.to_vec(), context.col_path.to_vec(), context.value_index)).or_insert_with(|| {
|
||||
let mut all: Vec<f64> = context
|
||||
.siblings
|
||||
.get(parent)
|
||||
.into_iter()
|
||||
.flatten()
|
||||
.filter_map(|sibling| {
|
||||
let mut path = parent.to_vec();
|
||||
path.push(*sibling);
|
||||
number(&lookup.get(&path, context.col_path, context.value_index))
|
||||
})
|
||||
.collect();
|
||||
all.sort_by(f64::total_cmp);
|
||||
all.dedup();
|
||||
all
|
||||
});
|
||||
let siblings = context.siblings.get(parent).cloned().unwrap_or_default();
|
||||
// Excel's ranks are dense: ties share a rank and the next value takes the next one
|
||||
// (5, 4, 3, 3, 4 rank 1, 2, 3, 3, 2), so count the distinct larger values.
|
||||
let larger = sorted.len() - sorted.partition_point(|value| *value <= me);
|
||||
Value::number(larger as f64 + 1.0)
|
||||
let mut larger: Vec<f64> = siblings
|
||||
.into_iter()
|
||||
.filter_map(|sibling| {
|
||||
let mut path = parent.to_vec();
|
||||
path.push(sibling);
|
||||
number(&lookup.get(&path, context.col_path, context.value_index)).filter(|value| *value > me)
|
||||
})
|
||||
.collect();
|
||||
larger.sort_by(f64::total_cmp);
|
||||
larger.dedup();
|
||||
Value::number(larger.len() as f64 + 1.0)
|
||||
}
|
||||
|
||||
fn show_value(definition: &PivotValue, raw: Value, lookup: &ValueLookup<'_>, context: &ShowContext<'_>) -> Value {
|
||||
@@ -698,7 +642,6 @@ pub(super) fn compute_from(wb: &Workbook, pt: &PivotTable, src: &Source, mut pro
|
||||
})
|
||||
.collect();
|
||||
let lookup = ValueLookup { wb, source: src, calc: &calc, vals: &vals, extra: &extra, aggs: &aggs, value_count: k };
|
||||
let show_cache = ShowCache::default();
|
||||
let mut chart_rows: Vec<u32> = Vec::new();
|
||||
let chart_cols: Vec<(u32, String)> = data_cols
|
||||
.iter()
|
||||
@@ -756,7 +699,7 @@ pub(super) fn compute_from(wb: &Workbook, pt: &PivotTable, src: &Source, mut pro
|
||||
let Some((_, vdef)) = vals.get(*vi) else { continue };
|
||||
let cp = spec.path.as_slice();
|
||||
let Some(raw) = lookup.get(rp, cp, *vi) else { continue };
|
||||
let context = ShowContext { row_path: rp, col_path: cp, value_index: *vi, level0: &level0, siblings: &siblings, cache: &show_cache };
|
||||
let context = ShowContext { row_path: rp, col_path: cp, value_index: *vi, level0: &level0, siblings: &siblings };
|
||||
let shown = show_value(vdef, raw, &lookup, &context);
|
||||
if let Some(pr) = probe.as_deref_mut() {
|
||||
if pr.at == Some((r, l + j)) {
|
||||
|
||||
@@ -21,7 +21,7 @@ pub(super) struct FieldItems {
|
||||
}
|
||||
|
||||
fn cmp_items(a: &Item, b: &Item) -> Ordering {
|
||||
a.class.cmp(&b.class).then(a.num.total_cmp(&b.num)).then_with(|| gridcraft_core::compare_text(&a.text, &b.text))
|
||||
a.class.cmp(&b.class).then(a.num.total_cmp(&b.num)).then_with(|| a.text.cmp(&b.text))
|
||||
}
|
||||
|
||||
/// (identity key, item) for a source value under a field's date grouping.
|
||||
|
||||
@@ -121,18 +121,6 @@ pub(super) enum ColKind {
|
||||
Grand,
|
||||
}
|
||||
|
||||
/// Longest item caption Excel keeps; a subtotal caption repeats its parent's label for every value
|
||||
/// field at every level, so a long source label is cut rather than multiplied.
|
||||
const MAX_CAPTION_LABEL_CHARS: usize = 255;
|
||||
|
||||
/// `label` (cut to [`MAX_CAPTION_LABEL_CHARS`]) and `suffix`, as in "East Total".
|
||||
pub(super) fn caption(label: &str, suffix: &str) -> String {
|
||||
let mut out: String = label.chars().take(MAX_CAPTION_LABEL_CHARS).collect();
|
||||
out.push(' ');
|
||||
out.push_str(suffix);
|
||||
out
|
||||
}
|
||||
|
||||
pub(super) struct ColSpec {
|
||||
pub(super) kind: ColKind,
|
||||
pub(super) path: Vec<u32>,
|
||||
@@ -174,7 +162,7 @@ impl Ctx<'_> {
|
||||
leaf: false,
|
||||
path: prefix.to_vec(),
|
||||
values: true,
|
||||
labels: vec![(col, Value::text(caption(label, name)), indent, None)],
|
||||
labels: vec![(col, Value::text(format!("{label} {name}")), indent, None)],
|
||||
vi: Some(vi),
|
||||
});
|
||||
}
|
||||
@@ -261,7 +249,7 @@ impl Ctx<'_> {
|
||||
leaf: false,
|
||||
path: prefix.clone(),
|
||||
values: true,
|
||||
labels: vec![(col, Value::text(caption(&label, "Total")), indent, None)],
|
||||
labels: vec![(col, Value::text(format!("{label} Total")), indent, None)],
|
||||
vi,
|
||||
});
|
||||
}
|
||||
@@ -328,7 +316,7 @@ impl Ctx<'_> {
|
||||
leaf: false,
|
||||
path: prefix.clone(),
|
||||
values: true,
|
||||
labels: vec![(col, Value::text(caption(&label, "Total")), 0, None)],
|
||||
labels: vec![(col, Value::text(format!("{label} Total")), 0, None)],
|
||||
vi,
|
||||
});
|
||||
}
|
||||
@@ -450,17 +438,3 @@ pub(super) fn cell_style(lk: &Look, rk: RowKind, cell: &OutCell, last_header: bo
|
||||
}
|
||||
s
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn captions_cut_a_long_parent_label() {
|
||||
assert_eq!(caption("East", "Total"), "East Total");
|
||||
let long = "é".repeat(100_000);
|
||||
let c = caption(&long, "Sum of Sales");
|
||||
assert_eq!(c.chars().count(), MAX_CAPTION_LABEL_CHARS + " Sum of Sales".len());
|
||||
assert!(c.ends_with(" Sum of Sales"));
|
||||
}
|
||||
}
|
||||
|
||||
@@ -153,11 +153,6 @@ pub fn compile(formula: &str, field: &dyn Fn(&str) -> Option<String>) -> Result<
|
||||
Expr::Name(n) if n.starts_with(PREFIX) => {}
|
||||
Expr::Name(n) => problem = Some(format!("There's no field named '{n}'.")),
|
||||
Expr::Ref(_) | Expr::Struct(_) => problem = Some("A calculated field can't refer to cells or tables; use field names.".into()),
|
||||
// Evaluated once per group and per report, with only the first value kept: Excel does not
|
||||
// allow functions that need arrays in a calculated field, and these build one from scalars.
|
||||
Expr::Call(f, _) if matches!(f.as_str(), "SEQUENCE" | "RANDARRAY" | "MAKEARRAY" | "EXPAND" | "MUNIT") => {
|
||||
problem = Some(format!("A calculated field can't use {f}: it builds an array."))
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
});
|
||||
@@ -337,16 +332,6 @@ mod tests {
|
||||
assert!(compile(&format!("={}", "1+".repeat(MAX_FORMULA_LEN)), &fields).is_err());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_calculated_field_cannot_build_arrays_from_scalars() {
|
||||
for f in
|
||||
["=SEQUENCE(5000,1000)", "=SUM(RANDARRAY(1000,1000))", "=MAKEARRAY(9,9,LAMBDA(r,c,r*c))+Sales", "=expand(Sales,1000,1000)", "=MUNIT(500)"]
|
||||
{
|
||||
assert!(compile(f, &fields).unwrap_err().contains("builds an array"), "{f}");
|
||||
}
|
||||
assert!(compile("=SUM({1,2,3})*Sales", &fields).is_ok(), "an array constant is bounded by the formula's own length");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn deep_and_cyclic_chains_are_bounded() {
|
||||
// 300 calculated fields, each using the previous one, and a two-field cycle.
|
||||
|
||||
@@ -18,44 +18,6 @@ fn v(s: &Session, a: &str) -> Value {
|
||||
s.doc().unwrap().wb.active().unwrap().value(CellRef::parse(a).unwrap())
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn format_cells_reading_order_survives_undo_and_xlsx() {
|
||||
let mut s = s();
|
||||
s.execute("cell.set", json!({"cell": "A1", "input": "مرحبا 123 ABC"})).unwrap();
|
||||
s.execute("home.bold", json!({"range": "A1"})).unwrap();
|
||||
s.execute("home.formatCells", json!({"range": "A1", "style": {"align": {"reading_order": "RightToLeft"}}})).unwrap();
|
||||
let alignment = |s: &Session| {
|
||||
let d = s.doc().unwrap();
|
||||
let style = d.wb.styles.get(d.wb.active().unwrap().style_id(CellRef::new(0, 0)));
|
||||
assert!(style.font.bold);
|
||||
serde_json::to_value(style.align).unwrap()
|
||||
};
|
||||
assert_eq!(alignment(&s)["reading_order"], "RightToLeft");
|
||||
s.execute("edit.undo", json!({})).unwrap();
|
||||
assert!(alignment(&s)["reading_order"].is_null());
|
||||
s.execute("edit.redo", json!({})).unwrap();
|
||||
assert_eq!(alignment(&s)["reading_order"], "RightToLeft");
|
||||
for order in [json!("invalid"), json!(2)] {
|
||||
assert!(s.execute("home.formatCells", json!({"range": "A1", "style": {"align": {"reading_order": order}}})).is_err());
|
||||
assert_eq!(alignment(&s)["reading_order"], "RightToLeft");
|
||||
}
|
||||
let wb = &s.doc().unwrap().wb;
|
||||
let (reopened, _) = gridcraft_xlsx::read_xlsx(&gridcraft_xlsx::write_xlsx(wb).unwrap()).unwrap();
|
||||
let cell = CellRef::new(0, 0);
|
||||
assert_eq!(reopened.styles.get(reopened.active().unwrap().style_id(cell)), wb.styles.get(wb.active().unwrap().style_id(cell)));
|
||||
assert_eq!(reopened.active().unwrap().value(cell), Value::text("مرحبا 123 ABC"));
|
||||
for order in ["LeftToRight", "Context"] {
|
||||
s.execute("home.formatCells", json!({"range": "A1", "style": {"align": {"reading_order": order}}})).unwrap();
|
||||
assert_eq!(alignment(&s)["reading_order"], order);
|
||||
}
|
||||
s.execute("home.formatCells", json!({"range": "A1", "style": {"align": {"reading_order": null}}})).unwrap();
|
||||
assert!(alignment(&s)["reading_order"].is_null());
|
||||
s.execute("edit.undo", json!({})).unwrap();
|
||||
assert_eq!(alignment(&s)["reading_order"], "Context");
|
||||
s.execute("edit.redo", json!({})).unwrap();
|
||||
assert!(alignment(&s)["reading_order"].is_null());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn entering_empty_cells_preserves_whole_sheet_formatting() {
|
||||
let mut s = s();
|
||||
@@ -1350,18 +1312,6 @@ fn scripts_leave_file_access_working_after_they_run() {
|
||||
let _ = std::fs::remove_dir_all(&dir);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn get_data_refuses_json_with_too_many_fields_or_rows_before_building_rows() {
|
||||
let mut s = s();
|
||||
let wide = format!("[{}]", (0..=gridcraft_core::MAX_COLS).map(|i| format!("{{\"k{i}\":1}}")).collect::<Vec<_>>().join(","));
|
||||
let err = s.execute("data.fromTextCsv", json!({"text": wide})).unwrap_err().to_string();
|
||||
assert!(err.contains("more fields"), "{err}");
|
||||
// MAX_ROWS objects plus the header row are one too many.
|
||||
let tall = format!("[{}]", vec!["{}"; gridcraft_core::MAX_ROWS as usize].join(","));
|
||||
let err = s.execute("data.fromTextCsv", json!({"text": tall})).unwrap_err().to_string();
|
||||
assert!(err.contains("more rows"), "{err}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn scripts_cannot_reach_outside_the_workbook_even_when_trusted() {
|
||||
let dir = std::env::temp_dir().join(format!("gridcraft-script-deny-{}", std::process::id()));
|
||||
@@ -1509,28 +1459,6 @@ fn case_sensitive_sort_puts_lowercase_first() {
|
||||
assert_eq!(col, ["a", "A", "aB", "Ab", "b", "B"].map(Value::from));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn data_sort_orders_vowelled_arabic_by_letters() {
|
||||
let mut s = s();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [["بيت"], ["بَاب"], ["تمر"], ["\u{200F}بحر"]]})).unwrap();
|
||||
s.execute("data.sort", json!({"range": "A1:A4", "header": false, "keys": [{"column": "A"}]})).unwrap();
|
||||
let col: Vec<Value> = (1..=4).map(|r| v(&s, &format!("A{r}"))).collect();
|
||||
assert_eq!(col, ["بَاب", "\u{200F}بحر", "بيت", "تمر"].map(Value::from));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn data_sort_orders_marks_and_digit_scripts_like_excel() {
|
||||
// Excel 16.0 (build 20527, en-US) sorts this column, as text, into exactly this order.
|
||||
let want = ["١٢", "۱۲", "ب\u{064E}1", "ب\u{064E}١", "بa", "\u{200F}باب", "ب\u{064B}اب", "ب\u{064E}اب", "ب\u{0652}اب", "ب\u{0651}اب"];
|
||||
let mut s = s();
|
||||
s.execute("home.numberFormat", json!({"range": "A1:A10", "code": "@"})).unwrap();
|
||||
let input: Vec<Vec<&str>> = want.iter().rev().map(|t| vec![*t]).collect();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": input})).unwrap();
|
||||
s.execute("data.sort", json!({"range": "A1:A10", "header": false, "keys": [{"column": "A"}]})).unwrap();
|
||||
let col: Vec<Value> = (1..=10).map(|r| v(&s, &format!("A{r}"))).collect();
|
||||
assert_eq!(col, want.map(Value::from));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn sort_with_a_sheet_prefix_sorts_the_named_sheet() {
|
||||
let col = |s: &Session, sheet: &str| -> Vec<Value> {
|
||||
@@ -3676,79 +3604,6 @@ fn xls_import_keeps_formulas_and_never_saves_back_to_xls() {
|
||||
assert!(s.execute("file.saveBytes", json!({"format": "xlsx"})).is_ok());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn xls_formulas_calculate_with_legacy_implicit_intersection() {
|
||||
// Data!A1 = 11, B1:G1 = 1, A2 = 5, A3 = 7. Report!B1:G1 are filled, so a formula in Report!A1
|
||||
// that spilled would show #SPILL!. Every cached result is 0, so the values below are GridCraft's.
|
||||
fn rec(id: u16, payload: &[u8]) -> Vec<u8> {
|
||||
[&id.to_le_bytes()[..], &(payload.len() as u16).to_le_bytes(), payload].concat()
|
||||
}
|
||||
let u16s = |v: &[u16]| v.iter().flat_map(|x| x.to_le_bytes()).collect::<Vec<u8>>();
|
||||
let bof = |dt: u16| rec(0x0809, &[&0x0600u16.to_le_bytes()[..], &dt.to_le_bytes(), &[0; 12]].concat());
|
||||
let number = |r: u16, c: u16, n: f64| rec(0x0203, &[&u16s(&[r, c, 0])[..], &n.to_le_bytes()].concat());
|
||||
let formula = |r: u16, c: u16, rgce: &[u8]| {
|
||||
rec(0x0006, &[&u16s(&[r, c, 0])[..], &0f64.to_le_bytes(), &[0; 6], &(rgce.len() as u16).to_le_bytes(), rgce].concat())
|
||||
};
|
||||
let area3d = |r0: u16, r1: u16, c0: u16, c1: u16| [&[0x3Bu8, 0, 0][..], &u16s(&[r0, r1, c0, c1])].concat();
|
||||
let a2a3 = area3d(1, 2, 0, 0);
|
||||
let mut data = bof(0x0010);
|
||||
data.extend([number(0, 0, 11.0), number(1, 0, 5.0), number(2, 0, 7.0)].concat());
|
||||
(1..7).for_each(|c| data.extend(number(0, c, 1.0)));
|
||||
data.extend(rec(0x000A, &[]));
|
||||
let mut report = bof(0x0010);
|
||||
(1..7).for_each(|c| report.extend(number(0, c, 0.0)));
|
||||
// A1 = Data!A1:G1; B2 = Data!$A$2:$A$3*10; C2 = SQRT(Data!$A$2:$A$3) (row 2: SQRT(5));
|
||||
// H1 = {SUM(Data!$A$2:$A$3*2)}, a one-cell Ctrl+Shift+Enter array.
|
||||
let cse = [&a2a3[..], &[0x1E, 2, 0, 0x05, 0x22, 1, 4, 0]].concat();
|
||||
report.extend(formula(0, 0, &area3d(0, 0, 0xC000, 6 | 0xC000)));
|
||||
report.extend(formula(1, 1, &[&a2a3[..], &[0x1E, 10, 0, 0x05]].concat()));
|
||||
report.extend(formula(1, 2, &[&a2a3[..], &[0x21, 20, 0]].concat()));
|
||||
report.extend(formula(0, 7, &[1, 0, 0, 7, 0]));
|
||||
report.extend(rec(0x0221, &[&u16s(&[0, 0])[..], &[7, 7], &[0; 6], &(cse.len() as u16).to_le_bytes(), &cse].concat()));
|
||||
report.extend(rec(0x000A, &[]));
|
||||
let boundsheet = |pos: u32, name: &str| rec(0x0085, &[&pos.to_le_bytes()[..], &[0, 0, name.len() as u8, 0], name.as_bytes()].concat());
|
||||
let globals = |p1: u32, p2: u32| {
|
||||
[
|
||||
bof(0x0005),
|
||||
boundsheet(p1, "Data"),
|
||||
boundsheet(p2, "Report"),
|
||||
rec(0x01AE, &u16s(&[2, 0x0401])),
|
||||
rec(0x0017, &u16s(&[1, 0, 0, 0])),
|
||||
rec(0x000A, &[]),
|
||||
]
|
||||
.concat()
|
||||
};
|
||||
let p1 = globals(0, 0).len() as u32;
|
||||
let stream = [globals(p1, p1 + data.len() as u32), data, report].concat();
|
||||
let bytes = gridcraft_xlsx::testing::compound_file(&[("Workbook", &stream)]);
|
||||
|
||||
let mut s = s();
|
||||
s.execute("file.open", json!({"name": "legacy.xls", "base64": crate::io::base64_encode(&bytes)})).unwrap();
|
||||
s.execute("formulas.calculateNow", json!({})).unwrap();
|
||||
let report = |s: &Session, a1: &str| s.doc().unwrap().wb.sheets[1].value(CellRef::parse(a1).unwrap());
|
||||
assert_eq!(report(&s, "A1"), Value::Number(11.0));
|
||||
assert_eq!(report(&s, "B2"), Value::Number(50.0));
|
||||
assert_eq!(report(&s, "C2"), Value::Number(5f64.sqrt()));
|
||||
assert_eq!(report(&s, "H1"), Value::Number(24.0));
|
||||
// Nothing spilled out of A1 or B2.
|
||||
assert_eq!(report(&s, "B1"), Value::Number(0.0));
|
||||
assert_eq!(report(&s, "B3"), Value::Empty);
|
||||
// Saved as .xlsx, the formulas carry no dynamic-array metadata.
|
||||
let r = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let saved = crate::io::base64_decode(r["base64"].as_str().unwrap()).unwrap();
|
||||
let (wb, _) = gridcraft_xlsx::read_xlsx(&saved).unwrap();
|
||||
let f = |a1: &str| wb.sheets[1].cell(CellRef::parse(a1).unwrap()).and_then(|c| c.formula.clone()).unwrap();
|
||||
assert_eq!(f("A1").text, "@Data!A1:G1");
|
||||
assert_eq!((f("H1").text.as_str(), f("H1").array), ("SUM(Data!$A$2:$A$3*2)", gridcraft_core::RangeRef::parse("H1")));
|
||||
let mut sheet = String::new();
|
||||
std::io::Read::read_to_string(
|
||||
&mut zip::ZipArchive::new(std::io::Cursor::new(&saved)).unwrap().by_name("xl/worksheets/sheet2.xml").unwrap(),
|
||||
&mut sheet,
|
||||
)
|
||||
.unwrap();
|
||||
assert!(!sheet.contains("cm=") && !sheet.contains("SINGLE"), "{sheet}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn malformed_xlsb_keeps_the_current_workbook() {
|
||||
let mut s = s();
|
||||
@@ -6480,36 +6335,6 @@ fn clearing_links_and_notes_keeps_allow_edit_ranges() {
|
||||
assert!(sh.hyperlinks.is_empty() && sh.comments.is_empty());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn replace_does_not_search_hidden_formula_source() {
|
||||
let mut s = s();
|
||||
s.execute("cell.set", json!({"cell": "A1", "input": "=SEQUENCE(3)*7"})).unwrap();
|
||||
s.execute("home.formatCells", json!({"range": "A1", "style": {"protection": {"locked": true, "hidden": true}}})).unwrap();
|
||||
s.execute("review.protectSheet", json!({})).unwrap();
|
||||
// Both give the same answer, and it is not an error: a hidden formula's source is not searched.
|
||||
let hit = s.execute("edit.replace", json!({"what": "SEQUENCE", "with": "X"})).unwrap();
|
||||
let miss = s.execute("edit.replace", json!({"what": "NOSUCHTEXT", "with": "X"})).unwrap();
|
||||
assert_eq!(hit, miss);
|
||||
assert_eq!(hit["replaced"], json!(0), "{hit}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn formulatext_does_not_reveal_a_hidden_formula() {
|
||||
let mut s = s();
|
||||
s.execute("cell.set", json!({"cell": "A1", "input": "=1+1"})).unwrap();
|
||||
s.execute("home.formatCells", json!({"range": "A1", "style": {"protection": {"locked": false, "hidden": true}}})).unwrap();
|
||||
s.execute("home.formatCells", json!({"range": "B1", "style": {"protection": {"locked": false}}})).unwrap();
|
||||
s.execute("cell.set", json!({"cell": "B1", "input": "=FORMULATEXT(A1)"})).unwrap();
|
||||
assert_eq!(v(&s, "B1"), Value::from("=1+1"), "not protected yet");
|
||||
s.execute("review.protectSheet", json!({})).unwrap();
|
||||
// Without entering the formula again: protecting recalculates what reads the protection.
|
||||
assert_eq!(v(&s, "B1"), Value::Error(gridcraft_core::CellError::NA));
|
||||
s.execute("cell.set", json!({"cell": "B1", "input": "=FORMULATEXT(A1)"})).unwrap();
|
||||
assert_eq!(v(&s, "B1"), Value::Error(gridcraft_core::CellError::NA));
|
||||
s.execute("review.unprotectSheet", json!({})).unwrap();
|
||||
assert_eq!(v(&s, "B1"), Value::from("=1+1"), "unprotecting shows it again without re-entering");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn protected_sheet_refuses_layout_commands_its_permissions_do_not_allow() {
|
||||
let mut s = s();
|
||||
@@ -6568,39 +6393,3 @@ fn pivot_new_sheet_destination_respects_workbook_structure_protection() {
|
||||
}
|
||||
assert!(s.execute("insert.pivotTable", json!({"source": "A1:B3", "location": "New"})).is_err());
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn circle_invalid_says_when_its_scan_allowance_ended_early() {
|
||||
let mut s = s();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": [[-1]]})).unwrap();
|
||||
s.execute("range.setValues", json!({"range": "A200000", "values": [[-1]]})).unwrap();
|
||||
s.execute("data.validation", json!({"range": "A1:A5", "type": "whole", "operator": "greater", "formula1": "0"})).unwrap();
|
||||
let out = s.execute("data.circleInvalid", json!({})).unwrap();
|
||||
assert_eq!((out["invalid"].as_array().map(Vec::len), out["truncated"].as_bool()), (Some(1), Some(false)), "{out}");
|
||||
s.execute("data.validation", json!({"range": "A1:A200000", "type": "whole", "operator": "greater", "formula1": "0"})).unwrap();
|
||||
let out = s.execute("data.circleInvalid", json!({})).unwrap();
|
||||
assert_eq!(out["truncated"].as_bool(), Some(true), "{out}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn history_charges_a_changed_sheets_comment_text() {
|
||||
let mut s = s();
|
||||
s.execute("review.newNote", json!({"cell": "A1", "text": "n".repeat(30_000), "author": "Example"})).unwrap();
|
||||
let older = (*s.doc().unwrap().wb).clone();
|
||||
s.execute("cell.set", json!({"cell": "B2", "input": "x"})).unwrap();
|
||||
let newer = (*s.doc().unwrap().wb).clone();
|
||||
// The older snapshot owns the comment text now that the sheet changed under it.
|
||||
assert!(crate::history_retained_bytes(&older, &newer) >= 30_000);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn history_charges_the_storage_of_empty_comment_replies() {
|
||||
let mut s = s();
|
||||
s.execute("review.newComment", json!({"cell": "A1", "text": "c", "author": "Ann"})).unwrap();
|
||||
for _ in 0..500 {
|
||||
s.execute("review.replyComment", json!({"cell": "A1", "text": "", "author": ""})).unwrap();
|
||||
}
|
||||
let sheet = s.doc().unwrap().wb.sheet(0).unwrap().clone();
|
||||
// 500 replies of no text are still 500 elements.
|
||||
assert!(crate::sheet_text_bytes(&sheet) >= 500 * std::mem::size_of::<(String, String)>());
|
||||
}
|
||||
|
||||
@@ -59,31 +59,3 @@ fn whole_row_format_keeps_hidden_rows_hidden() {
|
||||
assert!(info.size.unwrap_or(before) > before + 10.0, "hidden row still grows to its font: before={before}, size={:?}", info.size);
|
||||
assert!(info.hidden, "formatting a hidden row must not unhide it");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn many_merges_fit_rows_the_same_and_skip_merged_cells() {
|
||||
// Rows are fitted with one sweep over the sheet's merges; a merged cell still never grows
|
||||
// its row, and rows next to thousands of merges still fit.
|
||||
let mut s = gridcraft_engine::Session::new();
|
||||
s.new_workbook();
|
||||
let values: Vec<_> = (0..2_000).map(|i| json!([format!("v{i}")])).collect();
|
||||
s.execute("range.setValues", json!({"range": "A1", "values": values})).unwrap();
|
||||
{
|
||||
let sh = sheet_mut(&mut s);
|
||||
sh.merges.push(gridcraft_engine::core::RangeRef::parse("A3:B4").unwrap());
|
||||
for i in 0..20_000u32 {
|
||||
let row = 4 + i % 1_990;
|
||||
sh.merges.push(gridcraft_engine::core::RangeRef::new(
|
||||
gridcraft_engine::core::CellRef::new(row, 3 + (i % 7) * 2),
|
||||
gridcraft_engine::core::CellRef::new(row + 3, 4 + (i % 7) * 2),
|
||||
));
|
||||
}
|
||||
}
|
||||
s.execute("home.fontSize", json!({"range": "A:A", "size": 20})).unwrap();
|
||||
let sh = s.doc().unwrap().wb.active().unwrap();
|
||||
let default = sh.default_row_height;
|
||||
assert!(sh.row_height(0) > 30.0 && sh.row_height(1_999) > 30.0, "rows with big text grow");
|
||||
assert_eq!(sh.row_height(2), default, "a merged cell does not grow its row");
|
||||
assert_eq!(sh.row_height(3), default, "nor any row of the merge");
|
||||
assert!(sh.row_height(4) > 30.0 && sh.row_height(1_000) > 30.0, "rows beside other merges grow");
|
||||
}
|
||||
|
||||
@@ -1,366 +0,0 @@
|
||||
//! Legacy formulas (no dynamic-array metadata) keep one value through GridCraft (issue #410), on
|
||||
//! a calculated workbook:
|
||||
//! - every function that returns several values: a legacy cell formula calling it opens with
|
||||
//! Excel's implicit `@` and does not spill, and a typed `=@F(…)` saves and reopens with the same
|
||||
//! single value. The library's functions are found by calling each with sample arguments, so
|
||||
//! one added later is covered without editing this test.
|
||||
//! - a legacy formula the analysis cannot finish does not spill and saves as read, without `cm`.
|
||||
|
||||
use std::io::Write;
|
||||
|
||||
use gridcraft_engine::Session;
|
||||
use gridcraft_engine::core::{CellRef, DateSystem, Value};
|
||||
use gridcraft_functions::{Arg, Ctx, FnSpec};
|
||||
use serde_json::json;
|
||||
|
||||
struct Probe;
|
||||
|
||||
impl Ctx for Probe {
|
||||
fn date_system(&self) -> DateSystem {
|
||||
DateSystem::D1900
|
||||
}
|
||||
fn now_serial(&self) -> f64 {
|
||||
45_000.0
|
||||
}
|
||||
fn random(&mut self) -> f64 {
|
||||
0.5
|
||||
}
|
||||
}
|
||||
|
||||
/// The 2×2 block of numbers at `A1:B2` in the probe sheets.
|
||||
fn block() -> Value {
|
||||
Value::Array(std::sync::Arc::new(
|
||||
gridcraft_engine::core::Array::new(2, 2, vec![Value::Number(1.0), Value::Number(2.0), Value::Number(3.0), Value::Number(4.0)]).unwrap(),
|
||||
))
|
||||
}
|
||||
|
||||
/// Sample arguments (formula text, value). Single values only at a position the function lifts
|
||||
/// over, so an array result is the function's own, not lifting.
|
||||
fn candidates(scalar: bool) -> Vec<(&'static str, Arg)> {
|
||||
let mut out = vec![
|
||||
("3", Arg::val(Value::Number(3.0))),
|
||||
("1", Arg::val(Value::Number(1.0))),
|
||||
("\"a1,b22\"", Arg::val(Value::text("a1,b22"))),
|
||||
("\",\"", Arg::val(Value::text(","))),
|
||||
("\"[0-9]+\"", Arg::val(Value::text("[0-9]+"))),
|
||||
("TRUE", Arg::val(Value::Bool(true))),
|
||||
];
|
||||
if !scalar {
|
||||
out.insert(0, ("A1:B2", Arg::reference(block())));
|
||||
out.insert(1, ("{1,2;3,4}", Arg::val(block())));
|
||||
}
|
||||
out
|
||||
}
|
||||
|
||||
/// The first sample call of `spec` that returns more than one value, as formula text.
|
||||
fn array_call(spec: &FnSpec) -> Option<String> {
|
||||
let max = spec.max_args.min(spec.min_args.max(1) + 2).min(4).max(spec.min_args);
|
||||
for n in spec.min_args..=max {
|
||||
let per: Vec<_> = (0..n).map(|i| candidates(spec.is_scalar_arg(i))).collect();
|
||||
let per: Vec<_> = if n > 4 { per.into_iter().map(|c| c.into_iter().take(3).collect::<Vec<_>>()).collect() } else { per };
|
||||
let mut idx = vec![0usize; n];
|
||||
let mut tries = 0;
|
||||
loop {
|
||||
let args: Vec<Arg> = idx.iter().zip(&per).map(|(&i, c)| c[i].1.clone()).collect();
|
||||
if let Value::Array(a) = gridcraft_functions::call(spec, &args, &mut Probe)
|
||||
&& a.data.len() > 1
|
||||
{
|
||||
let text: Vec<&str> = idx.iter().zip(&per).map(|(&i, c)| c[i].0).collect();
|
||||
return Some(format!("{}({})", spec.name, text.join(",")));
|
||||
}
|
||||
tries += 1;
|
||||
// Next combination, last position fastest.
|
||||
let mut k = n;
|
||||
loop {
|
||||
if k == 0 || tries > 3000 {
|
||||
break;
|
||||
}
|
||||
k -= 1;
|
||||
idx[k] += 1;
|
||||
if idx[k] < per[k].len() {
|
||||
break;
|
||||
}
|
||||
idx[k] = 0;
|
||||
if k == 0 {
|
||||
k = usize::MAX;
|
||||
break;
|
||||
}
|
||||
}
|
||||
if n == 0 || k == usize::MAX || tries > 3000 {
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
/// Every library function with an array result, as a sample call, and the functions the
|
||||
/// evaluator implements itself that return several values or a range.
|
||||
fn array_calls() -> Vec<String> {
|
||||
let mut calls: Vec<String> = gridcraft_functions::all().iter().filter_map(array_call).collect();
|
||||
for found in ["MUNIT", "MODE.MULT", "TEXTSPLIT", "REGEXEXTRACT", "MMULT", "SEQUENCE"] {
|
||||
assert!(calls.iter().any(|c| c.starts_with(&format!("{found}("))), "{found} not found: {calls:?}");
|
||||
}
|
||||
calls.extend(
|
||||
["TRIMRANGE(A1:B2)", "OFFSET(A1:B2,0,0)", "INDIRECT(\"A1:B2\")", "INDEX(A1:B2,0,1)", "FILTERXML(\"<a><b>1</b><b>2</b></a>\",\"//b\")"]
|
||||
.map(String::from),
|
||||
);
|
||||
calls
|
||||
}
|
||||
|
||||
/// Where the `i`th sample call goes: column E, five rows apart, away from the data in `A1:B2`.
|
||||
fn slot(i: usize) -> CellRef {
|
||||
CellRef::new(9 + 5 * i as u32, 4)
|
||||
}
|
||||
|
||||
fn value(s: &Session, c: CellRef) -> Value {
|
||||
s.doc().unwrap().wb.active().unwrap().value(c)
|
||||
}
|
||||
|
||||
/// The cells right of and below `c` are empty and `c` is not #SPILL!: the formula did not spill.
|
||||
fn one_value(s: &Session, c: CellRef, what: &str) -> Result<Value, String> {
|
||||
let v = value(s, c);
|
||||
let right = value(s, CellRef::new(c.row, c.col + 1));
|
||||
let below = value(s, CellRef::new(c.row + 1, c.col));
|
||||
if right != Value::Empty || below != Value::Empty || v.display().contains("SPILL") {
|
||||
return Err(format!("{what}: {} spilled ({right:?} right, {below:?} below)", v.display()));
|
||||
}
|
||||
Ok(v)
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn legacy_formulas_with_array_functions_show_one_value() {
|
||||
let calls = array_calls();
|
||||
let formulas: Vec<_> = calls.iter().enumerate().map(|(i, f)| (slot(i).a1(), f.clone())).collect();
|
||||
let s = open(&legacy_xlsx(&formulas, ""));
|
||||
let wrong: Vec<_> = calls.iter().enumerate().filter_map(|(i, f)| one_value(&s, slot(i), f).err()).collect();
|
||||
assert!(wrong.is_empty(), "{wrong:#?}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn typed_at_of_array_functions_saves_one_value() {
|
||||
let calls = array_calls();
|
||||
let mut s = session();
|
||||
for (a, v) in [("A1", 1), ("B1", 2), ("A2", 3), ("B2", 4)] {
|
||||
s.execute("cell.set", json!({"cell": a, "input": v.to_string()})).unwrap();
|
||||
}
|
||||
for (i, f) in calls.iter().enumerate() {
|
||||
s.execute("cell.set", json!({"cell": slot(i).a1(), "input": format!("=@{f}")})).unwrap();
|
||||
}
|
||||
let before: Vec<_> = (0..calls.len()).map(|i| value(&s, slot(i))).collect();
|
||||
let saved = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let mut back = Session::new();
|
||||
back.execute("file.open", json!({"name": "typed.xlsx", "base64": saved["base64"]})).unwrap();
|
||||
let wrong: Vec<_> = calls
|
||||
.iter()
|
||||
.enumerate()
|
||||
.filter_map(|(i, f)| match one_value(&back, slot(i), f) {
|
||||
Ok(v) if v == before[i] => None,
|
||||
// A volatile function recalculates on open: one number again.
|
||||
Ok(Value::Number(_)) if volatile(f) => None,
|
||||
Ok(v) => Some(format!("@{f}: {} before saving, {} after", before[i].display(), v.display())),
|
||||
Err(e) => Some(e),
|
||||
})
|
||||
.collect();
|
||||
assert!(wrong.is_empty(), "{wrong:#?}");
|
||||
}
|
||||
|
||||
fn volatile(call: &str) -> bool {
|
||||
call.split('(').next().and_then(gridcraft_functions::lookup).is_some_and(|spec| spec.volatile)
|
||||
}
|
||||
|
||||
/// Formulas the analysis covers at Excel's nesting limits, and one it cannot finish (names too
|
||||
/// large to analyse, whose ranges would spill three rows): calculated, none spills, and saving
|
||||
/// writes each as read, as a plain `<f>`.
|
||||
#[test]
|
||||
fn deep_and_unfinished_legacy_formulas_calculate_and_save_as_read() {
|
||||
let nested = format!("{}A1:A3{}", "ABS(".repeat(200), ")".repeat(200));
|
||||
let parens = format!("{}A1:A3{}", "ABS((".repeat(64), "))".repeat(64));
|
||||
let parts = (1..=30).map(|i| format!("Part_{i}")).collect::<Vec<_>>().join("+");
|
||||
let part = format!("Sheet1!$A$1:$A$3+{}", vec!["Sheet1!$B$1"; 599].join("+"));
|
||||
let names: String = (1..=30).map(|i| format!(r#"<definedName name="Part_{i}">{part}</definedName>"#)).collect();
|
||||
let formulas = [("E10".to_string(), nested), ("E15".into(), parens), ("E20".into(), parts)];
|
||||
let mut s = open(&legacy_xlsx(&formulas, &names));
|
||||
for (a, _) in &formulas {
|
||||
one_value(&s, CellRef::parse(a).unwrap(), a).unwrap();
|
||||
}
|
||||
// Legacy Excel intersects A1:A3 with row 10 and 15: #VALUE!. The unfinished formula shows the
|
||||
// top-left value of its sum, 1 + 599 × 2 per name.
|
||||
assert_eq!(value(&s, CellRef::parse("E10").unwrap()).display(), "#VALUE!");
|
||||
assert_eq!(value(&s, CellRef::parse("E15").unwrap()).display(), "#VALUE!");
|
||||
assert_eq!(value(&s, CellRef::parse("E20").unwrap()), Value::Number(30.0 * (1.0 + 599.0 * 2.0)));
|
||||
let saved = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let bytes = gridcraft_engine::io::base64_decode(saved["base64"].as_str().unwrap()).unwrap();
|
||||
let mut zip = zip::ZipArchive::new(std::io::Cursor::new(bytes)).unwrap();
|
||||
let mut xml = String::new();
|
||||
std::io::Read::read_to_string(&mut zip.by_name("xl/worksheets/sheet1.xml").unwrap(), &mut xml).unwrap();
|
||||
for (a, f) in &formulas {
|
||||
assert!(xml.contains(&format!("<f>{f}</f>")), "{a} not saved as read: {xml}");
|
||||
}
|
||||
assert!(!xml.contains("cm=") && !xml.contains("SINGLE") && !xml.contains("t=\"array\""), "{xml}");
|
||||
}
|
||||
|
||||
/// Legacy unions where one value is expected, a name whose definition is IF of a range, and
|
||||
/// multi-valued results in the file's own parentheses: calculated, none spills (legacy Excel
|
||||
/// intersects a union to #VALUE!), and saving writes each as read, as a plain `<f>`.
|
||||
#[test]
|
||||
fn legacy_unions_names_and_parentheses_calculate_and_save_as_read() {
|
||||
let formulas: Vec<(String, String)> = ["SQRT((A1:A2,B1:B2))", "ABS(((A1:A2,B1:B2)))", "(A1:A2,B1:B2)+1", "SQRT(Flag)", "({1,2}+1)", "{1,2}+1"]
|
||||
.iter()
|
||||
.enumerate()
|
||||
.map(|(i, f)| (slot(i).a1(), f.to_string()))
|
||||
.collect();
|
||||
let names = r#"<definedName name="Flag">IF(Sheet1!$A$1:$A$3>1,1,0)</definedName>"#;
|
||||
let mut s = open(&legacy_xlsx(&formulas, names));
|
||||
for (a, _) in &formulas {
|
||||
one_value(&s, CellRef::parse(a).unwrap(), a).unwrap();
|
||||
}
|
||||
let shown: Vec<String> = formulas.iter().map(|(a, _)| value(&s, CellRef::parse(a).unwrap()).display()).collect();
|
||||
assert_eq!(shown, ["#VALUE!", "#VALUE!", "#VALUE!", "0", "2", "2"]);
|
||||
let saved = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let bytes = gridcraft_engine::io::base64_decode(saved["base64"].as_str().unwrap()).unwrap();
|
||||
let mut zip = zip::ZipArchive::new(std::io::Cursor::new(bytes)).unwrap();
|
||||
let mut xml = String::new();
|
||||
std::io::Read::read_to_string(&mut zip.by_name("xl/worksheets/sheet1.xml").unwrap(), &mut xml).unwrap();
|
||||
for (a, f) in &formulas {
|
||||
assert!(xml.contains(&format!("<f>{f}</f>")), "{a} not saved as read: {xml}");
|
||||
}
|
||||
assert!(!xml.contains("cm=") && !xml.contains("SINGLE") && !xml.contains("t=\"array\""), "{xml}");
|
||||
}
|
||||
|
||||
/// A whole column in a scalar argument of a legacy formula meets the formula's row (issue #1636,
|
||||
/// from PR #1660): `ISBLANK(B:B)` in row 4 tests `B4`, so the formula shows `C4` instead of
|
||||
/// spilling into `G5`, and in row 6, where `B6` is blank, shows 0.
|
||||
#[test]
|
||||
fn whole_column_in_a_legacy_scalar_argument_meets_the_formulas_row() {
|
||||
let formulas: Vec<(String, String)> = [
|
||||
("B4", r#""Thierry""#),
|
||||
("C4", "7"),
|
||||
("G4", r#"IF(ISBLANK(B:B),0,IF(B4="Thierry",C4,0))"#),
|
||||
("G5", "1"),
|
||||
("G6", r#"IF(ISBLANK(B:B),0,IF(B6="Thierry",C6,0))"#),
|
||||
]
|
||||
.iter()
|
||||
.map(|(a, f)| (a.to_string(), f.to_string()))
|
||||
.collect();
|
||||
let s = open(&legacy_xlsx(&formulas, ""));
|
||||
let at = |a: &str| CellRef::parse(a).unwrap();
|
||||
assert_eq!(value(&s, at("G4")), Value::Number(7.0), "B:B meets row 4 instead of spilling");
|
||||
assert_eq!(value(&s, at("G5")), Value::Number(1.0), "nothing spills below G4");
|
||||
assert_eq!(one_value(&s, at("G6"), "G6").unwrap(), Value::Number(0.0), "B6 is blank");
|
||||
}
|
||||
|
||||
/// One-cell ranges written as ranges (`B1:B1`) outside any `@`, and formulas whose reading
|
||||
/// reaches the parser's nesting limit (254 nested functions, far past Excel's 64): saved as read,
|
||||
/// without `@`. Only a one-cell range under an `@` counts as one an edit may have shrunk. All but
|
||||
/// the last are calculated without spilling and saved as a plain `<f>`.
|
||||
#[test]
|
||||
fn one_cell_ranges_and_formulas_at_the_nesting_limit_save_as_read() {
|
||||
let deep = |inner: &str| format!("{}{inner}{}", "ABS(".repeat(254), ")".repeat(254));
|
||||
let formulas: Vec<(String, String)> = [
|
||||
"SQRT(A1:A3)+B1:B1".to_string(),
|
||||
"A1:A3*B2:B2".into(),
|
||||
"SQRT(A1:A3)+$B$2:$B$2".into(),
|
||||
// Names with a one-cell range inside CHOOSE's index or IF's condition.
|
||||
"SQRT(NC)".into(),
|
||||
"SQRT(NI)".into(),
|
||||
"NC+1".into(),
|
||||
"NI*2".into(),
|
||||
deep("A1:A3%"),
|
||||
deep("-A1:A3"),
|
||||
]
|
||||
.into_iter()
|
||||
.enumerate()
|
||||
.map(|(i, f)| (slot(i).a1(), f))
|
||||
.collect();
|
||||
let names = r#"<definedName name="NC">CHOOSE(Sheet1!$B$1:$B$1,Sheet1!$A$1:$A$3,Sheet1!$A$1:$A$3)</definedName><definedName name="NI">IF(Sheet1!$B$1:$B$1>0,Sheet1!$A$1:$A$3)</definedName>"#;
|
||||
let (last, rest) = formulas.split_last().unwrap();
|
||||
let mut s = open(&legacy_xlsx(&formulas, names));
|
||||
for (a, _) in rest {
|
||||
one_value(&s, CellRef::parse(a).unwrap(), a).unwrap();
|
||||
}
|
||||
let saved = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let bytes = gridcraft_engine::io::base64_decode(saved["base64"].as_str().unwrap()).unwrap();
|
||||
let mut zip = zip::ZipArchive::new(std::io::Cursor::new(bytes)).unwrap();
|
||||
let mut xml = String::new();
|
||||
std::io::Read::read_to_string(&mut zip.by_name("xl/worksheets/sheet1.xml").unwrap(), &mut xml).unwrap();
|
||||
for (a, f) in rest {
|
||||
assert!(xml.contains(&format!("<f>{f}</f>")), "{a} not saved as read: {xml}");
|
||||
}
|
||||
// A reading the parser cannot nest leaves the formula as read: it calculates as a
|
||||
// dynamic-array formula would, but the file never gets an `@`.
|
||||
let (a, f) = last;
|
||||
assert!(xml.contains(&format!("{f}</f>")), "{a} not saved as read: {xml}");
|
||||
assert!(!xml.contains('@') && !xml.contains("SINGLE"), "{xml}");
|
||||
assert_eq!(xml.matches("cm=").count(), 1, "{xml}");
|
||||
}
|
||||
|
||||
/// Deleting rows shrinks the ranges names return (directly, through another name, CHOOSE's and
|
||||
/// IF's results, parentheses) to one cell: the `@` the reader put on each name is still the one
|
||||
/// legacy Excel implies, so the formulas save as legacy ones, without `_xlfn.SINGLE`.
|
||||
#[test]
|
||||
fn names_shrunk_by_deleting_rows_save_as_legacy() {
|
||||
let formulas: Vec<(String, String)> =
|
||||
["SQRT(MM)", "SQRT(CHM)", "SQRT(NCH)", "SQRT(NIF)", "SQRT(PM)"].into_iter().enumerate().map(|(i, f)| (slot(i).a1(), f.to_string())).collect();
|
||||
let names = r#"<definedName name="MM">Sheet1!$A$1:$A$3</definedName><definedName name="CHM">MM</definedName><definedName name="NCH">CHOOSE(1,Sheet1!$A$1:$A$3,Sheet1!$B$1:$B$3)</definedName><definedName name="NIF">IF(1,Sheet1!$A$1:$A$3,Sheet1!$B$1:$B$3)</definedName><definedName name="PM">(Sheet1!$A$1:$A$3)</definedName>"#;
|
||||
let mut s = open(&legacy_xlsx(&formulas, names));
|
||||
s.execute("home.deleteRows", json!({"rows": "2:3"})).unwrap();
|
||||
let saved = s.execute("file.saveBytes", json!({"format": "xlsx"})).unwrap();
|
||||
let bytes = gridcraft_engine::io::base64_decode(saved["base64"].as_str().unwrap()).unwrap();
|
||||
let mut zip = zip::ZipArchive::new(std::io::Cursor::new(bytes)).unwrap();
|
||||
let mut xml = String::new();
|
||||
std::io::Read::read_to_string(&mut zip.by_name("xl/worksheets/sheet1.xml").unwrap(), &mut xml).unwrap();
|
||||
for (a, f) in &formulas {
|
||||
assert!(xml.contains(&format!("<f>{f}</f>")), "{a} not saved as legacy {f}: {xml}");
|
||||
}
|
||||
assert!(!xml.contains("SINGLE") && !xml.contains("cm="), "{xml}");
|
||||
}
|
||||
|
||||
const CT: &str = r#"<?xml version="1.0"?><Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types"><Default Extension="xml" ContentType="application/xml"/><Default Extension="rels" ContentType="application/vnd.openxmlformats-package.relationships+xml"/><Override PartName="/xl/workbook.xml" ContentType="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet.main+xml"/><Override PartName="/xl/worksheets/sheet1.xml" ContentType="application/vnd.openxmlformats-officedocument.spreadsheetml.worksheet+xml"/></Types>"#;
|
||||
|
||||
/// A one-sheet package: numbers 1–4 in `A1:B2` and `formulas` (address, file text) saved as
|
||||
/// legacy formulas, without dynamic-array metadata; `names` is `<definedName>` elements.
|
||||
fn legacy_xlsx(formulas: &[(String, String)], names: &str) -> Vec<u8> {
|
||||
let esc = |s: &str| s.replace('&', "&").replace('<', "<").replace('>', ">").replace('"', """);
|
||||
let mut rows =
|
||||
String::from(r#"<row r="1"><c r="A1"><v>1</v></c><c r="B1"><v>2</v></c></row><row r="2"><c r="A2"><v>3</v></c><c r="B2"><v>4</v></c></row>"#);
|
||||
// Cells in one row share its `<row>`, in the order given.
|
||||
let mut by_row: Vec<(String, String)> = Vec::new();
|
||||
for (a, f) in formulas {
|
||||
let r: String = a.chars().filter(char::is_ascii_digit).collect();
|
||||
let c = format!(r#"<c r="{a}"><f>{}</f></c>"#, esc(f));
|
||||
match by_row.iter_mut().find(|(row, _)| *row == r) {
|
||||
Some((_, cells)) => cells.push_str(&c),
|
||||
None => by_row.push((r, c)),
|
||||
}
|
||||
}
|
||||
for (r, cells) in by_row {
|
||||
rows.push_str(&format!(r#"<row r="{r}">{cells}</row>"#));
|
||||
}
|
||||
let names = if names.is_empty() { String::new() } else { format!("<definedNames>{names}</definedNames>") };
|
||||
let parts: [(&str, String); 5] = [
|
||||
("[Content_Types].xml", CT.into()),
|
||||
("_rels/.rels", r#"<?xml version="1.0"?><Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships"><Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument" Target="xl/workbook.xml"/></Relationships>"#.into()),
|
||||
("xl/workbook.xml", format!(r#"<?xml version="1.0"?><workbook xmlns="http://schemas.openxmlformats.org/spreadsheetml/2006/main" xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships"><sheets><sheet name="Sheet1" sheetId="1" r:id="rId1"/></sheets>{names}</workbook>"#)),
|
||||
("xl/_rels/workbook.xml.rels", r#"<?xml version="1.0"?><Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships"><Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/worksheet" Target="worksheets/sheet1.xml"/></Relationships>"#.into()),
|
||||
("xl/worksheets/sheet1.xml", format!(r#"<?xml version="1.0"?><worksheet xmlns="http://schemas.openxmlformats.org/spreadsheetml/2006/main"><sheetData>{rows}</sheetData></worksheet>"#)),
|
||||
];
|
||||
let mut z = zip::ZipWriter::new(std::io::Cursor::new(Vec::new()));
|
||||
for (name, data) in parts {
|
||||
z.start_file(name, zip::write::SimpleFileOptions::default()).unwrap();
|
||||
z.write_all(data.as_bytes()).unwrap();
|
||||
}
|
||||
z.finish().unwrap().into_inner()
|
||||
}
|
||||
|
||||
fn session() -> Session {
|
||||
let mut s = Session::new();
|
||||
s.new_workbook();
|
||||
s
|
||||
}
|
||||
|
||||
fn open(bytes: &[u8]) -> Session {
|
||||
let mut s = Session::new();
|
||||
s.execute("file.open", json!({"name": "legacy.xlsx", "base64": gridcraft_engine::io::base64_encode(bytes)})).unwrap();
|
||||
s
|
||||
}
|
||||
@@ -69,35 +69,11 @@ fn parse_top(src: &str, locale: crate::FormulaLocale, allow_union: bool) -> Resu
|
||||
let mut p = Parser { toks, i: 0, depth: 0, input: locale == crate::FormulaLocale::Es };
|
||||
let e = p.expr(0, allow_union)?;
|
||||
match p.peek() {
|
||||
Tok::Eof if wrapper_height(&e) > MAX_DEPTH => Err(ParseError { msg: "formula is nested too deeply".into(), pos: 0 }),
|
||||
Tok::Eof => Ok(e),
|
||||
t => Err(p.err(format!("unexpected {}", describe(t)))),
|
||||
}
|
||||
}
|
||||
|
||||
/// The most unary operators, parentheses and invocations on one path of `e`. The parser bounds
|
||||
/// each level's postfix wrappers by its own depth, so levels inside levels can still stack them
|
||||
/// far past [`MAX_DEPTH`]; code that maps over the tree recurses once per wrapper.
|
||||
fn wrapper_height(e: &Expr) -> usize {
|
||||
let mut most = 0;
|
||||
let mut pending = vec![(e, 0usize)];
|
||||
while let Some((e, h)) = pending.pop() {
|
||||
most = most.max(h);
|
||||
match e {
|
||||
Expr::Unary(_, a) | Expr::Paren(a) => pending.push((a, h + 1)),
|
||||
Expr::Invoke(callee, args) => {
|
||||
pending.push((callee, h + 1));
|
||||
pending.extend(args.iter().map(|a| (a, h + 1)));
|
||||
}
|
||||
Expr::Binary(_, a, b) => pending.extend([(&**a, h), (&**b, h)]),
|
||||
Expr::Call(_, args) => pending.extend(args.iter().map(|a| (a, h))),
|
||||
Expr::Array(rows) => pending.extend(rows.iter().flatten().map(|a| (a, h))),
|
||||
_ => {}
|
||||
}
|
||||
}
|
||||
most
|
||||
}
|
||||
|
||||
fn describe(t: &Tok) -> String {
|
||||
match t {
|
||||
Tok::RParen => "`)`".into(),
|
||||
@@ -326,15 +302,11 @@ impl Parser {
|
||||
{
|
||||
return Ok(Expr::Unary(UnOp::Spill, Box::new(r)));
|
||||
}
|
||||
// Files write `@x` as `_xlfn.SINGLE(x)`; `SINGLE(A1:A3*2)` is `@(A1:A3*2)`.
|
||||
// Files write `@x` as `_xlfn.SINGLE(x)`.
|
||||
if name == "SINGLE"
|
||||
&& args.len() == 1
|
||||
&& let Some(x) = args.pop()
|
||||
{
|
||||
let x = match x {
|
||||
Expr::Binary(..) | Expr::Unary(UnOp::Neg | UnOp::Plus | UnOp::Percent, _) => Expr::Paren(Box::new(x)),
|
||||
x => x,
|
||||
};
|
||||
return Ok(Expr::Unary(UnOp::At, Box::new(x)));
|
||||
}
|
||||
// Files may spell a trim reference as a function call (`_xlfn._TRO_ALL(A1:E10)`).
|
||||
|
||||
@@ -106,15 +106,6 @@ fn spill_operator() -> Result<(), Box<dyn std::error::Error>> {
|
||||
Ok(())
|
||||
}
|
||||
|
||||
/// Files spell `@x` as `_xlfn.SINGLE(x)`; an operation inside keeps its parentheses.
|
||||
#[test]
|
||||
fn single_is_the_implicit_intersection_operator() {
|
||||
assert_eq!(rt("_xlfn.SINGLE(A1:A3)*2"), "@A1:A3*2");
|
||||
assert_eq!(rt("_xlfn.SINGLE(A1:A3*2)"), "@(A1:A3*2)");
|
||||
assert_eq!(rt("_xlfn.SINGLE(-A1:A3)"), "@(-A1:A3)");
|
||||
assert_eq!(rt("SQRT(_xlfn.SINGLE(Data[Qty]))"), "SQRT(@Data[Qty])");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn delete_sheet_shrinks_3d_span() -> Result<(), Box<dyn std::error::Error>> {
|
||||
let order: Vec<String> = ["Sheet1", "Sheet2", "Sheet3", "Results"].iter().map(|s| s.to_string()).collect();
|
||||
@@ -576,19 +567,3 @@ fn column_starting_with_at_round_trips() {
|
||||
assert_eq!(parse(&text).ok(), Some(Expr::Struct(s)), "printed as {text}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn wrappers_stacked_across_levels_are_bounded_as_a_whole() {
|
||||
// Each level adds 40 percents inside the parser's per-level bound, so no level is too deep alone.
|
||||
let stacked = |levels: usize| {
|
||||
let mut f = "1".to_string();
|
||||
for _ in 0..levels {
|
||||
f = format!("({f}){}", "%".repeat(40));
|
||||
}
|
||||
f
|
||||
};
|
||||
assert!(parse(&stacked(5)).is_ok());
|
||||
assert!(parse(&stacked(7)).is_err(), "7 x 41 wrappers on one path");
|
||||
assert!(parse(&format!("{}1{}", "(".repeat(200), ")".repeat(200))).is_ok());
|
||||
assert!(parse(&format!("={}", vec!["A1%"; 100].join("+"))).is_ok(), "siblings are not stacked");
|
||||
}
|
||||
|
||||
@@ -70,59 +70,39 @@ fn cell_criterion(v: &Value) -> Option<Criterion> {
|
||||
/// when the label is unknown, criterion).
|
||||
type CriteriaRow = (usize, Vec<(usize, Option<usize>, Criterion)>);
|
||||
|
||||
/// Most condition tests one selection may make (records times conditions): every record can be
|
||||
/// tested against every criteria row when none matches, so a modest database and criteria table
|
||||
/// multiply into billions of tests. Past it the function is `#NUM!`, as lifted lookups are; 5e8
|
||||
/// is a few seconds of matching.
|
||||
const SELECTION_WORK_BUDGET: usize = 500_000_000;
|
||||
|
||||
/// Rows (1-based within the database, excluding the header) of matching records.
|
||||
fn matching_records(db: &Array, crit: &Array, ctx: &mut dyn Ctx) -> R<Vec<usize>> {
|
||||
// The database column under each criteria label, found once per label rather than per cell.
|
||||
let columns: Vec<Option<usize>> = (0..crit.cols)
|
||||
.map(|c| match crit.get(0, c) {
|
||||
Some(Value::Text(t)) => find_column(db, t),
|
||||
Some(Value::Number(n)) => find_column(db, &gridcraft_core::number_to_text(*n)),
|
||||
_ => None,
|
||||
})
|
||||
.collect();
|
||||
fn matching_records(db: &Array, crit: &Array, ctx: &mut dyn Ctx) -> Vec<usize> {
|
||||
// Per criteria row: its index and a list of (criteria column, database column or None when
|
||||
// the label is unknown, criterion).
|
||||
let mut rows: Vec<CriteriaRow> = Vec::new();
|
||||
let mut tests = 0usize;
|
||||
for r in 1..crit.rows {
|
||||
let mut conds = Vec::new();
|
||||
for c in 0..crit.cols {
|
||||
let Some(cell) = crit.get(r, c) else { continue };
|
||||
let Some(cr) = cell_criterion(cell) else { continue };
|
||||
conds.push((c, columns.get(c).copied().flatten(), cr));
|
||||
let col = match crit.get(0, c) {
|
||||
Some(Value::Text(t)) => find_column(db, t),
|
||||
Some(Value::Number(n)) => find_column(db, &gridcraft_core::number_to_text(*n)),
|
||||
_ => None,
|
||||
};
|
||||
conds.push((c, col, cr));
|
||||
}
|
||||
tests = tests.saturating_add(conds.len().max(1));
|
||||
rows.push((r, conds));
|
||||
}
|
||||
// An empty criteria row matches every record without testing a condition, so nothing is charged.
|
||||
let all_match = rows.is_empty() || rows.iter().any(|(_, conds)| conds.is_empty());
|
||||
// Computed criteria (a formula evaluated per record) count like any other test.
|
||||
if !all_match && db.rows.saturating_sub(1).saturating_mul(tests) > SELECTION_WORK_BUDGET {
|
||||
return Err(CellError::Num);
|
||||
}
|
||||
Ok((1..db.rows)
|
||||
let all_match = rows.is_empty();
|
||||
(1..db.rows)
|
||||
.filter(|&rec| {
|
||||
all_match
|
||||
// Out of work: a computed criterion would fail at once, so records are left out
|
||||
// without scanning the criteria rows; the per-row check still stops mid-scan.
|
||||
|| (!ctx.exhausted()
|
||||
&& rows.iter().any(|(r, conds)| {
|
||||
!ctx.exhausted()
|
||||
&& conds.iter().all(|(c, col, cr)| !ctx.exhausted() && match col {
|
||||
Some(f) => cr.matches(db.get(rec, *f).unwrap_or(&Value::Empty)),
|
||||
// Under a label that names no field, only a formula (a computed
|
||||
// criterion) can hold.
|
||||
None => ctx.computed_criterion(*r, *c, db, rec).unwrap_or(false),
|
||||
})
|
||||
}))
|
||||
|| rows.iter().any(|(r, conds)| {
|
||||
conds.iter().all(|(c, col, cr)| match col {
|
||||
Some(f) => cr.matches(db.get(rec, *f).unwrap_or(&Value::Empty)),
|
||||
// Under a label that names no field, only a formula (a computed
|
||||
// criterion) can hold.
|
||||
None => ctx.computed_criterion(*r, *c, db, rec).unwrap_or(false),
|
||||
})
|
||||
})
|
||||
})
|
||||
.collect())
|
||||
.collect()
|
||||
}
|
||||
|
||||
struct Selection<'a> {
|
||||
@@ -143,7 +123,7 @@ fn select<'a>(a: &'a [Arg], ctx: &mut dyn Ctx) -> R<Selection<'a>> {
|
||||
let db = as_array(dbv);
|
||||
let crit = as_array(critv);
|
||||
let field = field_index(&db, a)?;
|
||||
let records = matching_records(&db, &crit, ctx)?;
|
||||
let records = matching_records(&db, &crit, ctx);
|
||||
Ok(Selection { db, field, records })
|
||||
}
|
||||
|
||||
@@ -372,63 +352,6 @@ mod tests {
|
||||
is_err(d("DSUM", n(9.0)), CellError::Value);
|
||||
}
|
||||
|
||||
/// Computed criteria are no longer asked for once the host is out of work.
|
||||
#[test]
|
||||
fn matching_stops_asking_computed_criteria_when_exhausted() {
|
||||
struct Spent {
|
||||
asked: usize,
|
||||
}
|
||||
impl crate::Ctx for Spent {
|
||||
fn date_system(&self) -> gridcraft_core::DateSystem {
|
||||
gridcraft_core::DateSystem::D1900
|
||||
}
|
||||
fn now_serial(&self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
fn random(&mut self) -> f64 {
|
||||
0.0
|
||||
}
|
||||
fn computed_criterion(&mut self, _: usize, _: usize, _: &gridcraft_core::Array, _: usize) -> Option<bool> {
|
||||
self.asked += 1;
|
||||
Some(true)
|
||||
}
|
||||
fn exhausted(&self) -> bool {
|
||||
self.asked > 0
|
||||
}
|
||||
}
|
||||
let criteria = arr(vec![vec![tv("Formula")], vec![nv(1.0)]]);
|
||||
let mut ctx = Spent { asked: 0 };
|
||||
let spec = crate::lookup("DCOUNT").unwrap();
|
||||
crate::call(spec, &[rf(db()), t("Age"), rf(criteria)], &mut ctx);
|
||||
assert_eq!(ctx.asked, 1);
|
||||
// Within one criteria row, the conditions after the one that spent the work are skipped too.
|
||||
let criteria = arr(vec![vec![tv("Formula"), tv("Other")], vec![nv(1.0), nv(1.0)]]);
|
||||
let mut ctx = Spent { asked: 0 };
|
||||
crate::call(spec, &[rf(db()), t("Age"), rf(criteria)], &mut ctx);
|
||||
assert_eq!(ctx.asked, 1);
|
||||
}
|
||||
|
||||
/// Security: records times criteria conditions past the selection budget is `#NUM!` before
|
||||
/// any record is tested; a nonmatching criteria table used to test every record against
|
||||
/// every criteria row.
|
||||
#[test]
|
||||
fn selection_work_is_bounded() {
|
||||
use gridcraft_core::Array;
|
||||
let column = |label: &str, len: usize, v: Value| {
|
||||
let mut data = vec![tv(label)];
|
||||
data.extend(std::iter::repeat_n(v, len));
|
||||
Value::from(Array::column(data))
|
||||
};
|
||||
let db = column("Tree", 100_000, tv("Apple"));
|
||||
let wide = column("Tree", 5_001, tv("=Plum"));
|
||||
is_err(ev("DCOUNTA", vec![rf(db.clone()), t("Tree"), rf(wide)]), CellError::Num);
|
||||
let narrow = column("Tree", 2, tv("=Apple"));
|
||||
close(ev("DCOUNTA", vec![rf(db.clone()), t("Tree"), rf(narrow)]), 100_000.0);
|
||||
// Empty criteria rows test no condition: the first one matches every record, so they cost nothing.
|
||||
let empty_rows = column("Tree", 5_001, Value::Empty);
|
||||
close(ev("DCOUNTA", vec![rf(db), t("Tree"), rf(empty_rows)]), 100_000.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn simple_criteria() {
|
||||
let c = |cells: Vec<Value>| arr(vec![vec![tv("Tree"), tv("Height")], cells]);
|
||||
|
||||
@@ -416,17 +416,13 @@ fn workday_core(a: &[Arg], c: &mut dyn Ctx, mask: [bool; 7], hol: Option<&Arg>)
|
||||
if days == 0 {
|
||||
return num_val(start as f64);
|
||||
}
|
||||
// The holidays on working days, still sorted (`holidays` sorts them), so each step counts
|
||||
// the ones it passed with two binary searches rather than a scan of every holiday: a run of
|
||||
// consecutive holidays takes one step per holiday.
|
||||
let hs: Vec<i64> = hs.into_iter().filter(|&h| !mask[weekday0(sys, h)]).collect();
|
||||
let dir = days.signum();
|
||||
let mut cur = start;
|
||||
let mut rem = days.abs();
|
||||
while rem > 0 {
|
||||
let next = advance(sys, cur, rem, dir, &mask, per_week);
|
||||
let (lo, hi) = if dir > 0 { (cur + 1, next) } else { (next, cur - 1) };
|
||||
rem = hs.partition_point(|&h| h <= hi).saturating_sub(hs.partition_point(|&h| h < lo)) as i64;
|
||||
rem = hs.iter().filter(|&&h| h >= lo && h <= hi && !mask[weekday0(sys, h)]).count() as i64;
|
||||
cur = next;
|
||||
if !(0..=MAX_SERIAL as i64).contains(&cur) {
|
||||
return Err(CellError::Num);
|
||||
@@ -759,23 +755,6 @@ mod tests {
|
||||
is_err(ev("TIMEVALUE", vec![t("x")]), CellError::Value);
|
||||
}
|
||||
|
||||
/// Security: WORKDAY counts the holidays each step passes by binary search, so a long run of
|
||||
/// consecutive holidays is linear work, not one scan of every holiday per holiday.
|
||||
#[test]
|
||||
fn workday_over_many_consecutive_holidays() {
|
||||
let start = d(2000.0, 1.0, 3.0);
|
||||
let hol: Vec<f64> = (1..=200_000).map(|i| start + f64::from(i)).collect();
|
||||
let out = ev("WORKDAY.INTL", vec![n(start), n(1.0), n(11.0), av(col(&hol))]);
|
||||
// Sundays are the only weekend (11): 200,000 days of holidays are skipped, landing on the
|
||||
// first working day after them.
|
||||
let after = start + 200_001.0;
|
||||
let sunday = ev("WEEKDAY", vec![n(after), n(2.0)]) == nv(7.0);
|
||||
close(out, if sunday { after + 1.0 } else { after });
|
||||
// Holidays on a weekend don't push the result out.
|
||||
let fri = d(2026.0, 10.0, 9.0);
|
||||
close(ev("WORKDAY", vec![n(fri), n(1.0), av(col(&[fri + 1.0, fri + 3.0]))]), fri + 4.0);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn day_counts() {
|
||||
close(ev("DAYS", vec![n(d(2021.0, 3.0, 15.0)), n(d(2021.0, 2.0, 1.0))]), 42.0);
|
||||
|
||||
@@ -171,17 +171,13 @@ fn sort(args: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
let a = if by_col { a.transpose() } else { a };
|
||||
let idx = if has(args, 1) { int_list(&arg(args, 1)?.value)? } else { vec![1] };
|
||||
let ord = if has(args, 2) { int_list(&arg(args, 2)?.value)? } else { vec![1] };
|
||||
if idx.is_empty() || idx.iter().any(|&i| i < 1 || i > a.cols as i64) || ord.iter().any(|&o| o != 1 && o != -1) {
|
||||
if idx.is_empty() || idx.iter().any(|&i| i < 1 || i as usize > a.cols) || ord.iter().any(|&o| o != 1 && o != -1) {
|
||||
return Err(CellError::Value);
|
||||
}
|
||||
if ord.len() > 1 && ord.len() != idx.len() {
|
||||
return Err(CellError::Value);
|
||||
}
|
||||
let orders: Vec<i64> = (0..idx.len()).map(|i| *ord.get(i).or(ord.first()).unwrap_or(&1)).collect();
|
||||
// A column repeated in sort_index can't break a tie its first use left, so only first uses
|
||||
// stay: the keys copied per row are then at most the row itself, however long sort_index is.
|
||||
let mut seen = std::collections::HashSet::new();
|
||||
let (idx, orders): (Vec<i64>, Vec<i64>) = idx.into_iter().zip(orders).filter(|(i, _)| seen.insert(*i)).unzip();
|
||||
let mut lines: Vec<(Vec<Value>, Vec<Value>)> =
|
||||
rows_of(&a).into_iter().map(|r| (idx.iter().map(|&i| r.get(i as usize - 1).cloned().unwrap_or(Value::Empty)).collect(), r)).collect();
|
||||
lines.sort_by(|x, y| cmp_keys(&x.0, &y.0, &orders));
|
||||
@@ -932,16 +928,9 @@ fn pivotby(args: &[Arg], ctx: &mut dyn Ctx) -> R<Value> {
|
||||
let ckeys: Vec<Vec<Value>> = p.rows.iter().map(|&r| key_row(&cf, r)).collect();
|
||||
let rk = distinct_keys(&rkeys, if row_sort.unsigned_abs() as usize <= rf.cols { row_sort } else { 0 });
|
||||
let ck = distinct_keys(&ckeys, if col_sort.unsigned_abs() as usize <= cf.cols { col_sort } else { 0 });
|
||||
// Source rows of every row key, column key and pair, built in one pass, so a total reads
|
||||
// its rows instead of rescanning the whole source once per total.
|
||||
let mut cell: HashMap<(String, String), Vec<usize>> = HashMap::new();
|
||||
let mut by_row: HashMap<String, Vec<usize>> = HashMap::new();
|
||||
let mut by_col: HashMap<String, Vec<usize>> = HashMap::new();
|
||||
for ((r, c), &i) in rkeys.iter().zip(ckeys.iter()).zip(p.rows.iter()) {
|
||||
let (r, c) = (row_key(r), row_key(c));
|
||||
by_row.entry(r.clone()).or_default().push(i);
|
||||
by_col.entry(c.clone()).or_default().push(i);
|
||||
cell.entry((r, c)).or_default().push(i);
|
||||
cell.entry((row_key(r), row_key(c))).or_default().push(i);
|
||||
}
|
||||
let mut agg = |rows: &[usize], c: usize| -> Value {
|
||||
if rows.is_empty() {
|
||||
@@ -951,12 +940,14 @@ fn pivotby(args: &[Arg], ctx: &mut dyn Ctx) -> R<Value> {
|
||||
aggregate_with(f, &vs, ctx)
|
||||
};
|
||||
let rows_for = |r: Option<&Vec<Value>>, c: Option<&Vec<Value>>| -> Vec<usize> {
|
||||
match (r, c) {
|
||||
(Some(r), Some(c)) => cell.get(&(row_key(r), row_key(c))).cloned().unwrap_or_default(),
|
||||
(Some(r), None) => by_row.get(&row_key(r)).cloned().unwrap_or_default(),
|
||||
(None, Some(c)) => by_col.get(&row_key(c)).cloned().unwrap_or_default(),
|
||||
(None, None) => p.rows.clone(),
|
||||
}
|
||||
let rkey = r.map(|k| row_key(k));
|
||||
let ckey = c.map(|k| row_key(k));
|
||||
p.rows
|
||||
.iter()
|
||||
.zip(rkeys.iter().zip(ckeys.iter()))
|
||||
.filter(|(_, (a, b))| rkey.as_ref().is_none_or(|k| *k == row_key(a)) && ckey.as_ref().is_none_or(|k| *k == row_key(b)))
|
||||
.map(|(&i, _)| i)
|
||||
.collect()
|
||||
};
|
||||
let mut cols: Vec<Option<&Vec<Value>>> = ck.iter().map(Some).collect();
|
||||
if col_total > 0 {
|
||||
@@ -1006,7 +997,10 @@ fn pivotby(args: &[Arg], ctx: &mut dyn Ctx) -> R<Value> {
|
||||
}
|
||||
};
|
||||
for c in &cols {
|
||||
let rows = rows_for(r, *c);
|
||||
let rows = match (r, c) {
|
||||
(Some(rk), Some(ck)) => cell.get(&(row_key(rk), row_key(ck))).cloned().unwrap_or_default(),
|
||||
_ => rows_for(r, *c),
|
||||
};
|
||||
for vc in 0..vals.cols {
|
||||
line.push(agg(&rows, vc));
|
||||
}
|
||||
@@ -1210,21 +1204,6 @@ mod tests {
|
||||
assert_eq!(rows_of(&r), vec![vec![nv(0.0), nv(5.0)], vec![nv(1.0), nv(2.0)], vec![nv(1.0), nv(1.0)]]);
|
||||
}
|
||||
|
||||
/// Security: a column repeated in sort_index keys the sort once. The keys copied per row stay
|
||||
/// within the row, however long sort_index is, and a repeat (even with the other order) can't
|
||||
/// change the result, since its first use already ordered the rows.
|
||||
#[test]
|
||||
fn sort_drops_repeated_sort_indices() {
|
||||
let m = || av(arr(vec![vec![nv(1.0), nv(2.0)], vec![nv(1.0), nv(1.0)], vec![nv(0.0), nv(5.0)]]));
|
||||
let want = vec![vec![nv(0.0), nv(5.0)], vec![nv(1.0), nv(2.0)], vec![nv(1.0), nv(1.0)]];
|
||||
let r = ev("SORT", vec![m(), av(row(&[1.0, 1.0, 2.0, 1.0])), av(row(&[1.0, -1.0, -1.0, 1.0]))]);
|
||||
assert_eq!(rows_of(&r), want);
|
||||
let idx = vec![1.0; 1_000_000];
|
||||
let tall = av(gridcraft_core::Array::filled(2_000, 1, nv(1.0)).into());
|
||||
let r = ev("SORT", vec![tall, av(row(&idx))]);
|
||||
assert_eq!(rows_of(&r).len(), 2_000);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn sortby() {
|
||||
let r = ev("SORTBY", vec![av(col(&[1.0, 2.0, 3.0])), av(col(&[30.0, 10.0, 20.0]))]);
|
||||
@@ -1468,16 +1447,6 @@ mod tests {
|
||||
);
|
||||
}
|
||||
|
||||
/// Security: PIVOTBY totals read their source rows from indexes built in the grouping pass,
|
||||
/// rather than rescanning the source once per total, so many distinct keys stay linear.
|
||||
#[test]
|
||||
fn pivotby_totals_with_many_groups() {
|
||||
let keys: Vec<f64> = (0..20_000).map(f64::from).collect();
|
||||
let r = rows_of(&ev("PIVOTBY", vec![av(col(&keys)), av(col(&vec![1.0; 20_000])), av(col(&vec![2.0; 20_000])), t("SUM")]));
|
||||
assert_eq!(r.get(1), Some(&vec![nv(0.0), nv(2.0), nv(2.0)]));
|
||||
assert_eq!(r.last(), Some(&vec![tv("Total"), nv(40_000.0), nv(40_000.0)]));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn pivotby_size_check_counts_column_header_rows() {
|
||||
// One row, n column fields and n value columns: n header rows + value-name row + body row + total row,
|
||||
|
||||
@@ -13,7 +13,6 @@ mod day_count;
|
||||
mod depreciation;
|
||||
mod securities;
|
||||
mod tvm;
|
||||
pub(crate) use tvm::cumulative_args;
|
||||
|
||||
use day_count::{coupdaybs, coupdays, coupdaysnc, coupncd, coupnum, couppcd};
|
||||
use depreciation::{amordegrc, amorlinc, db, ddb, sln, syd, vdb};
|
||||
|
||||
@@ -116,9 +116,7 @@ pub(super) fn ppmt(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
num_val(pmt_calc(rate, np, pv, fv, typ) - ipmt_calc(rate, per, np, pv, fv, typ))
|
||||
}
|
||||
|
||||
/// The validated (rate, periods, pv, start, end, type) of CUMIPMT/CUMPRINC: every check that
|
||||
/// refuses a call before its period loop, shared with the lifted-call work preflight.
|
||||
pub(crate) fn cumulative_args(a: &[Arg]) -> R<(f64, f64, f64, f64, f64, f64)> {
|
||||
fn cumulative(a: &[Arg], principal: bool) -> R<Value> {
|
||||
let (rate, np, pv) = (num(a, 0)?, num(a, 1)?, num(a, 2)?);
|
||||
let (start, end, typ) = (num(a, 3)?.trunc(), num(a, 4)?.trunc(), num(a, 5)?);
|
||||
let np_i = np.trunc();
|
||||
@@ -128,11 +126,6 @@ pub(crate) fn cumulative_args(a: &[Arg]) -> R<(f64, f64, f64, f64, f64, f64)> {
|
||||
if end - start > 1e7 || np_i > 1e15 {
|
||||
return Err(CellError::Num);
|
||||
}
|
||||
Ok((rate, np_i, pv, start, end, typ))
|
||||
}
|
||||
|
||||
fn cumulative(a: &[Arg], principal: bool) -> R<Value> {
|
||||
let (rate, np_i, pv, start, end, typ) = cumulative_args(a)?;
|
||||
let p = pmt_calc(rate, np_i, pv, 0.0, typ);
|
||||
let mut sum = 0.0;
|
||||
let mut per = start;
|
||||
|
||||
@@ -91,7 +91,7 @@ fn na(_a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
fn n_fn(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
// N(range) reads the range's first cell, but a computed array (N(A1:A3>0), N({...})) is
|
||||
// converted element by element, so N can stand in for `--` inside SUMPRODUCT.
|
||||
if let Some(Arg { value: Value::Array(arr), from_ref: false, .. }) = a.first() {
|
||||
if let Some(Arg { value: Value::Array(arr), from_ref: false }) = a.first() {
|
||||
let data = arr.data.iter().map(|v| n_of(v.clone()).unwrap_or_else(Value::Error)).collect();
|
||||
return Array::new(arr.rows, arr.cols, data).map(Value::from).ok_or(CellError::Calc);
|
||||
}
|
||||
|
||||
@@ -87,19 +87,16 @@ pub use stat::{aggregate_values, aggregate_values_k};
|
||||
pub struct Arg {
|
||||
pub value: Value,
|
||||
pub from_ref: bool,
|
||||
/// A written argument whose value came out blank (a lookup of an empty cell): given, not
|
||||
/// omitted, although its value is `Value::Empty`.
|
||||
pub blank_value: bool,
|
||||
}
|
||||
|
||||
impl Arg {
|
||||
/// A literal (non-reference) argument.
|
||||
pub fn val(value: impl Into<Value>) -> Arg {
|
||||
Arg { value: value.into(), from_ref: false, blank_value: false }
|
||||
Arg { value: value.into(), from_ref: false }
|
||||
}
|
||||
/// An argument that came from a cell or range reference.
|
||||
pub fn reference(value: impl Into<Value>) -> Arg {
|
||||
Arg { value: value.into(), from_ref: true, blank_value: false }
|
||||
Arg { value: value.into(), from_ref: true }
|
||||
}
|
||||
}
|
||||
|
||||
@@ -116,11 +113,6 @@ pub trait Ctx {
|
||||
fn computed_criterion(&mut self, _row: usize, _col: usize, _db: &Array, _record: usize) -> Option<bool> {
|
||||
None
|
||||
}
|
||||
/// Whether the cell's work budget has run out: the database functions stop asking for computed
|
||||
/// criteria then, as every one would fail at once.
|
||||
fn exhausted(&self) -> bool {
|
||||
false
|
||||
}
|
||||
/// For GROUPBY/PIVOTBY: whether their function argument is a LAMBDA (its value is then not used).
|
||||
fn has_lambda(&self) -> bool {
|
||||
false
|
||||
@@ -316,12 +308,6 @@ mod tests {
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("2026-10-22"), t("(\\d+)-(\\d+)-(\\d+)"), t("$3/$2/$1")]), Value::text("22/10/2026"));
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("abc"), t("x*"), t("-")]), Value::text("-a-b-c-"));
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("a1b2"), t("[0-9]"), t("#"), n(2.0)]), Value::text("a1b#"));
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("a1b2"), t("[0-9]"), t("#"), n(-2.0)]), Value::text("a#b2"));
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("a1b2"), t("[0-9]"), t("#"), n(-3.0)]), Value::text("a1b2"));
|
||||
// 2^32 is no occurrence the text has (and on 32-bit targets it must not wrap to 0).
|
||||
for occurrence in [4_294_967_296.0, -4_294_967_296.0] {
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t("a1b2"), t("[0-9]"), t("#"), n(occurrence)]), Value::text("a1b2"));
|
||||
}
|
||||
// One match whose replacement repeats it 10,000 times: refused before it is built.
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t(&many), t("^.*$"), t(&"$0".repeat(10_000))]), Value::Error(CellError::Value));
|
||||
assert_eq!(ev("REGEXREPLACE", vec![t(&many), t("^.*$"), t(&"$0".repeat(10_000)), n(1.0)]), Value::Error(CellError::Value));
|
||||
|
||||
+30
-196
@@ -50,7 +50,12 @@ fn reject_multi_cell_refs(spec: &FnSpec, args: &[Arg]) -> Option<Vec<Arg>> {
|
||||
if !args.iter().enumerate().any(|(i, a)| multi(i, a)) {
|
||||
return None;
|
||||
}
|
||||
Some(args.iter().enumerate().map(|(i, a)| if multi(i, a) { Arg::val(Value::Error(CellError::Value)) } else { a.clone() }).collect())
|
||||
Some(
|
||||
args.iter()
|
||||
.enumerate()
|
||||
.map(|(i, a)| if multi(i, a) { Arg { value: Value::Error(CellError::Value), from_ref: false } } else { a.clone() })
|
||||
.collect(),
|
||||
)
|
||||
}
|
||||
|
||||
/// Lookups whose every call searches a lookup vector from its start, so a call lifted over `n`
|
||||
@@ -63,86 +68,26 @@ const SCANS: &[&str] = &["VLOOKUP", "HLOOKUP", "LOOKUP", "MATCH", "XLOOKUP", "XM
|
||||
/// `XMATCH(SEQUENCE(1048576), -SEQUENCE(1048576))` runs for hours and hangs opening the file.
|
||||
/// Past the budget the call is `#NUM!`, as MMULT is past 2e9 multiply-adds. 5e8 keeps the worst
|
||||
/// case to a few seconds and still admits 5,000 needles over a 100,000-row table.
|
||||
///
|
||||
/// The same budget covers all the work of a lifted call added up across its elements: criteria
|
||||
/// aggregates rescanning their ranges, period loops, and array results built per element.
|
||||
const LIFTED_SCAN_BUDGET: usize = 500_000_000;
|
||||
|
||||
/// Criteria aggregates: every call tests each cell of its ranges against its criteria, so a call
|
||||
/// lifted over `n` criteria does `n` x (range cells) work, like the [`SCANS`] lookups.
|
||||
const CRITERIA: &[&str] = &["COUNTIF", "COUNTIFS", "SUMIF", "SUMIFS", "AVERAGEIF", "AVERAGEIFS", "MAXIFS", "MINIFS"];
|
||||
|
||||
/// Cumulative TVM functions that step through every period from `start_period` (argument 3) to
|
||||
/// `end_period` (argument 4) on each call.
|
||||
const PERIOD_LOOPS: &[&str] = &["CUMIPMT", "CUMPRINC"];
|
||||
|
||||
/// Budget units one period of a [`PERIOD_LOOPS`] call costs: each period computes a payment and a
|
||||
/// balance with powers, many times a scanned cell.
|
||||
const PERIOD_COST: usize = 16;
|
||||
|
||||
/// The work every lifted call of `spec` repeats over its non-lifted arguments: the longest vector
|
||||
/// a [`SCANS`] lookup searches, or every cell a [`CRITERIA`] aggregate tests. 0 for other
|
||||
/// functions.
|
||||
/// The length of the longest vector a lifted call of `spec` searches per needle: 0 unless `spec`
|
||||
/// is one of the [`SCANS`] lookups and a non-lifted argument is an array.
|
||||
fn scan_len(spec: &FnSpec, args: &[Arg]) -> usize {
|
||||
let scans = SCANS.contains(&spec.name);
|
||||
if !scans && !CRITERIA.contains(&spec.name) {
|
||||
if !SCANS.contains(&spec.name) {
|
||||
return 0;
|
||||
}
|
||||
// SUMIF and AVERAGEIF read the optional third range only at the coordinates of the first.
|
||||
let first = args.first().and_then(|a| if let Value::Array(arr) = &a.value { Some(arr.data.len()) } else { None });
|
||||
let lens = args.iter().enumerate().filter(|(i, _)| !spec.is_scalar_arg(*i)).filter_map(|(i, a)| match &a.value {
|
||||
Value::Array(arr) if scans => Some(arr.rows.max(arr.cols)),
|
||||
Value::Array(arr) if i == 2 && matches!(spec.name, "SUMIF" | "AVERAGEIF") => Some(arr.data.len().min(first.unwrap_or(usize::MAX))),
|
||||
Value::Array(arr) => Some(arr.data.len()),
|
||||
_ => None,
|
||||
});
|
||||
if scans { lens.max().unwrap_or(0) } else { lens.fold(0, usize::saturating_add) }
|
||||
}
|
||||
|
||||
/// The work of one lifted call that depends on its own (lifted) scalar arguments: the number of
|
||||
/// periods a [`PERIOD_LOOPS`] function steps through. 0 for other functions.
|
||||
fn call_cost(spec: &FnSpec, cell_args: &[Arg]) -> usize {
|
||||
if !PERIOD_LOOPS.contains(&spec.name) {
|
||||
return 0;
|
||||
}
|
||||
// Coercing numeric text reads all of it, even in a call the function then refuses, so each
|
||||
// character is charged as well: of the arguments it coerces in order, up to the first that fails.
|
||||
let mut parse = 0usize;
|
||||
for (i, a) in cell_args.iter().enumerate().take(6) {
|
||||
if let Value::Text(t) = &a.value {
|
||||
parse = parse.saturating_add(t.len());
|
||||
}
|
||||
if crate::util::num(cell_args, i).is_err() {
|
||||
break;
|
||||
}
|
||||
}
|
||||
// The function's own validation and coercion (text like "10000000" counts), so a call it
|
||||
// refuses before its loop costs nothing and no loop goes uncharged.
|
||||
let periods = crate::financial::cumulative_args(cell_args).map_or(0, |(.., start, end, _)| (end - start + 1.0) as usize);
|
||||
parse.saturating_add(periods.saturating_mul(PERIOD_COST))
|
||||
args.iter()
|
||||
.enumerate()
|
||||
.filter(|(i, _)| !spec.is_scalar_arg(*i))
|
||||
.filter_map(|(_, a)| match &a.value {
|
||||
Value::Array(arr) => Some(arr.rows.max(arr.cols)),
|
||||
_ => None,
|
||||
})
|
||||
.max()
|
||||
.unwrap_or(0)
|
||||
}
|
||||
|
||||
pub(crate) fn call_lifted(spec: &FnSpec, args: &[Arg], ctx: &mut dyn Ctx) -> Value {
|
||||
call_lifted_within(spec, args, ctx, LIFTED_SCAN_BUDGET)
|
||||
}
|
||||
|
||||
/// Sets the lifted arguments of `cell_args` to their elements at (`r`, `c`).
|
||||
fn element_args(spec: &FnSpec, args: &[Arg], r: usize, c: usize, cell_args: &mut [Arg]) {
|
||||
for (i, a) in args.iter().enumerate() {
|
||||
if spec.is_scalar_arg(i)
|
||||
&& let Value::Array(arr) = &a.value
|
||||
&& let Some(slot) = cell_args.get_mut(i)
|
||||
{
|
||||
let v = arr.get_broadcast(r, c);
|
||||
slot.value = if let Value::Array(inner) = &v { inner.data.first().cloned().unwrap_or(Value::Empty) } else { v };
|
||||
// A blank element of a written array argument is given, as a scalar blank is.
|
||||
slot.blank_value = !slot.from_ref && matches!(slot.value, Value::Empty);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// [`call_lifted`] with every call's work added up against `budget` (see [`LIFTED_SCAN_BUDGET`]).
|
||||
fn call_lifted_within(spec: &FnSpec, args: &[Arg], ctx: &mut dyn Ctx, budget: usize) -> Value {
|
||||
let rejected = if spec.value_args.is_empty() { None } else { reject_multi_cell_refs(spec, args) };
|
||||
let args = rejected.as_deref().unwrap_or(args);
|
||||
let mut rows = 0usize;
|
||||
@@ -161,42 +106,25 @@ fn call_lifted_within(spec: &FnSpec, args: &[Arg], ctx: &mut dyn Ctx, budget: us
|
||||
return (spec.imp)(args, ctx);
|
||||
}
|
||||
let needles = rows.saturating_mul(cols);
|
||||
// Every call is charged its repeated scan (at least 1), its own period loop, and the cells of
|
||||
// an array result (built in full, though only its first cell is kept), so lifting a scalar
|
||||
// argument can't multiply what one call may do into hours of work:
|
||||
// `COUNTIF(big_range, SEQUENCE(1e4))`, `CUMIPMT(0.1, 1e7, 1, 1, SEQUENCE(1e4,,1e7,0), 0)`,
|
||||
// `MUNIT(SEQUENCE(1e4,,3000,0))`.
|
||||
let per_call = scan_len(spec, args).max(1);
|
||||
let mut cell_args: Vec<Arg> = args.to_vec();
|
||||
let mut spent = needles.saturating_mul(per_call);
|
||||
if needles > MAX_CELLS || spent > budget {
|
||||
if needles > MAX_CELLS || needles.saturating_mul(scan_len(spec, args)) > LIFTED_SCAN_BUDGET {
|
||||
return Value::Error(CellError::Num);
|
||||
}
|
||||
// Period loops are known before any call runs, so a call over budget does none of them.
|
||||
if PERIOD_LOOPS.contains(&spec.name) {
|
||||
for r in 0..rows {
|
||||
for c in 0..cols {
|
||||
element_args(spec, args, r, c, &mut cell_args);
|
||||
spent = spent.saturating_add(call_cost(spec, &cell_args));
|
||||
if spent > budget {
|
||||
return Value::Error(CellError::Num);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
let mut cell_args: Vec<Arg> = args.to_vec();
|
||||
let mut data = Vec::with_capacity(rows * cols);
|
||||
for r in 0..rows {
|
||||
for c in 0..cols {
|
||||
element_args(spec, args, r, c, &mut cell_args);
|
||||
for (i, a) in args.iter().enumerate() {
|
||||
if spec.is_scalar_arg(i)
|
||||
&& let Value::Array(arr) = &a.value
|
||||
&& let Some(slot) = cell_args.get_mut(i)
|
||||
{
|
||||
let v = arr.get_broadcast(r, c);
|
||||
slot.value = if let Value::Array(inner) = &v { inner.data.first().cloned().unwrap_or(Value::Empty) } else { v };
|
||||
}
|
||||
}
|
||||
let out = (spec.imp)(&cell_args, ctx);
|
||||
data.push(match out {
|
||||
Value::Array(a) => {
|
||||
spent = spent.saturating_add(a.data.len());
|
||||
if spent > budget {
|
||||
return Value::Error(CellError::Num);
|
||||
}
|
||||
a.data.first().cloned().unwrap_or(Value::Empty)
|
||||
}
|
||||
Value::Array(a) => a.data.first().cloned().unwrap_or(Value::Empty),
|
||||
v => v,
|
||||
});
|
||||
}
|
||||
@@ -281,100 +209,6 @@ mod tests {
|
||||
assert_eq!(rows_of(&out), vec![vec![nv(-1.0)]; 1000]);
|
||||
}
|
||||
|
||||
/// Security: lifting a criteria aggregate rescans its range for every criterion, so criteria
|
||||
/// x range cells past the budget is `#NUM!` before any scan runs.
|
||||
#[test]
|
||||
fn lifted_criteria_aggregates_stop_past_their_work_budget() {
|
||||
use gridcraft_core::Array;
|
||||
let tall = |n: usize| av(Array::filled(n, 1, nv(-1.0)).into());
|
||||
is_err(ev("COUNTIF", vec![tall(1_000_000), tall(501)]), CellError::Num);
|
||||
is_err(ev("SUMIF", vec![tall(1_000_000), tall(501)]), CellError::Num);
|
||||
is_err(ev("COUNTIFS", vec![tall(1_000_000), tall(501)]), CellError::Num);
|
||||
// Two ranges of 250,000 cells each count together.
|
||||
is_err(ev("SUMIFS", vec![tall(250_000), tall(250_000), tall(1001)]), CellError::Num);
|
||||
// SUMIF reads its sum range only where the criteria range has cells: a one-cell range over a
|
||||
// million-cell sum range is one cell of work per criterion.
|
||||
let one = av(Array::filled(1, 1, nv(-1.0)).into());
|
||||
assert_eq!(rows_of(&ev("SUMIF", vec![one, tall(501), tall(1_000_000)])).len(), 501);
|
||||
let counts = ev("COUNTIF", vec![av(col(&[1.0, 2.0, 2.0])), av(col(&[2.0, 1.0, 5.0]))]);
|
||||
assert_eq!(rows_of(&counts), vec![vec![nv(2.0)], vec![nv(1.0)], vec![nv(0.0)]]);
|
||||
for name in super::CRITERIA.iter().chain(super::PERIOD_LOOPS) {
|
||||
assert!(crate::lookup(name).is_some(), "{name} isn't registered");
|
||||
}
|
||||
}
|
||||
|
||||
/// Security: CUMIPMT and CUMPRINC loop over their periods once per lifted element; the
|
||||
/// periods of all elements are added up first, so an over-budget call runs none of them.
|
||||
#[test]
|
||||
fn lifted_period_loops_stop_past_their_work_budget() {
|
||||
use gridcraft_core::Array;
|
||||
let ends = |len: usize, end: f64| av(Array::filled(len, 1, nv(end)).into());
|
||||
let args = |len: usize, end: f64| vec![n(0.01), n(1e7), n(1000.0), n(1.0), ends(len, end), n(0.0)];
|
||||
// 51 elements of 1e7 periods each is past 5e8.
|
||||
is_err(ev("CUMIPMT", args(51, 1e7)), CellError::Num);
|
||||
is_err(ev("CUMPRINC", args(51, 1e7)), CellError::Num);
|
||||
// A period costs more than a scanned cell, so fewer elements are enough.
|
||||
is_err(ev("CUMIPMT", args(49, 1e7)), CellError::Num);
|
||||
// Numeric text is coerced by the function, so its periods are charged the same.
|
||||
let text_ends = av(Array::filled(51, 1, Value::text("10000000")).into());
|
||||
is_err(ev("CUMIPMT", vec![n(0.01), n(1e7), n(1000.0), n(1.0), text_ends, n(0.0)]), CellError::Num);
|
||||
let out = ev("CUMIPMT", args(2, 12.0));
|
||||
assert_eq!(rows_of(&out).len(), 2);
|
||||
assert!(matches!(rows_of(&out).first().and_then(|r| r.first()), Some(Value::Number(x)) if *x < 0.0));
|
||||
}
|
||||
|
||||
/// Security: numeric text is parsed once per lifted element even when the call is refused; the
|
||||
/// characters read are charged, so a long text repeated down a column can't cost gigabytes of parsing.
|
||||
#[test]
|
||||
fn lifted_period_loops_charge_the_text_they_parse() {
|
||||
use gridcraft_core::Array;
|
||||
// 20,000 elements x 32,000 characters is past 5e8 though no call reaches its loop (rate is negative).
|
||||
let long = Value::text(format!("{}5", "0".repeat(32_000)));
|
||||
let ends = av(Array::filled(20_000, 1, long).into());
|
||||
is_err(ev("CUMIPMT", vec![n(-0.01), n(10.0), n(1000.0), n(1.0), ends, n(0.0)]), CellError::Num);
|
||||
// A short text period costs next to nothing.
|
||||
let ends = av(Array::filled(3, 1, Value::text("12")).into());
|
||||
assert_eq!(rows_of(&ev("CUMIPMT", vec![n(0.01), n(20.0), n(1000.0), n(1.0), ends, n(0.0)])).len(), 3);
|
||||
}
|
||||
|
||||
/// Text after the first argument that fails to coerce is never read, so it is not charged: the
|
||||
/// elements keep their own errors instead of one budget `#NUM!`.
|
||||
#[test]
|
||||
fn lifted_period_loops_charge_only_the_text_they_reach() {
|
||||
use gridcraft_core::Array;
|
||||
let rates = av(Array::filled(20_000, 1, Value::text("x")).into());
|
||||
let long = Value::text(format!("{}5", "0".repeat(32_000)));
|
||||
let out = ev("CUMIPMT", vec![rates, n(10.0), n(1000.0), n(1.0), av(long), n(0.0)]);
|
||||
assert!(rows_of(&out).iter().all(|r| r.first() == Some(&Value::Error(CellError::Value))));
|
||||
}
|
||||
|
||||
/// A call the function refuses before its loop (here: end period past the period count) is
|
||||
/// not charged, so rejected elements can't push a valid one over the budget.
|
||||
#[test]
|
||||
fn lifted_period_loops_do_not_charge_rejected_calls() {
|
||||
use gridcraft_core::Array;
|
||||
let mut ends = vec![nv(1e7 + 1.0); 60];
|
||||
ends.push(nv(12.0));
|
||||
let ends = av(Array::new(ends.len(), 1, ends).unwrap_or_else(|| panic!("shape")).into());
|
||||
let out = ev("CUMIPMT", vec![n(0.01), n(1e7), n(1000.0), n(1.0), ends, n(0.0)]);
|
||||
let rows = rows_of(&out);
|
||||
assert_eq!(rows.len(), 61);
|
||||
assert!(matches!(rows.last().and_then(|r| r.first()), Some(Value::Number(x)) if *x < 0.0));
|
||||
}
|
||||
|
||||
/// Security: an array-producing function lifted over a computed array builds each element's
|
||||
/// whole result before keeping one cell; their cells count against the budget.
|
||||
#[test]
|
||||
fn lifted_array_results_count_against_the_budget() {
|
||||
use gridcraft_core::Array;
|
||||
let spec = crate::lookup("MUNIT").unwrap_or_else(|| panic!("MUNIT isn't registered"));
|
||||
let dims = |n: usize| [av(Array::filled(n, 1, nv(10.0)).into())];
|
||||
// Ten 10x10 identity matrices are 1,000 cells.
|
||||
let out = super::call_lifted_within(spec, &dims(10), &mut TestCtx::default(), 1_010);
|
||||
assert_eq!(rows_of(&out), vec![vec![nv(1.0)]; 10]);
|
||||
is_err(super::call_lifted_within(spec, &dims(11), &mut TestCtx::default(), 1_010), CellError::Num);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn value_params_reject_multi_cell_references() {
|
||||
let s = FnSpec { value_args: &[true, false], ..spec() };
|
||||
|
||||
@@ -132,6 +132,11 @@ fn row(a: &Array, r: usize) -> Vec<Value> {
|
||||
(0..a.cols).map(|c| a.get(r, c).cloned().unwrap_or(Value::Empty)).collect()
|
||||
}
|
||||
|
||||
/// A looked-up cell value: blanks read as 0 like a reference to an empty cell.
|
||||
fn cell_result(v: Value) -> Value {
|
||||
if matches!(v, Value::Empty) { Value::Number(0.0) } else { v }
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------------
|
||||
// VLOOKUP / HLOOKUP / LOOKUP / MATCH
|
||||
|
||||
@@ -155,9 +160,7 @@ fn vh_lookup(args: &[Arg], vertical: bool) -> R<Value> {
|
||||
let pos = if approx { bsearch_last(&x, &keys, false) } else { find_exact(&x, &keys, true, false) };
|
||||
let pos = pos.ok_or(CellError::NA)?;
|
||||
let v = if vertical { table.get(pos, idx) } else { table.get(idx, pos) };
|
||||
// An empty target stays blank, as in Excel: ISBLANK sees it as blank and `&` /
|
||||
// CONCATENATE treat it as ""; only a formula's final result displays 0 (#374).
|
||||
Ok(v.cloned().unwrap_or(Value::Empty))
|
||||
Ok(cell_result(v.cloned().unwrap_or(Value::Empty)))
|
||||
}
|
||||
|
||||
fn vlookup(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
@@ -190,18 +193,18 @@ fn lookup(args: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
} else {
|
||||
None
|
||||
};
|
||||
return v.ok_or(CellError::NA);
|
||||
return v.map(cell_result).ok_or(CellError::NA);
|
||||
}
|
||||
// Array form: search the first row of a wide array, else the first column; return the
|
||||
// matching entry from the last row / column.
|
||||
if la.cols > la.rows {
|
||||
let keys = row(&la, 0);
|
||||
let pos = bsearch_last(&x, &keys, false).ok_or(CellError::NA)?;
|
||||
Ok(la.get(la.rows - 1, pos).cloned().unwrap_or(Value::Empty))
|
||||
Ok(cell_result(la.get(la.rows - 1, pos).cloned().unwrap_or(Value::Empty)))
|
||||
} else {
|
||||
let keys = column(&la, 0);
|
||||
let pos = bsearch_last(&x, &keys, false).ok_or(CellError::NA)?;
|
||||
Ok(la.get(pos, la.cols - 1).cloned().unwrap_or(Value::Empty))
|
||||
Ok(cell_result(la.get(pos, la.cols - 1).cloned().unwrap_or(Value::Empty)))
|
||||
}
|
||||
}
|
||||
|
||||
@@ -322,13 +325,13 @@ fn xlookup(args: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
if vertical {
|
||||
let r = row(&ra, p);
|
||||
if r.len() == 1 {
|
||||
return Ok(r.into_iter().next().unwrap_or(Value::Empty));
|
||||
return Ok(cell_result(r.into_iter().next().unwrap_or(Value::Empty)));
|
||||
}
|
||||
array_val(1, r.len(), r)
|
||||
} else {
|
||||
let c = column(&ra, p);
|
||||
if c.len() == 1 {
|
||||
return Ok(c.into_iter().next().unwrap_or(Value::Empty));
|
||||
return Ok(cell_result(c.into_iter().next().unwrap_or(Value::Empty)));
|
||||
}
|
||||
array_val(c.len(), 1, c)
|
||||
}
|
||||
@@ -592,8 +595,7 @@ mod tests {
|
||||
is_err(ev("VLOOKUP", vec![n(4.0), rf(table()), n(2.0), b(false)]), CellError::NA);
|
||||
is_err(ev("VLOOKUP", vec![n(4.0), rf(table()), n(2.0), empty()]), CellError::NA);
|
||||
close(ev("VLOOKUP", vec![n(3.0), rf(table()), n(3.0), b(false)]), 30.0);
|
||||
// An empty target stays blank (the cell displays 0 only as a final result) (#374).
|
||||
assert_eq!(ev("VLOOKUP", vec![n(5.0), rf(table()), n(3.0), b(false)]), Value::Empty);
|
||||
close(ev("VLOOKUP", vec![n(5.0), rf(table()), n(3.0), b(false)]), 0.0);
|
||||
is_err(ev("VLOOKUP", vec![n(3.0), rf(table()), n(4.0)]), CellError::Ref);
|
||||
is_err(ev("VLOOKUP", vec![n(3.0), rf(table()), n(0.0)]), CellError::Value);
|
||||
close(ev("VLOOKUP", vec![t("b*"), rf(names()), n(2.0), b(false)]), 2.0);
|
||||
|
||||
@@ -30,27 +30,17 @@ fn check_len(s: String) -> R<Value> {
|
||||
|
||||
/// Pushes `piece` onto `out`, failing once the result exceeds the cell limit.
|
||||
fn push_capped(out: &mut String, piece: &str) -> R<()> {
|
||||
// Checked before copying, so one oversized piece is refused without being appended.
|
||||
if out.len().saturating_add(piece.len()) > MAX_TEXT * 4 {
|
||||
out.push_str(piece);
|
||||
if out.len() > MAX_TEXT * 4 {
|
||||
return Err(CellError::Value);
|
||||
}
|
||||
out.push_str(piece);
|
||||
Ok(())
|
||||
}
|
||||
|
||||
/// `v.to_text()` without copying text values.
|
||||
fn piece(v: &Value) -> R<std::borrow::Cow<'_, str>> {
|
||||
match v {
|
||||
Value::Text(t) => Ok(std::borrow::Cow::Borrowed(t)),
|
||||
Value::Array(a) => a.data.first().map_or(Ok(std::borrow::Cow::Borrowed("")), piece),
|
||||
_ => v.to_text().map(std::borrow::Cow::Owned),
|
||||
}
|
||||
}
|
||||
|
||||
fn concatenate(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
let mut out = String::new();
|
||||
for x in a {
|
||||
push_capped(&mut out, &piece(&scalar(x))?)?;
|
||||
push_capped(&mut out, &scalar(x).to_text()?)?;
|
||||
}
|
||||
check_len(out)
|
||||
}
|
||||
@@ -59,49 +49,31 @@ fn concat(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
let mut out = String::new();
|
||||
for x in a {
|
||||
for v in as_array(&x.value).iter() {
|
||||
push_capped(&mut out, &piece(v)?)?;
|
||||
push_capped(&mut out, &v.to_text()?)?;
|
||||
}
|
||||
}
|
||||
check_len(out)
|
||||
}
|
||||
|
||||
/// The error `v.to_text()` would give, found without building any text.
|
||||
fn text_error(v: &Value) -> Option<CellError> {
|
||||
match v {
|
||||
Value::Error(e) => Some(*e),
|
||||
Value::Array(a) => a.data.first().and_then(text_error),
|
||||
_ => None,
|
||||
}
|
||||
}
|
||||
|
||||
fn textjoin(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
// Errors are found first, in argument order, so they win as before; text is then built one
|
||||
// piece at a time into the capped output, so many references to one long text (or a long
|
||||
// delimiter list) stop at the cap instead of first copying every piece.
|
||||
let delims = as_array(&arg(a, 0)?.value);
|
||||
if let Some(e) = delims.iter().find_map(text_error) {
|
||||
return Err(e);
|
||||
}
|
||||
let delims: Vec<String> = as_array(&arg(a, 0)?.value).iter().map(|v| v.to_text()).collect::<R<_>>()?;
|
||||
let ignore = boolean(a, 1)?;
|
||||
let items: Vec<_> = a.iter().skip(2).map(|x| as_array(&x.value)).collect();
|
||||
if let Some(e) = items.iter().flat_map(|x| x.iter()).find_map(text_error) {
|
||||
return Err(e);
|
||||
let mut items = Vec::new();
|
||||
for x in a.iter().skip(2) {
|
||||
for v in as_array(&x.value).iter() {
|
||||
let s = v.to_text()?;
|
||||
if ignore && s.is_empty() {
|
||||
continue;
|
||||
}
|
||||
items.push(s);
|
||||
}
|
||||
}
|
||||
let mut out = String::new();
|
||||
let mut joined = 0usize;
|
||||
for v in items.iter().flat_map(|x| x.iter()) {
|
||||
let s = piece(v)?;
|
||||
if ignore && s.is_empty() {
|
||||
continue;
|
||||
for (i, s) in items.iter().enumerate() {
|
||||
if i > 0 && !delims.is_empty() {
|
||||
push_capped(&mut out, &delims[(i - 1) % delims.len()])?;
|
||||
}
|
||||
if joined > 0
|
||||
&& !delims.data.is_empty()
|
||||
&& let Some(d) = delims.data.get((joined - 1) % delims.data.len())
|
||||
{
|
||||
push_capped(&mut out, &piece(d)?)?;
|
||||
}
|
||||
push_capped(&mut out, &s)?;
|
||||
joined += 1;
|
||||
push_capped(&mut out, s)?;
|
||||
}
|
||||
check_len(out)
|
||||
}
|
||||
@@ -942,25 +914,13 @@ fn regexreplace(a: &[Arg], _c: &mut dyn Ctx) -> R<Value> {
|
||||
out.push_str(s.get(last..).unwrap_or_default());
|
||||
return check_len(out);
|
||||
}
|
||||
// An occurrence past what `usize` holds (32-bit targets) names no match: the text is returned as it was.
|
||||
let Ok(k) = usize::try_from(occ.unsigned_abs()) else { return text_val(s) };
|
||||
// Keeps only the wanted match's captures: the nth from the front, or a window of the last `k`.
|
||||
let cap = if occ > 0 {
|
||||
re.captures_iter(&s).nth(k - 1)
|
||||
} else {
|
||||
let mut last = std::collections::VecDeque::new();
|
||||
for cap in re.captures_iter(&s) {
|
||||
if last.len() == k {
|
||||
last.pop_front();
|
||||
}
|
||||
last.push_back(cap);
|
||||
}
|
||||
if last.len() == k { last.pop_front() } else { None }
|
||||
};
|
||||
let Some(cap) = cap else { return text_val(s) };
|
||||
let caps: Vec<regex::Captures> = re.captures_iter(&s).collect();
|
||||
let k = occ.unsigned_abs() as usize;
|
||||
let idx = if occ > 0 { k.checked_sub(1) } else { caps.len().checked_sub(k) };
|
||||
let Some(cap) = idx.and_then(|i| caps.get(i)) else { return text_val(s) };
|
||||
let Some(m) = cap.get(0) else { return text_val(s) };
|
||||
out.push_str(s.get(..m.start()).unwrap_or_default());
|
||||
expand_capped(&cap, &rep, &mut out, limit).ok_or(CellError::Value)?;
|
||||
expand_capped(cap, &rep, &mut out, limit).ok_or(CellError::Value)?;
|
||||
out.push_str(s.get(m.end()..).unwrap_or_default());
|
||||
check_len(out)
|
||||
}
|
||||
@@ -1211,22 +1171,6 @@ mod tests {
|
||||
assert_eq!(rows_of(&ev("CONCATENATE", vec![av(row_t(&["a", "b"])), t("x")])), vec![vec![tv("ax"), tv("bx")]]);
|
||||
}
|
||||
|
||||
/// Security: TEXTJOIN builds its result piece by piece and stops at the text cap, rather
|
||||
/// than first copying every input (100,000 references to one 32,767-character text were
|
||||
/// 3 GB of copies). Errors among the inputs still win, in argument order.
|
||||
#[test]
|
||||
fn textjoin_stops_at_the_text_cap() {
|
||||
use gridcraft_core::Array;
|
||||
let long = tv(&"x".repeat(32_767));
|
||||
let many = || av(Array::filled(100_000, 1, long.clone()).into());
|
||||
is_err(ev("TEXTJOIN", vec![t(","), b(true), many()]), CellError::Value);
|
||||
is_err(ev("TEXTJOIN", vec![av(Array::filled(1, 100_000, long.clone()).into()), b(false), t("a"), t("b")]), CellError::Value);
|
||||
is_err(ev("TEXTJOIN", vec![t(","), b(true), many(), e(CellError::Div0)]), CellError::Div0);
|
||||
is_err(ev("TEXTJOIN", vec![av(arr(vec![vec![tv(","), Value::Error(CellError::NA)]])), b(true), t("a")]), CellError::NA);
|
||||
// Delimiters cycle and are converted only where used.
|
||||
is_text(ev("TEXTJOIN", vec![av(arr(vec![vec![nv(1.0), tv(";")]])), b(false), t("a"), t("b"), t("c"), t("d")]), "a1b;c1d");
|
||||
}
|
||||
|
||||
fn row_t(xs: &[&str]) -> Value {
|
||||
arr(vec![xs.iter().map(|x| tv(x)).collect()])
|
||||
}
|
||||
|
||||
@@ -51,7 +51,7 @@ pub(crate) fn arg(args: &[Arg], i: usize) -> R<&Arg> {
|
||||
|
||||
/// Whether argument `i` was supplied (present and not an empty literal like `F(1,)`).
|
||||
pub(crate) fn has(args: &[Arg], i: usize) -> bool {
|
||||
matches!(args.get(i), Some(a) if !matches!(a.value, Value::Empty) || a.from_ref || a.blank_value)
|
||||
matches!(args.get(i), Some(a) if !matches!(a.value, Value::Empty) || a.from_ref)
|
||||
}
|
||||
|
||||
/// Number coercion for a scalar parameter: empty = 0, TRUE = 1, numeric text parses.
|
||||
|
||||
@@ -53,15 +53,6 @@ impl Remote {
|
||||
self.conn = None;
|
||||
let mut last = std::io::Error::new(std::io::ErrorKind::NotFound, format!("cannot resolve {}", self.addr));
|
||||
for sa in self.addr.to_socket_addrs()? {
|
||||
// The token goes over plain TCP, so only to this machine: a network peer could read it,
|
||||
// and the app's control port listens on loopback only.
|
||||
if !sa.ip().is_loopback() {
|
||||
last = std::io::Error::new(
|
||||
std::io::ErrorKind::PermissionDenied,
|
||||
format!("{sa} is not a loopback address: the control token is sent only to 127.0.0.1 or ::1"),
|
||||
);
|
||||
continue;
|
||||
}
|
||||
match TcpStream::connect_timeout(&sa, Duration::from_secs(2)) {
|
||||
Ok(s) => {
|
||||
s.set_nodelay(true).ok();
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
//! JSON-RPC 2.0 framing and the MCP lifecycle / tools / resources / prompts methods.
|
||||
|
||||
use std::io::{BufRead, Read, Write};
|
||||
use std::io::{BufRead, Write};
|
||||
|
||||
use serde_json::{Value, json};
|
||||
|
||||
@@ -45,45 +45,6 @@ fn response(id: Value, result: Value) -> Value {
|
||||
json!({"jsonrpc": "2.0", "id": id, "result": result})
|
||||
}
|
||||
|
||||
/// Reads the next newline-delimited message into `buf`, keeping at most `MAX_LINE` bytes of it:
|
||||
/// a client sending no newline can't make the server allocate more (#2164). `None` at the end of
|
||||
/// `input`; `Some(false)` for a line over the limit, whose rest is read and dropped so the next
|
||||
/// message is read whole.
|
||||
pub(crate) fn read_message(input: &mut impl BufRead, buf: &mut Vec<u8>) -> std::io::Result<Option<bool>> {
|
||||
buf.clear();
|
||||
// One byte past the limit tells an over-long line from one of exactly `MAX_LINE` bytes.
|
||||
let limit = u64::try_from(MAX_LINE).unwrap_or(u64::MAX).saturating_add(1);
|
||||
if input.take(limit).read_until(b'\n', buf)? == 0 {
|
||||
return Ok(None);
|
||||
}
|
||||
if buf.len() > MAX_LINE && buf.last() != Some(&b'\n') {
|
||||
skip_line(input)?;
|
||||
buf.clear();
|
||||
return Ok(Some(false));
|
||||
}
|
||||
Ok(Some(true))
|
||||
}
|
||||
|
||||
/// Reads up to and including the next newline (or to the end of `input`) without keeping it.
|
||||
fn skip_line(input: &mut impl BufRead) -> std::io::Result<()> {
|
||||
loop {
|
||||
let chunk = input.fill_buf()?;
|
||||
if chunk.is_empty() {
|
||||
return Ok(());
|
||||
}
|
||||
match chunk.iter().position(|&b| b == b'\n') {
|
||||
Some(i) => {
|
||||
input.consume(i + 1);
|
||||
return Ok(());
|
||||
}
|
||||
None => {
|
||||
let n = chunk.len();
|
||||
input.consume(n);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn error(id: Value, code: i64, message: impl Into<String>) -> Value {
|
||||
json!({"jsonrpc": "2.0", "id": id, "error": {"code": code, "message": message.into()}})
|
||||
}
|
||||
@@ -107,13 +68,14 @@ impl Server {
|
||||
pub fn serve(&mut self, mut input: impl BufRead, mut output: impl Write) -> std::io::Result<()> {
|
||||
let mut buf = Vec::new();
|
||||
loop {
|
||||
let reply = match read_message(&mut input, &mut buf)? {
|
||||
None => return Ok(()),
|
||||
Some(false) => Some(error(Value::Null, INVALID_REQUEST, "message too large (16 MB max)").to_string()),
|
||||
Some(true) => match std::str::from_utf8(&buf) {
|
||||
Ok(line) => self.handle_line(line),
|
||||
Err(_) => Some(error(Value::Null, PARSE_ERROR, "message is not valid UTF-8").to_string()),
|
||||
},
|
||||
buf.clear();
|
||||
let n = input.read_until(b'\n', &mut buf)?;
|
||||
if n == 0 {
|
||||
return Ok(());
|
||||
}
|
||||
let reply = match std::str::from_utf8(&buf) {
|
||||
Ok(line) => self.handle_line(line),
|
||||
Err(_) => Some(error(Value::Null, PARSE_ERROR, "message is not valid UTF-8").to_string()),
|
||||
};
|
||||
if let Some(reply) = reply {
|
||||
output.write_all(reply.as_bytes())?;
|
||||
|
||||
@@ -359,71 +359,6 @@ fn serve_loop_over_buffers() {
|
||||
assert!(lines[2]["result"]["content"][0]["text"].as_str().unwrap().contains("42"));
|
||||
}
|
||||
|
||||
/// Streams `len` bytes of `a` from one reused buffer, then `tail`: a huge line costs no memory to
|
||||
/// produce, so only the server can make it expensive.
|
||||
struct HugeLine {
|
||||
left: u64,
|
||||
tail: &'static [u8],
|
||||
chunk: Vec<u8>,
|
||||
}
|
||||
|
||||
impl std::io::Read for HugeLine {
|
||||
fn read(&mut self, out: &mut [u8]) -> std::io::Result<usize> {
|
||||
let buf = std::io::BufRead::fill_buf(self)?;
|
||||
let n = buf.len().min(out.len());
|
||||
out[..n].copy_from_slice(&buf[..n]);
|
||||
std::io::BufRead::consume(self, n);
|
||||
Ok(n)
|
||||
}
|
||||
}
|
||||
|
||||
impl std::io::BufRead for HugeLine {
|
||||
fn fill_buf(&mut self) -> std::io::Result<&[u8]> {
|
||||
if self.left > 0 {
|
||||
let n = self.chunk.len().min(usize::try_from(self.left).unwrap_or(usize::MAX));
|
||||
return Ok(&self.chunk[..n]);
|
||||
}
|
||||
Ok(self.tail)
|
||||
}
|
||||
fn consume(&mut self, n: usize) {
|
||||
if self.left > 0 {
|
||||
self.left -= n as u64;
|
||||
} else {
|
||||
self.tail = &self.tail[n..];
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn a_line_without_newline_keeps_at_most_the_limit_in_memory() {
|
||||
// #2164: the whole line was buffered before its size was checked (4 GiB here; any size a
|
||||
// client cares to send).
|
||||
let mut input =
|
||||
HugeLine { left: 4 << 30, tail: concat!("\n", r#"{"jsonrpc":"2.0","id":7,"method":"ping"}"#, "\n").as_bytes(), chunk: vec![b'a'; 1 << 20] };
|
||||
let mut buf = Vec::new();
|
||||
assert_eq!(crate::server::read_message(&mut input, &mut buf).unwrap(), Some(false));
|
||||
assert!(buf.capacity() <= 2 * (16 << 20) + 1, "kept {} bytes", buf.capacity());
|
||||
assert_eq!(crate::server::read_message(&mut input, &mut buf).unwrap(), Some(true));
|
||||
assert!(buf.starts_with(br#"{"jsonrpc""#));
|
||||
assert_eq!(crate::server::read_message(&mut input, &mut buf).unwrap(), None);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn serve_loop_bounds_a_line_without_newline() {
|
||||
// #2164: a 4 GiB line was read whole before its size was checked. Now the server keeps at
|
||||
// most 16 MB of it, answers that it is too large, and reads the next message normally.
|
||||
let input =
|
||||
HugeLine { left: 4 << 30, tail: concat!("\n", r#"{"jsonrpc":"2.0","id":7,"method":"ping"}"#, "\n").as_bytes(), chunk: vec![b'a'; 1 << 20] };
|
||||
let mut out = Vec::new();
|
||||
server().serve(input, &mut out).unwrap();
|
||||
let lines: Vec<Value> = String::from_utf8(out).unwrap().lines().map(|l| serde_json::from_str(l).unwrap()).collect();
|
||||
assert_eq!(lines.len(), 2);
|
||||
assert_eq!(lines[0]["error"]["code"], -32600);
|
||||
assert!(lines[0]["error"]["message"].as_str().unwrap().contains("too large"));
|
||||
assert_eq!(lines[1]["id"], 7);
|
||||
assert!(lines[1].get("result").is_some(), "{}", lines[1]);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn cell_set_without_input_is_error() {
|
||||
let mut s = server();
|
||||
@@ -482,9 +417,6 @@ fn remote_backend_over_tcp() {
|
||||
drop(s);
|
||||
handle.join().unwrap();
|
||||
assert!(Remote::connect_with_token("127.0.0.1:1", "x".into()).is_err());
|
||||
// A network address is refused before anything, the token included, is sent.
|
||||
let err = Remote::connect_with_token("192.0.2.1:7979", "x".into()).err().unwrap();
|
||||
assert_eq!(err.kind(), std::io::ErrorKind::PermissionDenied, "{err}");
|
||||
}
|
||||
|
||||
#[test]
|
||||
|
||||
@@ -230,30 +230,6 @@ pub enum CfRule {
|
||||
},
|
||||
}
|
||||
|
||||
impl CfRule {
|
||||
/// The differential style a matching rule applies; colour scales, data bars and icon sets have none.
|
||||
pub fn style(&self) -> Option<&Style> {
|
||||
match self {
|
||||
CfRule::CellIs { style, .. }
|
||||
| CfRule::Expression { style, .. }
|
||||
| CfRule::ContainsText { style, .. }
|
||||
| CfRule::NotContainsText { style, .. }
|
||||
| CfRule::BeginsWith { style, .. }
|
||||
| CfRule::EndsWith { style, .. }
|
||||
| CfRule::Blanks { style }
|
||||
| CfRule::NoBlanks { style }
|
||||
| CfRule::Errors { style }
|
||||
| CfRule::NoErrors { style }
|
||||
| CfRule::Duplicate { style }
|
||||
| CfRule::Unique { style }
|
||||
| CfRule::Top10 { style, .. }
|
||||
| CfRule::AboveAverage { style, .. }
|
||||
| CfRule::TimePeriod { style, .. } => Some(style),
|
||||
CfRule::ColorScale { .. } | CfRule::DataBar { .. } | CfRule::IconSet { .. } => None,
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Number of icons in an icon set (`3Arrows` → 3, `5Quarters` → 5), always 3 to 5.
|
||||
pub fn icon_count(set: &str) -> usize {
|
||||
set.chars().next().and_then(|c| c.to_digit(10)).unwrap_or(3).clamp(3, 5) as usize
|
||||
|
||||
@@ -26,8 +26,8 @@ pub use python::{PyCode, PyOutput};
|
||||
pub use sheet::{DEFAULT_COL_WIDTH, DEFAULT_ROW_HEIGHT, LineIndex, Sheet, SheetUid, SheetViewMode, Visibility, points_to_chars};
|
||||
pub use store::CellStore;
|
||||
pub use style::{
|
||||
Alignment, BorderLine, BorderStyle, Borders, Color, Fill, Font, HAlign, NumFmt, PatternType, ReadingOrder, Style, StyleId, StyleTable, Theme,
|
||||
Underline, VAlign, VertAlign,
|
||||
Alignment, BorderLine, BorderStyle, Borders, Color, Fill, Font, HAlign, NumFmt, PatternType, Style, StyleId, StyleTable, Theme, Underline,
|
||||
VAlign, VertAlign,
|
||||
};
|
||||
pub use workbook::{
|
||||
CalcMode, CalcSettings, DefinedName, DocProps, EQUATION_BUDGET_MESSAGE, MAX_EQUATION_LAYOUT_BYTES, PhoneticProps, PhoneticRun, PhoneticString,
|
||||
|
||||
@@ -350,14 +350,6 @@ pub enum VAlign {
|
||||
Distributed,
|
||||
}
|
||||
|
||||
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq, Hash, Serialize, Deserialize)]
|
||||
pub enum ReadingOrder {
|
||||
#[default]
|
||||
Context,
|
||||
LeftToRight,
|
||||
RightToLeft,
|
||||
}
|
||||
|
||||
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq, Hash, Serialize, Deserialize)]
|
||||
pub struct Alignment {
|
||||
pub h: HAlign,
|
||||
@@ -367,9 +359,6 @@ pub struct Alignment {
|
||||
pub indent: u8,
|
||||
/// Degrees -90..=90, or 255 for vertical stacked text.
|
||||
pub rotation: i16,
|
||||
/// Unspecified inherits in differential styles and uses context for ordinary cells.
|
||||
#[serde(default)]
|
||||
pub reading_order: Option<ReadingOrder>,
|
||||
}
|
||||
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash, Serialize, Deserialize)]
|
||||
@@ -499,25 +488,6 @@ static DEFAULT_STYLE: std::sync::LazyLock<Style> = std::sync::LazyLock::new(Styl
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn reading_order_defaults_for_older_json_and_distinguishes_styles() {
|
||||
let align: Alignment = serde_json::from_str(r#"{"h":"General","v":"Bottom","wrap":false,"shrink":false,"indent":0,"rotation":0}"#).unwrap();
|
||||
assert_eq!(align.reading_order, None);
|
||||
let mut styles = StyleTable::default();
|
||||
let context = styles.derive(StyleId::DEFAULT, |s| s.align.reading_order = Some(ReadingOrder::Context));
|
||||
let ltr = styles.derive(StyleId::DEFAULT, |s| s.align.reading_order = Some(ReadingOrder::LeftToRight));
|
||||
let rtl = styles.derive(StyleId::DEFAULT, |s| s.align.reading_order = Some(ReadingOrder::RightToLeft));
|
||||
assert_ne!(ltr, rtl);
|
||||
assert_ne!(context, StyleId::DEFAULT);
|
||||
assert_ne!(context, ltr);
|
||||
assert_ne!(context, rtl);
|
||||
assert_ne!(ltr, StyleId::DEFAULT);
|
||||
assert_ne!(rtl, StyleId::DEFAULT);
|
||||
assert_eq!(styles.derive(ltr, |s| s.align.reading_order = Some(ReadingOrder::RightToLeft)), rtl);
|
||||
assert_eq!(styles.derive(rtl, |s| s.align.reading_order = Some(ReadingOrder::Context)), context);
|
||||
assert_eq!(styles.derive(context, |s| s.align.reading_order = None), StyleId::DEFAULT);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn interning() {
|
||||
let mut t = StyleTable::default();
|
||||
|
||||
@@ -268,11 +268,6 @@ pub struct Workbook {
|
||||
/// `preserved` are relative to. Saving into another folder rewrites them from there. Not in JSON.
|
||||
#[serde(skip)]
|
||||
pub source_dir: Option<std::path::PathBuf>,
|
||||
/// The region whose formats this workbook is typed and shown in (Options › Region). A runtime
|
||||
/// setting of the application, not of the file: it is not saved, and the engine session sets
|
||||
/// it on every document it holds. en-US until then.
|
||||
#[serde(skip, default = "gridcraft_core::region::default_region")]
|
||||
pub region: &'static gridcraft_core::Region,
|
||||
/// The file this workbook reports as its own, which is what `CELL("filename")` returns.
|
||||
/// `None` = no file yet, which Excel reports as an empty string until the workbook has been
|
||||
/// saved once. Deliberately separate from `source_dir`: that one names the folder relative
|
||||
@@ -370,7 +365,6 @@ impl Workbook {
|
||||
macros_unread: Arc::default(),
|
||||
preserved: None,
|
||||
source_dir: None,
|
||||
region: gridcraft_core::region::default_region(),
|
||||
source_path: None,
|
||||
}
|
||||
}
|
||||
|
||||
@@ -0,0 +1,56 @@
|
||||
//! Built-in number formats (ECMA-376 §18.8.30, en-US).
|
||||
|
||||
const BUILTINS: [(u32, &str); 36] = [
|
||||
(0, "General"),
|
||||
(1, "0"),
|
||||
(2, "0.00"),
|
||||
(3, "#,##0"),
|
||||
(4, "#,##0.00"),
|
||||
(5, "\"$\"#,##0_);\\(\"$\"#,##0\\)"),
|
||||
(6, "\"$\"#,##0_);[Red]\\(\"$\"#,##0\\)"),
|
||||
(7, "\"$\"#,##0.00_);\\(\"$\"#,##0.00\\)"),
|
||||
(8, "\"$\"#,##0.00_);[Red]\\(\"$\"#,##0.00\\)"),
|
||||
(9, "0%"),
|
||||
(10, "0.00%"),
|
||||
(11, "0.00E+00"),
|
||||
(12, "# ?/?"),
|
||||
(13, "# ??/??"),
|
||||
(14, "m/d/yyyy"),
|
||||
(15, "d-mmm-yy"),
|
||||
(16, "d-mmm"),
|
||||
(17, "mmm-yy"),
|
||||
(18, "h:mm AM/PM"),
|
||||
(19, "h:mm:ss AM/PM"),
|
||||
(20, "h:mm"),
|
||||
(21, "h:mm:ss"),
|
||||
(22, "m/d/yyyy h:mm"),
|
||||
(37, "#,##0 ;(#,##0)"),
|
||||
(38, "#,##0 ;[Red](#,##0)"),
|
||||
(39, "#,##0.00;(#,##0.00)"),
|
||||
(40, "#,##0.00;[Red](#,##0.00)"),
|
||||
(41, "_(* #,##0_);_(* \\(#,##0\\);_(* \"-\"_);_(@_)"),
|
||||
(42, "_(\"$\"* #,##0_);_(\"$\"* \\(#,##0\\);_(\"$\"* \"-\"_);_(@_)"),
|
||||
(43, "_(* #,##0.00_);_(* \\(#,##0.00\\);_(* \"-\"??_);_(@_)"),
|
||||
(44, "_(\"$\"* #,##0.00_);_(\"$\"* \\(#,##0.00\\);_(\"$\"* \"-\"??_);_(@_)"),
|
||||
(45, "mm:ss"),
|
||||
(46, "[h]:mm:ss"),
|
||||
(47, "mmss.0"),
|
||||
(48, "##0.0E+0"),
|
||||
(49, "@"),
|
||||
];
|
||||
|
||||
/// Built-in format code for an id (0..=49, en-US), e.g. 14 => "m/d/yyyy". Locale-specific ids
|
||||
/// (23–36) and unknown ids return `None`.
|
||||
pub fn builtin_format(id: u32) -> Option<&'static str> {
|
||||
BUILTINS.iter().find(|(i, _)| *i == id).map(|(_, c)| *c)
|
||||
}
|
||||
|
||||
/// Id of a built-in format code. Matching ignores case and backslash escapes.
|
||||
pub fn builtin_id(code: &str) -> Option<u32> {
|
||||
if let Some((id, _)) = BUILTINS.iter().find(|(_, c)| *c == code) {
|
||||
return Some(*id);
|
||||
}
|
||||
let norm = |s: &str| s.chars().filter(|&c| c != '\\').flat_map(char::to_lowercase).collect::<String>();
|
||||
let n = norm(code);
|
||||
BUILTINS.iter().find(|(_, c)| norm(c) == n).map(|(id, _)| *id)
|
||||
}
|
||||
@@ -1,206 +0,0 @@
|
||||
//! Calendars a format section can show dates in: the `B1` (Gregorian) and `B2` (Hijri) prefixes.
|
||||
|
||||
use gridcraft_core::date::{DateTime, MONTHS, WEEKDAYS, days_from_civil, days_in_month};
|
||||
|
||||
/// The calendar a section's year, month and day parts count in.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq, Default)]
|
||||
pub(crate) enum Calendar {
|
||||
#[default]
|
||||
Gregorian,
|
||||
/// The tabular (arithmetic) Islamic calendar.
|
||||
Hijri,
|
||||
}
|
||||
|
||||
/// Arabic names of the Hijri months, Muharram first. Two-word names are joined by a no-break
|
||||
/// space (U+00A0), as Excel writes them.
|
||||
const HIJRI_MONTHS: [&str; 12] = [
|
||||
"محرم",
|
||||
"صفر",
|
||||
"ربيع\u{a0}الأول",
|
||||
"ربيع\u{a0}الثاني",
|
||||
"جمادى\u{a0}الأولى",
|
||||
"جمادى\u{a0}الثانية",
|
||||
"رجب",
|
||||
"شعبان",
|
||||
"رمضان",
|
||||
"شوال",
|
||||
"ذو\u{a0}القعدة",
|
||||
"ذو\u{a0}الحجة",
|
||||
];
|
||||
|
||||
/// Arabic weekday names, Sunday first, shown for `ddd` and `dddd` under `B2`.
|
||||
const ARABIC_WEEKDAYS: [&str; 7] = ["الأحد", "الإثنين", "الثلاثاء", "الأربعاء", "الخميس", "الجمعة", "السبت"];
|
||||
|
||||
/// Year, month and day of a date in a section's calendar, with the month and weekday names.
|
||||
pub(crate) struct CalendarDate {
|
||||
pub year: i64,
|
||||
pub month: u32,
|
||||
pub day: u32,
|
||||
/// `mmmm`.
|
||||
pub month_name: &'static str,
|
||||
/// `mmm`.
|
||||
pub month_short: &'static str,
|
||||
/// `dddd`.
|
||||
pub day_name: &'static str,
|
||||
/// `ddd`.
|
||||
pub day_short: &'static str,
|
||||
}
|
||||
|
||||
impl Calendar {
|
||||
/// The date of a day in this calendar, `None` when it has none: under `B2`, the 1900 system's
|
||||
/// fictitious 29 February 1900 (serial 60, which Excel shows as `#VALUE!`) and days before
|
||||
/// 1 Muharram 1 AH.
|
||||
pub(crate) fn date(self, dt: &DateTime) -> Option<CalendarDate> {
|
||||
match self {
|
||||
Calendar::Hijri => hijri_date(dt),
|
||||
Calendar::Gregorian => Some(gregorian_date(dt)),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn gregorian_date(dt: &DateTime) -> CalendarDate {
|
||||
let month_name = MONTHS.get((dt.month as usize).wrapping_sub(1)).copied().unwrap_or("");
|
||||
let day_name = WEEKDAYS.get(dt.weekday as usize).copied().unwrap_or("");
|
||||
CalendarDate {
|
||||
year: dt.year.into(),
|
||||
month: dt.month,
|
||||
day: dt.day,
|
||||
month_name,
|
||||
month_short: month_name.get(..3).unwrap_or(month_name),
|
||||
day_name,
|
||||
day_short: day_name.get(..3).unwrap_or(day_name),
|
||||
}
|
||||
}
|
||||
|
||||
/// Days from 1 Muharram 1 AH to the first of a Hijri month. Leap years (30 days in the last
|
||||
/// month) are years 2, 5, 7, 10, 13, 15, 18, 21, 24, 26 and 29 of each 30-year cycle (10,631
|
||||
/// days), which is what Excel shows; odd months have 30 days and even months 29.
|
||||
fn days_before_month(year: i64, month: i64) -> i64 {
|
||||
// `(4 + 11 * year) / 30` is the number of leap years before `year`.
|
||||
(year - 1) * 354 + (4 + 11 * year).div_euclid(30) + 29 * (month - 1) + month / 2
|
||||
}
|
||||
|
||||
/// Hijri year, month and day of the day `days` after 1 Muharram 1 AH (`days >= 0`).
|
||||
fn hijri_from_days(days: i64) -> Option<(i64, i64, i64)> {
|
||||
if days < 0 {
|
||||
return None;
|
||||
}
|
||||
// The estimate is the mean year length rounded; it can be one year off at a year boundary.
|
||||
// Estimate and `days_before_month` both move by 30 years per 10,631 days, so one step is
|
||||
// enough everywhere if it is within one cycle (`every_day_round_trips` checks many).
|
||||
let mut year = (30 * days + 10646) / 10631;
|
||||
if days < days_before_month(year, 1) {
|
||||
year -= 1;
|
||||
} else if days >= days_before_month(year + 1, 1) {
|
||||
year += 1;
|
||||
}
|
||||
let month = ((11 * (days - days_before_month(year, 1)) + 330) / 325).clamp(1, 12);
|
||||
let day = days - days_before_month(year, month) + 1;
|
||||
Some((year, month, day))
|
||||
}
|
||||
|
||||
/// 1 Muharram 1 AH is 18 July 622 (proleptic Gregorian), so that 1 January 2001 is 6 Shawwal 1421.
|
||||
fn hijri_date(dt: &DateTime) -> Option<CalendarDate> {
|
||||
// 29 February 1900 (serial 60) is no real day and has no Hijri date. The 1900 system's day 0
|
||||
// (`1900-01-00`, serial 0) shows as 1 January 1900, as Excel shows it.
|
||||
if dt.day > days_in_month(dt.year, dt.month) {
|
||||
return None;
|
||||
}
|
||||
let (year, month, day) = hijri_from_days(days_from_civil(dt.year, dt.month, dt.day.max(1)) - days_from_civil(622, 7, 18))?;
|
||||
let month_name = HIJRI_MONTHS.get(usize::try_from(month - 1).ok()?).copied()?;
|
||||
// Excel shows the full Arabic name for `ddd` as well as `dddd`, as it does for `mmm`.
|
||||
let day_name = ARABIC_WEEKDAYS.get(dt.weekday as usize).copied()?;
|
||||
Some(CalendarDate {
|
||||
year,
|
||||
month: u32::try_from(month).ok()?,
|
||||
day: u32::try_from(day).ok()?,
|
||||
month_name,
|
||||
month_short: month_name,
|
||||
day_name,
|
||||
day_short: day_name,
|
||||
})
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
use gridcraft_core::DateSystem;
|
||||
use gridcraft_core::date::{datetime_from_serial, max_serial};
|
||||
|
||||
/// Leap years of the 30-year cycle (Excel's set, per the reviewer's 43,512-serial comparison).
|
||||
const LEAP: [i64; 11] = [2, 5, 7, 10, 13, 15, 18, 21, 24, 26, 29];
|
||||
|
||||
fn is_leap(year: i64) -> bool {
|
||||
LEAP.contains(&((year - 1).rem_euclid(30) + 1))
|
||||
}
|
||||
|
||||
fn month_len(year: i64, month: i64) -> i64 {
|
||||
if month == 12 { 29 + i64::from(is_leap(year)) } else { 30 - (month + 1) % 2 }
|
||||
}
|
||||
|
||||
/// Years 1 to 10,000 (every cycle up to and past 31 December 9999, which is 9666 AH): year,
|
||||
/// month and cycle lengths follow the leap-year set.
|
||||
#[test]
|
||||
fn year_and_month_lengths_follow_the_leap_set() {
|
||||
assert_eq!(days_before_month(1, 1), 0);
|
||||
for year in 1..=10_000 {
|
||||
let start = days_before_month(year, 1);
|
||||
let len = days_before_month(year + 1, 1) - start;
|
||||
assert_eq!(len, if is_leap(year) { 355 } else { 354 }, "year {year}");
|
||||
for month in 1..=12 {
|
||||
let next = if month == 12 { days_before_month(year + 1, 1) } else { days_before_month(year, month + 1) };
|
||||
assert_eq!(next - days_before_month(year, month), month_len(year, month), "{year}-{month}");
|
||||
}
|
||||
assert_eq!(days_before_month(year + 30, 1) - start, 10_631, "cycle from {year}");
|
||||
}
|
||||
// One cycle: exactly the eleven leap years, 10,631 days.
|
||||
let cycle: Vec<i64> = (1..=30).filter(|&y| days_before_month(y + 1, 1) - days_before_month(y, 1) == 355).collect();
|
||||
assert_eq!(cycle, LEAP);
|
||||
}
|
||||
|
||||
/// Every day from 1 Muharram 1 AH to one cycle past 31 December 9999: day -> Hijri -> day
|
||||
/// round-trips, the day is within its month, and consecutive days follow each other.
|
||||
#[test]
|
||||
fn every_day_round_trips() {
|
||||
let last = days_from_civil(9999, 12, 31) - days_from_civil(622, 7, 18) + 10_631;
|
||||
let mut prev: Option<(i64, i64, i64)> = None;
|
||||
for days in 0..=last {
|
||||
let (y, m, d) = hijri_from_days(days).unwrap();
|
||||
assert!((1..=12).contains(&m) && (1..=month_len(y, m)).contains(&d), "{days}: {y}-{m}-{d}");
|
||||
assert_eq!(days_before_month(y, m) + d - 1, days, "{y}-{m}-{d}");
|
||||
if let Some((py, pm, pd)) = prev {
|
||||
let expected = if pd < month_len(py, pm) {
|
||||
(py, pm, pd + 1)
|
||||
} else if pm < 12 {
|
||||
(py, pm + 1, 1)
|
||||
} else {
|
||||
(py + 1, 1, 1)
|
||||
};
|
||||
assert_eq!((y, m, d), expected, "after {py}-{pm}-{pd}");
|
||||
}
|
||||
prev = Some((y, m, d));
|
||||
}
|
||||
assert_eq!(hijri_from_days(-1), None);
|
||||
}
|
||||
|
||||
/// Every serial of the 1900 system: the Hijri date names the serial's day, one day apart.
|
||||
#[test]
|
||||
fn every_serial_has_the_right_hijri_day() {
|
||||
let epoch = days_from_civil(622, 7, 18);
|
||||
let mut prev = None;
|
||||
for serial in 1..=max_serial(DateSystem::D1900) {
|
||||
if serial == 60 {
|
||||
continue;
|
||||
}
|
||||
let dt = datetime_from_serial(DateSystem::D1900, serial as f64).unwrap();
|
||||
let h = Calendar::Hijri.date(&dt).unwrap();
|
||||
assert!((1..=month_len(h.year, i64::from(h.month))).contains(&i64::from(h.day)), "serial {serial}");
|
||||
let days = days_before_month(h.year, i64::from(h.month)) + i64::from(h.day) - 1;
|
||||
assert_eq!(days, days_from_civil(dt.year, dt.month, dt.day) - epoch, "serial {serial}");
|
||||
if let Some(p) = prev {
|
||||
assert_eq!(days, p + 1, "serial {serial}");
|
||||
}
|
||||
prev = Some(days);
|
||||
}
|
||||
}
|
||||
}
|
||||
+50
-143
@@ -4,8 +4,6 @@
|
||||
//! digits and then rounded half away from zero on that decimal string. That makes 2.675 with
|
||||
//! "0.00" show 2.68 and 1.005 show 1.01, as in Excel, even though neither is exact in binary.
|
||||
|
||||
use std::fmt::Write as _;
|
||||
|
||||
/// Most decimals we ever compute; placeholders beyond this are padded with zeros.
|
||||
pub(crate) const MAX_DECIMALS: usize = 400;
|
||||
|
||||
@@ -23,8 +21,7 @@ pub(crate) fn decimal_digits(a: f64) -> ([u8; 15], i32) {
|
||||
}
|
||||
// 16 digits, rounded. Unless the 16th is a 5 it decides the 15th exactly (4 or less rounds
|
||||
// down, 6 or more up), so one format serves almost every number.
|
||||
let mut d16 = [0u8; 16];
|
||||
let exp = significant_digits(a, &mut d16);
|
||||
let (d16, exp) = significant_digits(a, 16);
|
||||
for (slot, d) in digits.iter_mut().zip(&d16) {
|
||||
*slot = *d;
|
||||
}
|
||||
@@ -33,8 +30,11 @@ pub(crate) fn decimal_digits(a: f64) -> ([u8; 15], i32) {
|
||||
5 if is_exact_tie(a) => (digits, exp),
|
||||
5 => {
|
||||
// Not a tie: round the exact value (the 16-digit rounding above may be a double rounding).
|
||||
let (d15, exp15) = significant_digits(a, 15);
|
||||
let mut out = [0u8; 15];
|
||||
let exp15 = significant_digits(a, &mut out);
|
||||
for (slot, d) in out.iter_mut().zip(&d15) {
|
||||
*slot = *d;
|
||||
}
|
||||
(out, exp15)
|
||||
}
|
||||
_ => {
|
||||
@@ -59,54 +59,12 @@ pub(crate) fn decimal_digits(a: f64) -> ([u8; 15], i32) {
|
||||
}
|
||||
}
|
||||
|
||||
/// Writes `out.len()` significant digits of `a` (rounded half to even on the exact binary value)
|
||||
/// into `out` and returns the power of ten of the first.
|
||||
///
|
||||
/// Up to 32 digits are formatted in a stack buffer; more (the 800 of the tie check) use a string.
|
||||
fn significant_digits(a: f64, out: &mut [u8]) -> i32 {
|
||||
fn parse(s: &str, out: &mut [u8]) -> i32 {
|
||||
let (mant, exp) = s.split_once('e').unwrap_or((s, "0"));
|
||||
let digits = mant.bytes().filter(u8::is_ascii_digit).map(|b| b - b'0');
|
||||
for (slot, d) in out.iter_mut().zip(digits) {
|
||||
*slot = d;
|
||||
}
|
||||
exp.parse().unwrap_or(0)
|
||||
}
|
||||
let prec = out.len().saturating_sub(1);
|
||||
if out.len() <= 32 {
|
||||
let mut buf = StackStr::default();
|
||||
if write!(buf, "{:.*e}", prec, a).is_ok() {
|
||||
return parse(buf.as_str(), out);
|
||||
}
|
||||
}
|
||||
parse(&format!("{:.*e}", prec, a), out)
|
||||
}
|
||||
|
||||
/// A fixed-size string buffer for short number formatting; a write that does not fit fails.
|
||||
struct StackStr {
|
||||
buf: [u8; 64],
|
||||
len: usize,
|
||||
}
|
||||
|
||||
impl Default for StackStr {
|
||||
fn default() -> Self {
|
||||
StackStr { buf: [0; 64], len: 0 }
|
||||
}
|
||||
}
|
||||
|
||||
impl StackStr {
|
||||
fn as_str(&self) -> &str {
|
||||
self.buf.get(..self.len).and_then(|b| std::str::from_utf8(b).ok()).unwrap_or("")
|
||||
}
|
||||
}
|
||||
|
||||
impl std::fmt::Write for StackStr {
|
||||
fn write_str(&mut self, s: &str) -> std::fmt::Result {
|
||||
let end = self.len.checked_add(s.len()).ok_or(std::fmt::Error)?;
|
||||
self.buf.get_mut(self.len..end).ok_or(std::fmt::Error)?.copy_from_slice(s.as_bytes());
|
||||
self.len = end;
|
||||
Ok(())
|
||||
}
|
||||
/// `n` significant digits of `a` (rounded half to even on the exact binary value) and the power of
|
||||
/// ten of the first.
|
||||
fn significant_digits(a: f64, n: usize) -> (Vec<u8>, i32) {
|
||||
let s = format!("{:.*e}", n.saturating_sub(1), a);
|
||||
let (mant, exp) = s.split_once('e').unwrap_or((s.as_str(), "0"));
|
||||
(mant.bytes().filter(u8::is_ascii_digit).map(|b| b - b'0').collect(), exp.parse().unwrap_or(0))
|
||||
}
|
||||
|
||||
/// True when `a`'s exact decimal expansion has a 5 as its 16th significant digit and only zeros after.
|
||||
@@ -114,59 +72,27 @@ fn is_exact_tie(a: f64) -> bool {
|
||||
let is_tie = |ds: &[u8]| ds.get(15) == Some(&5) && ds.get(16..).is_some_and(|t| t.iter().all(|&d| d == 0));
|
||||
// A cheap screen first (30 digits), then the exact expansion: a double has at most 767
|
||||
// significant digits and Rust prints every one of them.
|
||||
let mut screen = [0u8; 30];
|
||||
significant_digits(a, &mut screen);
|
||||
if !is_tie(&screen) {
|
||||
return false;
|
||||
}
|
||||
let mut exact = vec![0u8; 800];
|
||||
significant_digits(a, &mut exact);
|
||||
is_tie(&exact)
|
||||
is_tie(&significant_digits(a, 30).0) && is_tie(&significant_digits(a, 800).0)
|
||||
}
|
||||
|
||||
/// Rounds the number `digits × 10^exp` (see [`decimal_digits`]) to `decimals` places, half away
|
||||
/// from zero. Returns the integer digits ("" when the integer part is zero) and exactly
|
||||
/// `decimals` fraction digits.
|
||||
pub(crate) fn fixed_from_digits(digits: &[u8; 15], exp: i32, decimals: usize) -> (String, String) {
|
||||
let (mut int, mut frac) = (String::new(), String::new());
|
||||
fixed_into(digits, exp, decimals, &mut int, &mut frac);
|
||||
(int, frac)
|
||||
}
|
||||
|
||||
/// [`fixed_from_digits`] into caller-owned strings (cleared first), so a loop can reuse them.
|
||||
///
|
||||
/// The rounded digits are `lead` (at most 16, a carry can add a leading 1) followed by `zeros`
|
||||
/// zero digits (placeholders beyond the 15 known digits), which is all a `Vec` of them held.
|
||||
fn fixed_into(digits: &[u8; 15], exp: i32, decimals: usize, int: &mut String, frac: &mut String) {
|
||||
int.clear();
|
||||
frac.clear();
|
||||
let decimals_calc = decimals.min(MAX_DECIMALS);
|
||||
let keep = exp as i64 + decimals_calc as i64 + 1;
|
||||
let mut head = [0u8; 16];
|
||||
let mut head_len = 0usize;
|
||||
let mut zeros = 0usize;
|
||||
if keep <= 0 {
|
||||
if keep == 0 && digits.first().copied().unwrap_or(0) >= 5 {
|
||||
if let Some(h) = head.first_mut() {
|
||||
*h = 1;
|
||||
}
|
||||
head_len = 1;
|
||||
}
|
||||
let mut m: Vec<u8> = if keep <= 0 {
|
||||
if keep == 0 && digits.first().copied().unwrap_or(0) >= 5 { vec![1] } else { Vec::new() }
|
||||
} else if keep as usize >= digits.len() {
|
||||
for (h, d) in head.iter_mut().zip(digits) {
|
||||
*h = *d;
|
||||
}
|
||||
head_len = digits.len();
|
||||
zeros = keep as usize - digits.len();
|
||||
let mut v = digits.to_vec();
|
||||
v.resize(keep as usize, 0);
|
||||
v
|
||||
} else {
|
||||
let keep = keep as usize;
|
||||
for (h, d) in head.iter_mut().zip(digits.iter().take(keep)) {
|
||||
*h = *d;
|
||||
}
|
||||
head_len = keep;
|
||||
let mut v = digits.get(..keep).map(<[u8]>::to_vec).unwrap_or_default();
|
||||
if digits.get(keep).copied().unwrap_or(0) >= 5 {
|
||||
let mut carry = true;
|
||||
for d in head.iter_mut().take(keep).rev() {
|
||||
for d in v.iter_mut().rev() {
|
||||
if *d == 9 {
|
||||
*d = 0;
|
||||
} else {
|
||||
@@ -176,36 +102,26 @@ fn fixed_into(digits: &[u8; 15], exp: i32, decimals: usize, int: &mut String, fr
|
||||
}
|
||||
}
|
||||
if carry {
|
||||
head.copy_within(0..keep, 1);
|
||||
if let Some(h) = head.first_mut() {
|
||||
*h = 1;
|
||||
}
|
||||
head_len += 1;
|
||||
v.insert(0, 1);
|
||||
}
|
||||
}
|
||||
}
|
||||
let total = head_len + zeros;
|
||||
let mut lead = head.iter().take(head_len).take_while(|&&d| d == 0).count();
|
||||
if lead == head_len {
|
||||
lead = total; // all zeros
|
||||
}
|
||||
let len = total - lead;
|
||||
// Digit `k` of the rounded number after its leading zeros.
|
||||
let digit_at = |k: usize| {
|
||||
let pos = lead + k;
|
||||
if pos < head_len { head.get(pos).map_or('0', |&d| char::from(b'0' + d)) } else { '0' }
|
||||
v
|
||||
};
|
||||
if len <= decimals_calc {
|
||||
frac.extend(std::iter::repeat_n('0', decimals_calc - len));
|
||||
frac.extend((0..len).map(digit_at));
|
||||
let lead = m.iter().take_while(|&&d| d == 0).count();
|
||||
m.drain(..lead);
|
||||
let to_s = |ds: &[u8]| ds.iter().map(|d| char::from(b'0' + d)).collect::<String>();
|
||||
let (int, mut frac) = if m.len() <= decimals_calc {
|
||||
let mut f = "0".repeat(decimals_calc - m.len());
|
||||
f.push_str(&to_s(&m));
|
||||
(String::new(), f)
|
||||
} else {
|
||||
let split = len - decimals_calc;
|
||||
int.extend((0..split).map(digit_at));
|
||||
frac.extend((split..len).map(digit_at));
|
||||
}
|
||||
let split = m.len() - decimals_calc;
|
||||
(to_s(m.get(..split).unwrap_or(&[])), to_s(m.get(split..).unwrap_or(&[])))
|
||||
};
|
||||
if decimals > decimals_calc {
|
||||
frac.extend(std::iter::repeat_n('0', decimals - decimals_calc));
|
||||
frac.push_str(&"0".repeat(decimals - decimals_calc));
|
||||
}
|
||||
(int, frac)
|
||||
}
|
||||
|
||||
/// Rounds a non-negative number to `decimals` places (see [`fixed_from_digits`]).
|
||||
@@ -232,25 +148,21 @@ pub fn format_general_fit(n: f64, max_chars: usize) -> Option<String> {
|
||||
let sign = if n < 0.0 { "-" } else { "" };
|
||||
let (digits, x) = decimal_digits(n.abs());
|
||||
let int_len = sign.len() + (x.max(0) as usize) + 1;
|
||||
// Fixed notation for magnitudes from 1E-4 up to what fits. The 15 digits are computed once;
|
||||
// each width only re-rounds them, into strings that are reused.
|
||||
let (mut i, mut f, mut s) = (String::new(), String::new(), String::new());
|
||||
// Fixed notation for magnitudes from 1E-4 up to what fits.
|
||||
if x >= -4 && int_len <= max_chars {
|
||||
let avail = max_chars - int_len;
|
||||
// Zeros past the 400th decimal are trimmed below anyway, so a huge width needs no more.
|
||||
let mut decimals = (if avail >= 2 { avail - 1 } else { 0 }).min(MAX_DECIMALS);
|
||||
let mut decimals = if avail >= 2 { avail - 1 } else { 0 };
|
||||
loop {
|
||||
fixed_into(&digits, x, decimals, &mut i, &mut f);
|
||||
let ft = f.trim_end_matches('0');
|
||||
if i.is_empty() && ft.is_empty() {
|
||||
let (i, f) = fixed_from_digits(&digits, x, decimals);
|
||||
let f = f.trim_end_matches('0');
|
||||
if i.is_empty() && f.is_empty() {
|
||||
break; // rounded away to zero: use scientific
|
||||
}
|
||||
s.clear();
|
||||
s.push_str(sign);
|
||||
let mut s = String::from(sign);
|
||||
s.push_str(if i.is_empty() { "0" } else { &i });
|
||||
if !ft.is_empty() {
|
||||
if !f.is_empty() {
|
||||
s.push('.');
|
||||
s.push_str(ft);
|
||||
s.push_str(f);
|
||||
}
|
||||
if s.chars().count() <= max_chars {
|
||||
return Some(s);
|
||||
@@ -263,20 +175,19 @@ pub fn format_general_fit(n: f64, max_chars: usize) -> Option<String> {
|
||||
}
|
||||
for md in (0..=max_chars.min(14)).rev() {
|
||||
let mut e = x;
|
||||
fixed_into(&digits, 0, md, &mut i, &mut f);
|
||||
if i.len() > 1 {
|
||||
let (mut mi, mut mf) = fixed_from_digits(&digits, 0, md);
|
||||
if mi.len() > 1 {
|
||||
e += 1;
|
||||
fixed_into(&digits, -1, md, &mut i, &mut f);
|
||||
(mi, mf) = fixed_from_digits(&digits, -1, md);
|
||||
}
|
||||
let ft = f.trim_end_matches('0');
|
||||
s.clear();
|
||||
s.push_str(sign);
|
||||
s.push_str(if i.is_empty() { "0" } else { &i });
|
||||
if !ft.is_empty() {
|
||||
let mf = mf.trim_end_matches('0');
|
||||
let mut s = String::from(sign);
|
||||
s.push_str(if mi.is_empty() { "0" } else { &mi });
|
||||
if !mf.is_empty() {
|
||||
s.push('.');
|
||||
s.push_str(ft);
|
||||
s.push_str(mf);
|
||||
}
|
||||
let _ = write!(s, "E{}{:02}", if e < 0 { '-' } else { '+' }, e.unsigned_abs());
|
||||
s.push_str(&format!("E{}{:02}", if e < 0 { '-' } else { '+' }, e.unsigned_abs()));
|
||||
if s.chars().count() <= max_chars {
|
||||
return Some(s);
|
||||
}
|
||||
@@ -300,10 +211,6 @@ mod tests {
|
||||
assert_eq!(fixed(0.0005, 3), ("".into(), "001".into()));
|
||||
assert_eq!(fixed(1e20, 0).0, "100000000000000000000");
|
||||
assert_eq!(fixed(0.0, 2), ("".into(), "00".into()));
|
||||
// Past 400 decimals only zeros are added; huge exponents keep all their digits.
|
||||
assert_eq!(fixed(1.5, 450), ("1".into(), format!("5{}", "0".repeat(449))));
|
||||
assert_eq!(fixed(1e300, 1), (format!("1{}", "0".repeat(300)), "0".into()));
|
||||
assert_eq!(fixed(123456789012345678.0, 0).0, "123456789012346000");
|
||||
}
|
||||
|
||||
#[test]
|
||||
|
||||
@@ -3,7 +3,7 @@
|
||||
#![deny(clippy::unwrap_used, clippy::expect_used, clippy::panic, clippy::unimplemented, clippy::todo, clippy::unreachable)]
|
||||
#![forbid(unsafe_code)]
|
||||
|
||||
mod calendar;
|
||||
mod builtin;
|
||||
mod decimal;
|
||||
mod parse;
|
||||
mod render;
|
||||
@@ -12,6 +12,7 @@ use std::sync::Arc;
|
||||
|
||||
use gridcraft_core::{CellError, DateSystem, Value};
|
||||
|
||||
pub use builtin::{builtin_format, builtin_id};
|
||||
pub use decimal::format_general_fit;
|
||||
|
||||
use parse::{DatePart, SecKind, Section, Tok};
|
||||
@@ -95,9 +96,6 @@ struct Inner {
|
||||
num_sections: usize,
|
||||
/// Index of the section that applies to text, if any.
|
||||
text_section: Option<usize>,
|
||||
/// A code Excel rejects (a misplaced `B1`/`B2` calendar prefix): shown as General, and
|
||||
/// `TEXT` returns `#VALUE!`, as for a code that is too long.
|
||||
invalid: bool,
|
||||
}
|
||||
|
||||
/// A parsed number format code. Cheap to clone.
|
||||
@@ -132,8 +130,7 @@ fn too_long(code: &str) -> bool {
|
||||
|
||||
impl NumberFormat {
|
||||
/// Parses a format code. Never fails: unknown or malformed parts degrade to literal text, and
|
||||
/// a code longer than [`MAX_FORMAT_CODE_CHARS`] characters, or one Excel rejects for a
|
||||
/// misplaced `B1`/`B2` calendar prefix, degrades to General.
|
||||
/// a code longer than [`MAX_FORMAT_CODE_CHARS`] characters degrades to General.
|
||||
pub fn parse(code: &str) -> NumberFormat {
|
||||
Self::parse_inner(code, false)
|
||||
}
|
||||
@@ -150,16 +147,12 @@ impl NumberFormat {
|
||||
let trimmed = code.trim();
|
||||
// `checkbox` marks a checkbox cell (Insert › Checkbox): TRUE/FALSE draw as a box, any other
|
||||
// value displays as General.
|
||||
let general = |invalid: bool| NumberFormat {
|
||||
inner: Arc::new(Inner { code: code.to_string(), general: true, sections: Vec::new(), num_sections: 0, text_section: None, invalid }),
|
||||
};
|
||||
if trimmed.is_empty() || trimmed.eq_ignore_ascii_case("general") || trimmed == "checkbox" || (!uncapped && too_long(code)) {
|
||||
return general(false);
|
||||
return NumberFormat {
|
||||
inner: Arc::new(Inner { code: code.to_string(), general: true, sections: Vec::new(), num_sections: 0, text_section: None }),
|
||||
};
|
||||
}
|
||||
let sections = parse::parse_sections(code);
|
||||
if sections.iter().any(|s| s.invalid) {
|
||||
return general(true);
|
||||
}
|
||||
let len = sections.len();
|
||||
let last_is_text = sections.last().is_some_and(|s| s.kind == SecKind::Text);
|
||||
let text_section = if len >= 4 {
|
||||
@@ -176,7 +169,7 @@ impl NumberFormat {
|
||||
} else {
|
||||
len
|
||||
};
|
||||
NumberFormat { inner: Arc::new(Inner { code: code.to_string(), general: false, sections, num_sections, text_section, invalid: false }) }
|
||||
NumberFormat { inner: Arc::new(Inner { code: code.to_string(), general: false, sections, num_sections, text_section }) }
|
||||
}
|
||||
|
||||
pub fn general() -> NumberFormat {
|
||||
@@ -393,14 +386,6 @@ impl NumberFormat {
|
||||
None => Formatted { text: t.to_string(), color: None, fill: None, numeric: false },
|
||||
}
|
||||
}
|
||||
|
||||
/// [`Self::format_text`]'s text, or `None` once it passes `max_units` UTF-16 code units.
|
||||
fn format_text_within(&self, t: &str, max_units: usize) -> Option<String> {
|
||||
match self.inner.text_section.and_then(|i| self.inner.sections.get(i)) {
|
||||
Some(sec) => render::render_text_within(sec, t, max_units).map(|out| out.text),
|
||||
None => (t.encode_utf16().count() <= max_units).then(|| t.to_string()),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Formats a cell value for display.
|
||||
@@ -494,10 +479,6 @@ impl NumberFormat {
|
||||
/// The `TEXT()` worksheet function: formats a value with a format code. Numeric text is
|
||||
/// converted to a number first; other text and booleans (as `TRUE`/`FALSE`) go through the text
|
||||
/// section, or are returned as is.
|
||||
/// Longest text a cell holds, in UTF-16 code units (Excel's 32,767 characters, which XLSB also
|
||||
/// stores that way), and so the longest `TEXT` result.
|
||||
const MAX_CELL_TEXT_CHARS: usize = 32_767;
|
||||
|
||||
pub fn text_function(v: &Value, code: &str, sys: DateSystem) -> Result<String, CellError> {
|
||||
if code.is_empty() {
|
||||
return Ok(String::new());
|
||||
@@ -506,9 +487,6 @@ pub fn text_function(v: &Value, code: &str, sys: DateSystem) -> Result<String, C
|
||||
return Err(CellError::Value);
|
||||
}
|
||||
let fmt = NumberFormat::parse_cached(code);
|
||||
if fmt.inner.invalid {
|
||||
return Err(CellError::Value);
|
||||
}
|
||||
let n = match v.scalar() {
|
||||
Value::Error(e) => return Err(e),
|
||||
Value::Empty => 0.0,
|
||||
@@ -516,9 +494,7 @@ pub fn text_function(v: &Value, code: &str, sys: DateSystem) -> Result<String, C
|
||||
Value::Bool(b) => return Ok(fmt.format_text(if b { "TRUE" } else { "FALSE" }).text),
|
||||
Value::Text(t) => match gridcraft_core::parse::parse_number_text(&t) {
|
||||
Some(n) => n,
|
||||
// Every `@` repeats the text: a result past a cell's limit is `#VALUE!`, so nested
|
||||
// TEXT calls can't multiply it further.
|
||||
None => return fmt.format_text_within(&t, MAX_CELL_TEXT_CHARS).ok_or(CellError::Value),
|
||||
None => return Ok(fmt.format_text(&t).text),
|
||||
},
|
||||
Value::Array(_) => return Err(CellError::Value),
|
||||
};
|
||||
|
||||
+12
-100
@@ -1,7 +1,6 @@
|
||||
//! Format code parsing: sections, bracket tags and tokens.
|
||||
|
||||
use crate::FormatColor;
|
||||
use crate::calendar::Calendar;
|
||||
|
||||
/// At most this many sections are honoured (`pos;neg;zero;text`); extra ones are ignored.
|
||||
const MAX_SECTIONS: usize = 4;
|
||||
@@ -25,11 +24,6 @@ pub(crate) enum DatePart {
|
||||
}
|
||||
|
||||
impl DatePart {
|
||||
/// `[h]`, `[m]` or `[s]`.
|
||||
pub(crate) fn is_elapsed(self) -> bool {
|
||||
matches!(self, DatePart::ElapsedH(_) | DatePart::ElapsedM(_) | DatePart::ElapsedS(_))
|
||||
}
|
||||
|
||||
pub(crate) fn is_calendar(self) -> bool {
|
||||
matches!(self, DatePart::Year(_) | DatePart::Month(_) | DatePart::Day(_))
|
||||
}
|
||||
@@ -60,28 +54,18 @@ pub(crate) enum Tok {
|
||||
}
|
||||
|
||||
impl Tok {
|
||||
/// Appends the text this token shows when it has no special meaning in its section.
|
||||
pub(crate) fn push_literal(&self, out: &mut String) {
|
||||
match self {
|
||||
Tok::Lit(s) | Tok::FixedDenom(s) => out.push_str(s),
|
||||
Tok::Digit(c) => out.push(*c),
|
||||
Tok::Point => out.push('.'),
|
||||
Tok::Comma => out.push(','),
|
||||
Tok::Percent => out.push('%'),
|
||||
Tok::Exp { upper, plus } => {
|
||||
out.push(if *upper { 'E' } else { 'e' });
|
||||
out.push(if *plus { '+' } else { '-' });
|
||||
}
|
||||
Tok::Slash => out.push('/'),
|
||||
Tok::Fill(_) | Tok::At | Tok::General | Tok::Date(_) => {}
|
||||
}
|
||||
}
|
||||
|
||||
/// The text this token shows when it has no special meaning in its section.
|
||||
pub(crate) fn literal_text(&self) -> String {
|
||||
let mut s = String::new();
|
||||
self.push_literal(&mut s);
|
||||
s
|
||||
match self {
|
||||
Tok::Lit(s) | Tok::FixedDenom(s) => s.clone(),
|
||||
Tok::Digit(c) => c.to_string(),
|
||||
Tok::Point => ".".into(),
|
||||
Tok::Comma => ",".into(),
|
||||
Tok::Percent => "%".into(),
|
||||
Tok::Exp { upper, plus } => format!("{}{}", if *upper { 'E' } else { 'e' }, if *plus { '+' } else { '-' }),
|
||||
Tok::Slash => "/".into(),
|
||||
Tok::Fill(_) | Tok::At | Tok::General | Tok::Date(_) => String::new(),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -138,17 +122,6 @@ pub(crate) struct Section {
|
||||
pub fraction: bool,
|
||||
pub currency: bool,
|
||||
pub fill: bool,
|
||||
/// Date sections: the most digits of any `.0`/`.00`/`.000` part (at most 3), which decides how
|
||||
/// the time of day is rounded.
|
||||
pub sub_digits: u8,
|
||||
/// Date sections: has a year, month or day part, so the serial's date is needed.
|
||||
pub needs_date: bool,
|
||||
/// Date sections: has `AM/PM` or `A/P`, so hours show on the 12-hour clock.
|
||||
pub ampm: bool,
|
||||
pub calendar: Calendar,
|
||||
/// Excel rejects the whole code because of this section: a `B` and digit that is not a
|
||||
/// calendar prefix, or a prefix before a number format (`B20.00`).
|
||||
pub invalid: bool,
|
||||
}
|
||||
|
||||
/// Splits a code into sections at unquoted, unescaped, unbracketed semicolons.
|
||||
@@ -278,60 +251,17 @@ struct Raw {
|
||||
color: Option<FormatColor>,
|
||||
cond: Option<Cond>,
|
||||
currency: bool,
|
||||
calendar: Calendar,
|
||||
/// The section's calendar prefix as written, `'1'` or `'2'` (whether or not a locale tag
|
||||
/// cancels it).
|
||||
prefix: Option<char>,
|
||||
/// An unquoted `B` and digit that is not a calendar prefix: Excel rejects the code.
|
||||
invalid: bool,
|
||||
}
|
||||
|
||||
/// Where a section's body starts: after its leading `[...]` tags, read as `tokenize` reads them.
|
||||
/// Also whether one of those tags is a `[$...]` locale tag.
|
||||
fn after_leading_brackets(cs: &[char]) -> (usize, bool) {
|
||||
let mut i = 0usize;
|
||||
let mut locale = false;
|
||||
while cs.get(i) == Some(&'[') {
|
||||
match cs.get(i + 1..).and_then(|r| r.iter().position(|&d| d == ']')) {
|
||||
Some(p) => {
|
||||
locale |= cs.get(i + 1) == Some(&'$');
|
||||
i += p + 2;
|
||||
}
|
||||
None => break,
|
||||
}
|
||||
}
|
||||
(i, locale)
|
||||
}
|
||||
|
||||
fn tokenize(src: &str) -> Raw {
|
||||
let cs: Vec<char> = src.chars().collect();
|
||||
let mut raw = Raw { toks: Vec::new(), color: None, cond: None, currency: false, calendar: Calendar::Gregorian, prefix: None, invalid: false };
|
||||
let mut raw = Raw { toks: Vec::new(), color: None, cond: None, currency: false };
|
||||
let toks = &mut raw.toks;
|
||||
// With no `]` at or after a `[`, that `[` is a literal; checking this once keeps the scan linear.
|
||||
let last_close = cs.iter().rposition(|&c| c == ']');
|
||||
let (body_start, locale_first) = after_leading_brackets(&cs);
|
||||
// Where the next calendar prefix may stand: the body's start, or right after another prefix.
|
||||
let mut prefix_at = body_start;
|
||||
let mut i = 0usize;
|
||||
while let Some(&c) = cs.get(i) {
|
||||
match c {
|
||||
// A calendar prefix, never shown: `B1` or `B2`, any case, which Excel accepts repeated
|
||||
// (`B2B2`) but not mixed (`B2B1`) or with more digits (`B12`, `B02`).
|
||||
'B' | 'b' if i == prefix_at && cs.get(i + 1).is_some_and(char::is_ascii_digit) => {
|
||||
let d = cs.get(i + 1).copied();
|
||||
let ok = matches!(d, Some('1' | '2')) && !cs.get(i + 2).is_some_and(char::is_ascii_digit) && raw.prefix.is_none_or(|p| Some(p) == d);
|
||||
raw.invalid |= !ok;
|
||||
raw.prefix = d;
|
||||
i += 2;
|
||||
prefix_at = i;
|
||||
continue;
|
||||
}
|
||||
// Any other `B` and digit (`B3dd`, `dd/mm/yyyyB2`, `dd B2 mm`, `B2[Red]B2`) makes Excel
|
||||
// reject the code.
|
||||
'B' | 'b' if cs.get(i + 1).is_some_and(char::is_ascii_digit) => {
|
||||
raw.invalid = true;
|
||||
push_lit(toks, c.encode_utf8(&mut [0; 4]));
|
||||
}
|
||||
'"' => {
|
||||
let mut s = String::new();
|
||||
i += 1;
|
||||
@@ -353,8 +283,6 @@ fn tokenize(src: &str) -> Raw {
|
||||
if is_currency_char(d) {
|
||||
raw.currency = true;
|
||||
}
|
||||
// Excel rejects an escaped `\B2` as well.
|
||||
raw.invalid |= matches!(d, 'B' | 'b') && cs.get(i + 2).is_some_and(char::is_ascii_digit);
|
||||
push_lit(toks, d.encode_utf8(&mut [0; 4]));
|
||||
}
|
||||
i += 2;
|
||||
@@ -455,11 +383,6 @@ fn tokenize(src: &str) -> Raw {
|
||||
}
|
||||
i += 1;
|
||||
}
|
||||
// Excel keeps the date Gregorian when a `[$-401]`-style locale or currency tag comes before
|
||||
// `B2` (`B2[$-401]` is still Hijri).
|
||||
if raw.prefix == Some('2') && !locale_first {
|
||||
raw.calendar = Calendar::Hijri;
|
||||
}
|
||||
raw
|
||||
}
|
||||
|
||||
@@ -515,7 +438,7 @@ fn date_part(t: &Tok) -> Option<DatePart> {
|
||||
}
|
||||
|
||||
fn analyze(raw: Raw) -> Section {
|
||||
let Raw { toks, color, cond, currency, calendar, prefix, invalid } = raw;
|
||||
let Raw { toks, color, cond, currency } = raw;
|
||||
let is_date = toks.iter().any(|t| matches!(t, Tok::Date(_)));
|
||||
let has_general = toks.iter().any(|t| matches!(t, Tok::General));
|
||||
let mut sec = Section {
|
||||
@@ -532,11 +455,6 @@ fn analyze(raw: Raw) -> Section {
|
||||
fraction: false,
|
||||
currency,
|
||||
fill: toks.iter().any(|t| matches!(t, Tok::Fill(_))),
|
||||
sub_digits: 0,
|
||||
needs_date: false,
|
||||
ampm: false,
|
||||
calendar,
|
||||
invalid,
|
||||
};
|
||||
if is_date {
|
||||
sec.kind = SecKind::Date;
|
||||
@@ -551,9 +469,6 @@ fn analyze(raw: Raw) -> Section {
|
||||
sec.kind = SecKind::Text;
|
||||
return analyze_text(toks, sec);
|
||||
}
|
||||
// Number section. Excel rejects a calendar prefix before digits (`B20.00`, `B1#,##0`);
|
||||
// `B2General`, `B2@` and `B2"x"` show as without the prefix.
|
||||
sec.invalid |= prefix.is_some() && sec.has_digits;
|
||||
analyze_number(toks, sec)
|
||||
}
|
||||
|
||||
@@ -593,9 +508,6 @@ fn analyze_date(mut toks: Vec<Tok>, mut sec: Section) -> Section {
|
||||
}
|
||||
i += 1;
|
||||
}
|
||||
sec.sub_digits = out.iter().filter_map(|t| if let Tok::Date(DatePart::SubSec(n)) = t { Some(*n) } else { None }).max().unwrap_or(0).min(3);
|
||||
sec.needs_date = out.iter().any(|t| matches!(t, Tok::Date(d) if d.is_calendar()));
|
||||
sec.ampm = out.iter().any(|t| matches!(t, Tok::Date(DatePart::AmPm | DatePart::AP(..))));
|
||||
sec.toks = out;
|
||||
sec
|
||||
}
|
||||
|
||||
+133
-242
@@ -1,11 +1,8 @@
|
||||
//! Rendering a value with a parsed section.
|
||||
|
||||
use std::fmt::Write as _;
|
||||
|
||||
use gridcraft_core::DateSystem;
|
||||
use gridcraft_core::date::{datetime_from_serial, max_serial};
|
||||
use gridcraft_core::date::{MONTHS, WEEKDAYS, datetime_from_serial, max_serial};
|
||||
|
||||
use crate::calendar::Calendar;
|
||||
use crate::decimal::{decimal_digits, fixed, fixed_from_digits, format_general_fit};
|
||||
use crate::parse::{DatePart, Section, Tok};
|
||||
|
||||
@@ -27,10 +24,6 @@ impl Out {
|
||||
self.fill = Some((c, self.text.len()));
|
||||
}
|
||||
}
|
||||
/// Appends what `t` shows when it has no special meaning (no temporary string).
|
||||
fn push_tok(&mut self, t: &Tok) {
|
||||
t.push_literal(&mut self.text);
|
||||
}
|
||||
pub(crate) fn prepend(&mut self, s: &str) {
|
||||
self.text.insert_str(0, s);
|
||||
if let Some((_, p)) = &mut self.fill {
|
||||
@@ -46,41 +39,12 @@ pub(crate) fn render_text(sec: &Section, text: &str) -> Out {
|
||||
match t {
|
||||
Tok::At => out.push(text),
|
||||
Tok::Fill(c) => out.fill(*c),
|
||||
other => out.push_tok(other),
|
||||
other => out.push(&other.literal_text()),
|
||||
}
|
||||
}
|
||||
out
|
||||
}
|
||||
|
||||
/// [`render_text`], but `None` once the output passes `max_units` UTF-16 code units (a cell's
|
||||
/// limit), checked before each piece is appended, so an oversized result is never built.
|
||||
pub(crate) fn render_text_within(sec: &Section, text: &str, max_units: usize) -> Option<Out> {
|
||||
let text_units = text.encode_utf16().count();
|
||||
let mut out = Out::default();
|
||||
let mut units = 0usize;
|
||||
for t in &sec.toks {
|
||||
match t {
|
||||
Tok::At => {
|
||||
units = units.saturating_add(text_units);
|
||||
if units > max_units {
|
||||
return None;
|
||||
}
|
||||
out.push(text);
|
||||
}
|
||||
Tok::Fill(c) => out.fill(*c),
|
||||
other => {
|
||||
let literal = other.literal_text();
|
||||
units = units.saturating_add(literal.encode_utf16().count());
|
||||
if units > max_units {
|
||||
return None;
|
||||
}
|
||||
out.push(&literal);
|
||||
}
|
||||
}
|
||||
}
|
||||
Some(out)
|
||||
}
|
||||
|
||||
/// Renders a number through a General section (literals around the General keyword).
|
||||
pub(crate) fn render_general(sec: &Section, v: f64, width: usize) -> Out {
|
||||
let mut out = Out::default();
|
||||
@@ -88,7 +52,7 @@ pub(crate) fn render_general(sec: &Section, v: f64, width: usize) -> Out {
|
||||
match t {
|
||||
Tok::General => out.push(&format_general_fit(v, width).unwrap_or_else(|| "#".repeat(width.clamp(1, 11)))),
|
||||
Tok::Fill(c) => out.fill(*c),
|
||||
other => out.push_tok(other),
|
||||
other => out.push(&other.literal_text()),
|
||||
}
|
||||
}
|
||||
out
|
||||
@@ -99,7 +63,8 @@ pub(crate) fn render_date(sec: &Section, v: f64, sys: DateSystem) -> Option<Out>
|
||||
if !v.is_finite() || !(0.0..max_serial(sys) as f64 + 1.0).contains(&v) {
|
||||
return None;
|
||||
}
|
||||
let unit: i64 = 10i64.pow(3 - u32::from(sec.sub_digits.min(3)));
|
||||
let sub_digits = sec.toks.iter().filter_map(|t| if let Tok::Date(DatePart::SubSec(n)) = t { Some(*n) } else { None }).max().unwrap_or(0).min(3);
|
||||
let unit: i64 = 10i64.pow(3 - sub_digits as u32);
|
||||
let mut day = v.floor() as i64;
|
||||
let mut ms = (((v - v.floor()) * 86_400_000.0) / unit as f64).round() as i64 * unit;
|
||||
if ms >= 86_400_000 {
|
||||
@@ -107,42 +72,38 @@ pub(crate) fn render_date(sec: &Section, v: f64, sys: DateSystem) -> Option<Out>
|
||||
day += 1;
|
||||
}
|
||||
let total_ms = day.saturating_mul(86_400_000).saturating_add(ms);
|
||||
// Under `B2`, Excel shows the 1900 system's fictitious 29 February (serial 60) as `#VALUE!`
|
||||
// whenever a clock or calendar part is shown, even for `hh:mm` (`[h]` alone still works).
|
||||
if sec.calendar == Calendar::Hijri
|
||||
&& sys == DateSystem::D1900
|
||||
&& day == 60
|
||||
&& sec.toks.iter().any(|t| matches!(t, Tok::Date(d) if !d.is_elapsed()))
|
||||
{
|
||||
return None;
|
||||
}
|
||||
let dt = if sec.needs_date { datetime_from_serial(sys, day as f64)? } else { datetime_from_serial(sys, 0.0).unwrap_or_default() };
|
||||
let ampm = sec.ampm;
|
||||
let needs_date = sec.toks.iter().any(|t| matches!(t, Tok::Date(d) if d.is_calendar()));
|
||||
let dt = if needs_date { datetime_from_serial(sys, day as f64)? } else { datetime_from_serial(sys, 0.0).unwrap_or_default() };
|
||||
let ampm = sec.toks.iter().any(|t| matches!(t, Tok::Date(DatePart::AmPm | DatePart::AP(..))));
|
||||
let hour = ms / 3_600_000;
|
||||
let minute = (ms / 60_000) % 60;
|
||||
let second = (ms / 1000) % 60;
|
||||
let milli = ms % 1000;
|
||||
// No date in the section's calendar (serial 60 under `B2`): shown like an out-of-range date.
|
||||
let date = sec.calendar.date(&dt)?;
|
||||
let pad = |n: i64, w: u8| if w >= 2 { format!("{:0w$}", n, w = w.min(20) as usize) } else { n.to_string() };
|
||||
let month_name = MONTHS.get((dt.month as usize).wrapping_sub(1)).copied().unwrap_or("");
|
||||
let day_name = WEEKDAYS.get(dt.weekday as usize).copied().unwrap_or("");
|
||||
let mut out = Out::default();
|
||||
for t in &sec.toks {
|
||||
match t {
|
||||
Tok::Date(d) => {
|
||||
let text = &mut out.text;
|
||||
match *d {
|
||||
let s = match *d {
|
||||
DatePart::Year(n) => {
|
||||
let _ = if n <= 2 { write!(text, "{:02}", date.year.rem_euclid(100)) } else { write!(text, "{:04}", date.year) };
|
||||
if n <= 2 {
|
||||
format!("{:02}", dt.year.rem_euclid(100))
|
||||
} else {
|
||||
format!("{:04}", dt.year)
|
||||
}
|
||||
}
|
||||
DatePart::Month(n) => match n {
|
||||
1 | 2 => push_padded(text, date.month as i64, n),
|
||||
3 => text.push_str(date.month_short),
|
||||
5 => text.extend(date.month_name.chars().take(1)),
|
||||
_ => text.push_str(date.month_name),
|
||||
1 | 2 => pad(dt.month as i64, n),
|
||||
3 => month_name.chars().take(3).collect(),
|
||||
5 => month_name.chars().take(1).collect(),
|
||||
_ => month_name.to_string(),
|
||||
},
|
||||
DatePart::Day(n) => match n {
|
||||
1 | 2 => push_padded(text, date.day as i64, n),
|
||||
3 => text.push_str(date.day_short),
|
||||
_ => text.push_str(date.day_name),
|
||||
1 | 2 => pad(dt.day as i64, n),
|
||||
3 => day_name.chars().take(3).collect(),
|
||||
_ => day_name.to_string(),
|
||||
},
|
||||
DatePart::Hour(n) => {
|
||||
let h = if ampm {
|
||||
@@ -151,168 +112,78 @@ pub(crate) fn render_date(sec: &Section, v: f64, sys: DateSystem) -> Option<Out>
|
||||
} else {
|
||||
hour
|
||||
};
|
||||
push_padded(text, h, n);
|
||||
pad(h, n)
|
||||
}
|
||||
DatePart::Minute(n) => push_padded(text, minute, n),
|
||||
DatePart::Second(n) => push_padded(text, second, n),
|
||||
DatePart::Minute(n) => pad(minute, n),
|
||||
DatePart::Second(n) => pad(second, n),
|
||||
DatePart::SubSec(n) => {
|
||||
// `.` and the first `n` (at most 3) of the three millisecond digits, then
|
||||
// zeros for the digits beyond milliseconds.
|
||||
text.push('.');
|
||||
for (k, div) in [100, 10, 1].into_iter().enumerate() {
|
||||
if k < n as usize {
|
||||
text.push(char::from(b'0' + (milli / div % 10) as u8));
|
||||
}
|
||||
}
|
||||
for _ in 3..n {
|
||||
text.push('0');
|
||||
let mut s = format!(".{:03}", milli);
|
||||
s.truncate(1 + (n as usize).min(3));
|
||||
if n > 3 {
|
||||
s.push_str(&"0".repeat(n as usize - 3));
|
||||
}
|
||||
s
|
||||
}
|
||||
DatePart::ElapsedH(n) => push_padded(text, total_ms / 3_600_000, n),
|
||||
DatePart::ElapsedM(n) => push_padded(text, total_ms / 60_000, n),
|
||||
DatePart::ElapsedS(n) => push_padded(text, total_ms / 1000, n),
|
||||
DatePart::AmPm => text.push_str(if hour >= 12 { "PM" } else { "AM" }),
|
||||
DatePart::AP(a, p) => text.push(if hour >= 12 { p } else { a }),
|
||||
}
|
||||
DatePart::ElapsedH(n) => pad(total_ms / 3_600_000, n),
|
||||
DatePart::ElapsedM(n) => pad(total_ms / 60_000, n),
|
||||
DatePart::ElapsedS(n) => pad(total_ms / 1000, n),
|
||||
DatePart::AmPm => (if hour >= 12 { "PM" } else { "AM" }).to_string(),
|
||||
DatePart::AP(a, p) => (if hour >= 12 { p } else { a }).to_string(),
|
||||
};
|
||||
out.push(&s);
|
||||
}
|
||||
Tok::Fill(c) => out.fill(*c),
|
||||
other => out.push_tok(other),
|
||||
other => out.push(&other.literal_text()),
|
||||
}
|
||||
}
|
||||
Some(out)
|
||||
}
|
||||
|
||||
/// Appends `n`, zero padded to `w` digits (at most 20) when `w` is 2 or more.
|
||||
fn push_padded(out: &mut String, n: i64, w: u8) {
|
||||
let _ = if w >= 2 { write!(out, "{:0w$}", n, w = w.min(20) as usize) } else { write!(out, "{n}") };
|
||||
}
|
||||
|
||||
/// Scratch text of one number rendering: every token's own output (`piece`, replaces the token's
|
||||
/// text), the text inserted ahead of a token (`before`) and the replacement of a literal
|
||||
/// (`lit_spaces`: fractions blank the literals of a hidden part) live in ONE buffer, addressed
|
||||
/// by byte ranges, instead of one string per token. Writing a piece again appends the new text
|
||||
/// and points the slot at it.
|
||||
struct Parts {
|
||||
buf: String,
|
||||
slots: Vec<Slot>,
|
||||
}
|
||||
|
||||
#[derive(Clone, Copy, Default)]
|
||||
struct Slot {
|
||||
piece: (usize, usize),
|
||||
before: (usize, usize),
|
||||
/// `Some(k)`: show `k` spaces instead of the literal.
|
||||
lit_spaces: Option<usize>,
|
||||
}
|
||||
|
||||
impl Parts {
|
||||
fn new(n: usize) -> Parts {
|
||||
Parts { buf: String::new(), slots: vec![Slot::default(); n + 1] }
|
||||
}
|
||||
/// Start of a piece written straight into the buffer; finish it with [`Parts::end_piece`].
|
||||
fn mark(&self) -> usize {
|
||||
self.buf.len()
|
||||
}
|
||||
fn end_piece(&mut self, ti: usize, start: usize) {
|
||||
let end = self.buf.len();
|
||||
if let Some(sl) = self.slots.get_mut(ti) {
|
||||
sl.piece = (start, end);
|
||||
}
|
||||
}
|
||||
fn set_piece(&mut self, ti: usize, text: &str) {
|
||||
let start = self.mark();
|
||||
self.buf.push_str(text);
|
||||
self.end_piece(ti, start);
|
||||
}
|
||||
fn set_spaces(&mut self, ti: usize, k: usize) {
|
||||
let start = self.mark();
|
||||
self.buf.extend(std::iter::repeat_n(' ', k));
|
||||
self.end_piece(ti, start);
|
||||
}
|
||||
fn set_before(&mut self, ti: usize, text: &str) {
|
||||
let start = self.mark();
|
||||
self.buf.push_str(text);
|
||||
let end = self.buf.len();
|
||||
if let Some(sl) = self.slots.get_mut(ti) {
|
||||
sl.before = (start, end);
|
||||
}
|
||||
}
|
||||
fn set_lit_spaces(&mut self, ti: usize, k: usize) {
|
||||
if let Some(sl) = self.slots.get_mut(ti) {
|
||||
sl.lit_spaces = Some(k);
|
||||
}
|
||||
}
|
||||
fn text(&self, (a, b): (usize, usize)) -> &str {
|
||||
self.buf.get(a..b).unwrap_or("")
|
||||
}
|
||||
fn piece(&self, ti: usize) -> &str {
|
||||
self.slots.get(ti).map_or("", |sl| self.text(sl.piece))
|
||||
}
|
||||
fn before(&self, ti: usize) -> &str {
|
||||
self.slots.get(ti).map_or("", |sl| self.text(sl.before))
|
||||
}
|
||||
}
|
||||
|
||||
/// Assigns integer digits right-to-left to the placeholders at `idxs`; the leftmost placeholder
|
||||
/// takes all remaining digits. `force_zero` shows `0` in the units placeholder when there are no
|
||||
/// digits. A blank `?` pads with a space, and so does the thousands separator it would carry.
|
||||
///
|
||||
/// `digits` is ASCII (digit strings made by this crate), so bytes are characters. Each
|
||||
/// placeholder's text is written left to right: a digit, then the separator that follows it when
|
||||
/// its position counted from the right is a multiple of 3.
|
||||
fn fill_int(toks: &[Tok], idxs: &[usize], digits: &str, thousands: bool, force_zero: bool, parts: &mut Parts) {
|
||||
let ds = digits.as_bytes();
|
||||
fn fill_int(toks: &[Tok], idxs: &[usize], digits: &str, thousands: bool, force_zero: bool, pieces: &mut [String]) {
|
||||
let ds: Vec<char> = digits.chars().collect();
|
||||
let mut rem = ds.len();
|
||||
let mut pos = 0usize;
|
||||
let count = idxs.len();
|
||||
let sep = |p: usize| thousands && p > 0 && p.is_multiple_of(3);
|
||||
for (k, &ti) in idxs.iter().rev().enumerate() {
|
||||
let leftmost = k + 1 == count;
|
||||
let ph = match toks.get(ti) {
|
||||
Some(Tok::Digit(c)) => *c,
|
||||
_ => '0',
|
||||
};
|
||||
let start = parts.mark();
|
||||
let mut rev = String::new();
|
||||
let push_digit = |c: char, rev: &mut String, pos: &mut usize| {
|
||||
if thousands && *pos > 0 && pos.is_multiple_of(3) {
|
||||
rev.push(',');
|
||||
}
|
||||
rev.push(c);
|
||||
*pos += 1;
|
||||
};
|
||||
let take = if leftmost { rem } else { rem.min(1) };
|
||||
if take == 0 {
|
||||
match ph {
|
||||
'0' => {
|
||||
parts.buf.push('0');
|
||||
if sep(pos) {
|
||||
parts.buf.push(',');
|
||||
}
|
||||
pos += 1;
|
||||
}
|
||||
_ if force_zero && k == 0 => {
|
||||
parts.buf.push('0');
|
||||
if sep(pos) {
|
||||
parts.buf.push(',');
|
||||
}
|
||||
pos += 1;
|
||||
}
|
||||
'0' => push_digit('0', &mut rev, &mut pos),
|
||||
_ if force_zero && k == 0 => push_digit('0', &mut rev, &mut pos),
|
||||
'?' => {
|
||||
parts.buf.push(' ');
|
||||
if sep(pos) {
|
||||
parts.buf.push(' ');
|
||||
if thousands && pos > 0 && pos.is_multiple_of(3) {
|
||||
rev.push(' ');
|
||||
}
|
||||
rev.push(' ');
|
||||
pos += 1;
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
} else {
|
||||
// The `take` digits just left of `rem`, most significant first; the digit `s` places
|
||||
// from the right end of the run sits at position `pos + s`.
|
||||
for s in (0..take).rev() {
|
||||
let d = ds.get(rem.saturating_sub(1 + s)).copied().unwrap_or(b'0');
|
||||
parts.buf.push(char::from(d));
|
||||
if sep(pos + s) {
|
||||
parts.buf.push(',');
|
||||
}
|
||||
for _ in 0..take {
|
||||
rem -= 1;
|
||||
push_digit(ds.get(rem).copied().unwrap_or('0'), &mut rev, &mut pos);
|
||||
}
|
||||
rem -= take;
|
||||
pos += take;
|
||||
}
|
||||
parts.end_piece(ti, start);
|
||||
if let Some(p) = pieces.get_mut(ti) {
|
||||
*p = rev.chars().rev().collect();
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -332,26 +203,26 @@ fn with_commas(digits: &str, thousands: bool) -> String {
|
||||
}
|
||||
|
||||
/// Fraction-digit placeholders: trailing zeros show as `0`, nothing (`#`) or a space (`?`).
|
||||
/// `frac` is ASCII digits.
|
||||
fn fill_frac(toks: &[Tok], idxs: &[usize], frac: &str, parts: &mut Parts) {
|
||||
let fd = frac.as_bytes();
|
||||
let last_sig = fd.iter().rposition(|&c| c != b'0');
|
||||
fn fill_frac(toks: &[Tok], idxs: &[usize], frac: &str, pieces: &mut [String]) {
|
||||
let fd: Vec<char> = frac.chars().collect();
|
||||
let last_sig = fd.iter().rposition(|&c| c != '0');
|
||||
for (i, &ti) in idxs.iter().enumerate() {
|
||||
let ph = match toks.get(ti) {
|
||||
Some(Tok::Digit(c)) => *c,
|
||||
_ => '0',
|
||||
};
|
||||
let start = parts.mark();
|
||||
if last_sig.is_some_and(|l| i <= l) {
|
||||
parts.buf.push(char::from(fd.get(i).copied().unwrap_or(b'0')));
|
||||
let s = if last_sig.is_some_and(|l| i <= l) {
|
||||
fd.get(i).copied().unwrap_or('0').to_string()
|
||||
} else {
|
||||
match ph {
|
||||
'0' => parts.buf.push('0'),
|
||||
'?' => parts.buf.push(' '),
|
||||
_ => {}
|
||||
'0' => "0".into(),
|
||||
'?' => " ".into(),
|
||||
_ => String::new(),
|
||||
}
|
||||
};
|
||||
if let Some(p) = pieces.get_mut(ti) {
|
||||
*p = s;
|
||||
}
|
||||
parts.end_piece(ti, start);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -466,12 +337,15 @@ pub(crate) fn render_number(sec: &Section, v: f64) -> Out {
|
||||
}
|
||||
let toks = &sec.toks;
|
||||
let n = toks.len();
|
||||
// Pieces replace the token's own output; `before` is inserted ahead of a token; literal
|
||||
// replacements blank the literals of a hidden fraction part (see [`Parts`]).
|
||||
let mut parts = Parts::new(n);
|
||||
// Pieces replace the token's own output; `before` is inserted ahead of a token; `lits`
|
||||
// replaces a literal's text (fractions blank the literals of a hidden part).
|
||||
let mut pieces: Vec<String> = vec![String::new(); n];
|
||||
let mut before: Vec<String> = vec![String::new(); n + 1];
|
||||
let mut lits: Vec<Option<String>> = Vec::new();
|
||||
let mut zero = false;
|
||||
if sec.fraction {
|
||||
render_fraction(sec, x, &mut parts);
|
||||
lits = vec![None; n];
|
||||
render_fraction(sec, x, &mut pieces, &mut lits);
|
||||
} else {
|
||||
let NumberLayout { point, exp, int_idx, frac_idx } = number_layout(toks);
|
||||
let (int_s, frac_s) = if let Some(e_at) = exp {
|
||||
@@ -485,6 +359,7 @@ pub(crate) fn render_number(sec: &Section, v: f64) -> Out {
|
||||
if int_s.is_empty() && frac_s.chars().all(|c| c == '0') {
|
||||
e = 0;
|
||||
}
|
||||
let head = if upper { "E" } else { "e" };
|
||||
let sign = if e < 0 {
|
||||
"-"
|
||||
} else if plus {
|
||||
@@ -497,16 +372,18 @@ pub(crate) fn render_number(sec: &Section, v: f64) -> Out {
|
||||
// `E` (observed: `|0|e+|0|` shows `|1|e|+5|`); `?` pads the digits with spaces.
|
||||
match exp_idx.first() {
|
||||
Some(&first) => {
|
||||
parts.set_piece(e_at, if upper { "E" } else { "e" });
|
||||
parts.set_before(first, sign);
|
||||
fill_int(toks, &exp_idx, &digits, false, false, &mut parts);
|
||||
if let Some(p) = pieces.get_mut(e_at) {
|
||||
*p = head.to_string();
|
||||
}
|
||||
if let Some(b) = before.get_mut(first) {
|
||||
*b = sign.to_string();
|
||||
}
|
||||
fill_int(toks, &exp_idx, &digits, false, false, &mut pieces);
|
||||
}
|
||||
None => {
|
||||
let start = parts.mark();
|
||||
parts.buf.push(if upper { 'E' } else { 'e' });
|
||||
parts.buf.push_str(sign);
|
||||
parts.buf.push_str(&digits);
|
||||
parts.end_piece(e_at, start);
|
||||
if let Some(p) = pieces.get_mut(e_at) {
|
||||
*p = format!("{head}{sign}{digits}");
|
||||
}
|
||||
}
|
||||
}
|
||||
(int_s, frac_s)
|
||||
@@ -514,32 +391,32 @@ pub(crate) fn render_number(sec: &Section, v: f64) -> Out {
|
||||
fixed(x, frac_idx.len())
|
||||
};
|
||||
zero = int_s.chars().chain(frac_s.chars()).all(|c| c == '0');
|
||||
fill_int(toks, &int_idx, &int_s, sec.thousands, false, &mut parts);
|
||||
fill_int(toks, &int_idx, &int_s, sec.thousands, false, &mut pieces);
|
||||
if int_idx.is_empty()
|
||||
&& !int_s.is_empty()
|
||||
&& let Some(p) = point.or(exp)
|
||||
&& let Some(b) = before.get_mut(p)
|
||||
{
|
||||
parts.set_before(p, &with_commas(&int_s, sec.thousands));
|
||||
*b = with_commas(&int_s, sec.thousands);
|
||||
}
|
||||
fill_frac(toks, &frac_idx, &frac_s, &mut parts);
|
||||
if let Some(p) = point {
|
||||
parts.set_piece(p, ".");
|
||||
fill_frac(toks, &frac_idx, &frac_s, &mut pieces);
|
||||
if let Some(p) = point
|
||||
&& let Some(s) = pieces.get_mut(p)
|
||||
{
|
||||
*s = ".".into();
|
||||
}
|
||||
}
|
||||
let mut out = Out::default();
|
||||
for (i, t) in toks.iter().enumerate() {
|
||||
out.push(parts.before(i));
|
||||
if let Some(b) = before.get(i) {
|
||||
out.push(b);
|
||||
}
|
||||
match t {
|
||||
Tok::Lit(s) => match parts.slots.get(i).and_then(|sl| sl.lit_spaces) {
|
||||
Some(k) => out.text.extend(std::iter::repeat_n(' ', k)),
|
||||
None => out.push(s),
|
||||
},
|
||||
Tok::Lit(s) => out.push(lits.get(i).and_then(Option::as_deref).unwrap_or(s)),
|
||||
Tok::Percent => out.push("%"),
|
||||
Tok::Fill(c) => out.fill(*c),
|
||||
// Every other token shows its piece. `At` (`0@`) is reachable here, `Comma`, `General`
|
||||
// and `Date` are not (analysis turns commas into literals and routes sections with
|
||||
// dates or General elsewhere); none of them is ever given a piece, so they print "".
|
||||
_ => out.push(parts.piece(i)),
|
||||
Tok::At | Tok::General | Tok::Date(_) | Tok::Comma => {}
|
||||
_ => out.push(pieces.get(i).map(String::as_str).unwrap_or("")),
|
||||
}
|
||||
}
|
||||
out.zero = zero;
|
||||
@@ -595,14 +472,22 @@ fn shown_width(t: &Tok) -> usize {
|
||||
}
|
||||
|
||||
/// Replaces the literals in `range` with spaces (`spaces`) or nothing; placeholders and the bar
|
||||
/// in it get the same treatment as pieces.
|
||||
fn blank_range(toks: &[Tok], range: std::ops::Range<usize>, spaces: bool, parts: &mut Parts) {
|
||||
/// in it get the same treatment through `pieces`.
|
||||
fn blank_range(toks: &[Tok], range: std::ops::Range<usize>, spaces: bool, pieces: &mut [String], lits: &mut [Option<String>]) {
|
||||
for i in range {
|
||||
let Some(t) = toks.get(i) else { break };
|
||||
let width = if spaces { shown_width(t) } else { 0 };
|
||||
let s = if spaces { " ".repeat(shown_width(t)) } else { String::new() };
|
||||
match t {
|
||||
Tok::Lit(_) => parts.set_lit_spaces(i, width),
|
||||
Tok::Digit(_) | Tok::Slash | Tok::FixedDenom(_) => parts.set_spaces(i, width),
|
||||
Tok::Lit(_) => {
|
||||
if let Some(l) = lits.get_mut(i) {
|
||||
*l = Some(s);
|
||||
}
|
||||
}
|
||||
Tok::Digit(_) | Tok::Slash | Tok::FixedDenom(_) => {
|
||||
if let Some(p) = pieces.get_mut(i) {
|
||||
*p = s;
|
||||
}
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
}
|
||||
@@ -686,7 +571,7 @@ fn fraction_parts(lay: &FractionLayout, x: f64) -> (f64, f64, u64) {
|
||||
/// becomes spaces when the whole part or the numerator has a `?`); when the fraction is zero and
|
||||
/// the numerator has no `0` placeholder, the separator, numerator, bar and denominator disappear
|
||||
/// (or become spaces when any part has a `?`) and the whole part shows at least `0`.
|
||||
fn render_fraction(sec: &Section, x: f64, parts: &mut Parts) {
|
||||
fn render_fraction(sec: &Section, x: f64, pieces: &mut [String], lits: &mut [Option<String>]) {
|
||||
let toks = &sec.toks;
|
||||
let Some(lay) = fraction_layout(toks) else { return };
|
||||
let (whole, num, den) = fraction_parts(&lay, x);
|
||||
@@ -696,28 +581,32 @@ fn render_fraction(sec: &Section, x: f64, parts: &mut Parts) {
|
||||
let whole_s = if whole >= 1.0 { fixed(whole, 0).0 } else { String::new() };
|
||||
let num_zero_ph = num_idx.iter().any(|&i| toks.get(i) == Some(&Tok::Digit('0')));
|
||||
if has_whole && num == 0.0 && !num_zero_ph {
|
||||
fill_int(toks, &whole_idx, &whole_s, sec.thousands, true, parts);
|
||||
fill_int(toks, &whole_idx, &whole_s, sec.thousands, true, pieces);
|
||||
let spaces = has_q(&whole_idx) || has_q(&num_idx) || has_q(&den_idx);
|
||||
blank_range(toks, sep_start..den_end, spaces, parts);
|
||||
blank_range(toks, sep_start..den_end, spaces, pieces, lits);
|
||||
return;
|
||||
}
|
||||
if has_whole {
|
||||
// A zero value shows `0` in a `?` or `0` units place, nothing in a `#` one.
|
||||
let units_hash = whole_idx.last().is_some_and(|&i| toks.get(i) == Some(&Tok::Digit('#')));
|
||||
let force = whole_s.is_empty() && num == 0.0 && !units_hash;
|
||||
fill_int(toks, &whole_idx, &whole_s, sec.thousands, force, parts);
|
||||
let whole_shown = whole_idx.iter().any(|&i| parts.piece(i).chars().any(|c| c.is_ascii_digit()));
|
||||
fill_int(toks, &whole_idx, &whole_s, sec.thousands, force, pieces);
|
||||
let whole_shown = whole_idx.iter().any(|&i| pieces.get(i).is_some_and(|p| p.chars().any(|c| c.is_ascii_digit())));
|
||||
if !whole_shown {
|
||||
let spaces = has_q(&whole_idx) || has_q(&num_idx);
|
||||
blank_range(toks, sep_start..num_start, spaces, parts);
|
||||
blank_range(toks, sep_start..num_start, spaces, pieces, lits);
|
||||
}
|
||||
}
|
||||
let num_s = if num < 1e15 { (num as u64).to_string() } else { fixed(num, 0).0 };
|
||||
fill_int(toks, &num_idx, &num_s, false, true, parts);
|
||||
parts.set_piece(slash, "/");
|
||||
fill_int(toks, &num_idx, &num_s, false, true, pieces);
|
||||
if let Some(p) = pieces.get_mut(slash) {
|
||||
*p = "/".into();
|
||||
}
|
||||
let den_s = den.to_string();
|
||||
if fixed_den.is_some() {
|
||||
parts.set_piece(den_start, &den_s);
|
||||
if let Some(p) = pieces.get_mut(den_start) {
|
||||
*p = den_s;
|
||||
}
|
||||
} else {
|
||||
// Denominator digits are zero padded to the `0` placeholders, then left aligned; `?`
|
||||
// pads with spaces after them.
|
||||
@@ -742,7 +631,9 @@ fn render_fraction(sec: &Section, x: f64, parts: &mut Parts) {
|
||||
} else {
|
||||
s
|
||||
};
|
||||
parts.set_piece(ti, &s);
|
||||
if let Some(p) = pieces.get_mut(ti) {
|
||||
*p = s;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
+22
-251
@@ -15,9 +15,6 @@ fn format_colors_resolve_named_and_indexed_rgb_safely() {
|
||||
|
||||
const S: DateSystem = DateSystem::D1900;
|
||||
|
||||
/// Built-in format 44 (accounting with a currency symbol, two decimals).
|
||||
const ACCOUNTING_2DP: &str = "_(\"$\"* #,##0.00_);_(\"$\"* \\(#,##0.00\\);_(\"$\"* \"-\"??_);_(@_)";
|
||||
|
||||
fn f(code: &str, n: f64) -> String {
|
||||
format_value(&Value::Number(n), &NumberFormat::parse(code), S).text
|
||||
}
|
||||
@@ -259,7 +256,7 @@ fn literals_and_fill() -> Result<(), Box<dyn std::error::Error>> {
|
||||
assert_eq!(f("(0)", 5.0), "(5)");
|
||||
let r = full("*-0", Value::Number(5.0));
|
||||
assert_eq!((r.text.as_str(), r.fill), ("5", Some(('-', 0))));
|
||||
let acct = ACCOUNTING_2DP;
|
||||
let acct = builtin_format(44).ok_or("none")?;
|
||||
let r = full(acct, Value::Number(0.0));
|
||||
assert_eq!((r.text.as_str(), r.fill), (" $- ", Some((' ', 2))));
|
||||
let r = full(acct, Value::Number(1234.5));
|
||||
@@ -359,7 +356,7 @@ fn kinds() -> Result<(), Box<dyn std::error::Error>> {
|
||||
assert_eq!(NumberFormat::parse("0.00").kind(), FormatKind::Number);
|
||||
assert_eq!(NumberFormat::parse("#,##0.00;[Red](#,##0.00)").kind(), FormatKind::Number);
|
||||
assert_eq!(NumberFormat::parse("\"$\"#,##0.00").kind(), FormatKind::Currency);
|
||||
assert_eq!(NumberFormat::parse(ACCOUNTING_2DP).kind(), FormatKind::Accounting);
|
||||
assert_eq!(NumberFormat::parse(builtin_format(44).ok_or("none")?).kind(), FormatKind::Accounting);
|
||||
assert_eq!(NumberFormat::parse("m/d/yyyy").kind(), FormatKind::Date);
|
||||
assert_eq!(NumberFormat::parse("h:mm AM/PM").kind(), FormatKind::Time);
|
||||
assert_eq!(NumberFormat::parse("0%").kind(), FormatKind::Percentage);
|
||||
@@ -372,11 +369,26 @@ fn kinds() -> Result<(), Box<dyn std::error::Error>> {
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn builtin_codes_render() {
|
||||
// Built-in ids 5, 12 and 47 (the id tables live with the file formats, in gridcraft-xlsx).
|
||||
assert_eq!(f("\"$\"#,##0_);\\(\"$\"#,##0\\)", -1234.0), "($1,234)");
|
||||
assert_eq!(f("# ?/?", 0.5), " 1/2");
|
||||
assert_eq!(f("mmss.0", 61.2 / 86400.0), "0101.2");
|
||||
fn builtins() -> Result<(), Box<dyn std::error::Error>> {
|
||||
assert_eq!(builtin_format(0), Some("General"));
|
||||
assert_eq!(builtin_format(14), Some("m/d/yyyy"));
|
||||
assert_eq!(builtin_format(49), Some("@"));
|
||||
assert_eq!(builtin_format(30), None);
|
||||
assert_eq!(builtin_format(50), None);
|
||||
assert_eq!(builtin_id("m/d/yyyy"), Some(14));
|
||||
assert_eq!(builtin_id("0.00%"), Some(10));
|
||||
assert_eq!(builtin_id("GENERAL"), Some(0));
|
||||
assert_eq!(builtin_id("[h]:mm:ss"), Some(46));
|
||||
assert_eq!(builtin_id("#,##0.00_);(#,##0.00)"), None);
|
||||
for id in 0..=60 {
|
||||
if let Some(code) = builtin_format(id) {
|
||||
assert_eq!(builtin_id(code), Some(id));
|
||||
}
|
||||
}
|
||||
assert_eq!(f(builtin_format(5).ok_or("none")?, -1234.0), "($1,234)");
|
||||
assert_eq!(f(builtin_format(12).ok_or("none")?, 0.5), " 1/2");
|
||||
assert_eq!(f(builtin_format(47).ok_or("none")?, 61.2 / 86400.0), "0101.2");
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[test]
|
||||
@@ -384,13 +396,6 @@ fn text_fn() {
|
||||
assert_eq!(text_function(&Value::Number(1234.5), "$#,##0.00", S), Ok("$1,234.50".into()));
|
||||
assert_eq!(text_function(&Value::from("5"), "0.00", S), Ok("5.00".into()));
|
||||
assert_eq!(text_function(&Value::from("abc"), "0.00", S), Ok("abc".into()));
|
||||
// Each `@` repeats the text: up to a cell's 32,767 characters, then #VALUE!.
|
||||
let at = "@".repeat(255);
|
||||
assert_eq!(text_function(&Value::from("x".repeat(128).as_str()), &at, S).map(|s| s.len()), Ok(128 * 255));
|
||||
assert_eq!(text_function(&Value::from("x".repeat(32_000).as_str()), &at, S), Err(CellError::Value));
|
||||
// The limit counts UTF-16 units: 16,384 emoji are 32,768 of them, one too many.
|
||||
assert_eq!(text_function(&Value::from("\u{1F600}".repeat(16_383).as_str()), "@", S).map(|s| s.chars().count()), Ok(16_383));
|
||||
assert_eq!(text_function(&Value::from("\u{1F600}".repeat(16_384).as_str()), "@", S), Err(CellError::Value));
|
||||
assert_eq!(text_function(&Value::Error(CellError::NA), "0", S), Err(CellError::NA));
|
||||
assert_eq!(text_function(&Value::Number(46302.0), "yyyy", S), Ok("2026".into()));
|
||||
assert_eq!(text_function(&Value::Number(0.5), "General", S), Ok("0.5".into()));
|
||||
@@ -897,240 +902,6 @@ fn closed_brackets_still_hold_their_semicolons() {
|
||||
assert_eq!(f("\"[\"0;0", 5.0), "[5");
|
||||
}
|
||||
|
||||
const JAN_1_2001: f64 = 36892.0;
|
||||
|
||||
#[test]
|
||||
fn b1_prefix_is_hidden_and_gregorian() {
|
||||
assert_eq!(f("B1dd/mm/yyyy", 46314.0), "19/10/2026");
|
||||
assert_eq!(f("b1dddd, mmmm d, yyyy", 46314.0), "Monday, October 19, 2026");
|
||||
assert_eq!(f("B1yyyy-mm-dd hh:mm", 46314.5), "2026-10-19 12:00");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn b2_prefix_renders_hijri_date() {
|
||||
assert_eq!(f("B2dd/mm/yyyy", JAN_1_2001), "06/10/1421");
|
||||
assert_eq!(f("b2d/m/yy", JAN_1_2001), "6/10/21");
|
||||
assert_eq!(f("B2dddd dd/mm/yyyy hh:mm", JAN_1_2001 + 0.75), "الإثنين 06/10/1421 18:00");
|
||||
assert_eq!(text_function(&Value::Number(46314.0), "B2dd/mm/yyyy", S).as_deref(), Ok("08/05/1448"));
|
||||
assert_eq!(f("B2e", 46314.0), "1448");
|
||||
assert_eq!(f("B2ee", 46314.0), "1448");
|
||||
assert_eq!(f("B2d/m/yyyy;@", 46314.0), "8/5/1448");
|
||||
}
|
||||
|
||||
/// Under `B2`, Excel shows Arabic weekday names, the full name for `ddd` as well as `dddd`.
|
||||
#[test]
|
||||
fn b2_prefix_names_weekdays_in_arabic() {
|
||||
// Monday 19 and Tuesday 20 October 2026.
|
||||
assert_eq!(f("B2dddd", 46314.0), "الإثنين");
|
||||
assert_eq!(f("B2ddd", 46314.0), "الإثنين");
|
||||
assert_eq!(f("B2dddd", 46315.0), "الثلاثاء");
|
||||
assert_eq!(f("B1dddd", 46314.0), "Monday");
|
||||
assert_eq!(f("B1ddd", 46315.0), "Tue");
|
||||
// Every weekday has a name.
|
||||
for v in 46314..46321 {
|
||||
assert!(!f("B2dddd", v as f64).is_empty());
|
||||
}
|
||||
}
|
||||
|
||||
/// `B2` on a General or text section changes nothing: Excel shows the plain serial.
|
||||
#[test]
|
||||
fn b2_prefix_on_general_and_text_sections() {
|
||||
assert_eq!(f("B2General", 46314.0), "46314");
|
||||
assert_eq!(f("B2@", 46314.0), "46314");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn b2_prefix_names_hijri_months() {
|
||||
assert_eq!(f("B2d mmmm yyyy", JAN_1_2001), "6 شوال 1421");
|
||||
assert_eq!(f("B2mmm", 37034.0), "ربيع\u{a0}الأول");
|
||||
assert_eq!(f("B2mmmm", JAN_1_2001 + 23.0), "شوال");
|
||||
assert_eq!(f("B2mmmm", JAN_1_2001 + 24.0), "ذو\u{a0}القعدة");
|
||||
assert_eq!(f("B2mmmmm", JAN_1_2001 + 24.0), "ذ");
|
||||
}
|
||||
|
||||
/// All twelve month names and seven weekday names as Excel 16.0 (build 20527, en-US) shows them,
|
||||
/// read with TEXT and a cell format: two-word month names are joined by a no-break space.
|
||||
#[test]
|
||||
fn b2_names_match_excel() {
|
||||
let months = [
|
||||
(46219.0, "صفر"),
|
||||
(46248.0, "ربيع\u{a0}الأول"),
|
||||
(46278.0, "ربيع\u{a0}الثاني"),
|
||||
(46307.0, "جمادى\u{a0}الأولى"),
|
||||
(46337.0, "جمادى\u{a0}الثانية"),
|
||||
(46366.0, "رجب"),
|
||||
(46396.0, "شعبان"),
|
||||
(46425.0, "رمضان"),
|
||||
(46455.0, "شوال"),
|
||||
(46484.0, "ذو\u{a0}القعدة"),
|
||||
(46514.0, "ذو\u{a0}الحجة"),
|
||||
(46543.0, "محرم"),
|
||||
];
|
||||
for (serial, name) in months {
|
||||
assert_eq!(f("B2dd", serial), "01", "{serial}");
|
||||
assert_eq!(f("B2mmmm", serial), name, "{serial}");
|
||||
assert_eq!(f("B2mmm", serial), name, "{serial}");
|
||||
assert_eq!(f("B2mmmmmm", serial), name, "{serial}");
|
||||
}
|
||||
// Sunday 18 to Saturday 24 October 2026.
|
||||
let days = ["الأحد", "الإثنين", "الثلاثاء", "الأربعاء", "الخميس", "الجمعة", "السبت"];
|
||||
for (k, name) in days.iter().enumerate() {
|
||||
let serial = 46313.0 + k as f64;
|
||||
assert_eq!(f("B2dddd", serial), *name, "{serial}");
|
||||
assert_eq!(f("B2ddd", serial), *name, "{serial}");
|
||||
}
|
||||
assert_eq!(f("B2yy", 46314.0), "48");
|
||||
assert_eq!(f("B2y", 46314.0), "48");
|
||||
assert_eq!(f("B2yyy", 46314.0), "1448");
|
||||
assert_eq!(f("B2yyyyy", 46314.0), "1448");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn b2_month_and_year_rollover() {
|
||||
assert_eq!(f("B2dd/mm/yyyy", 36915.0), "29/10/1421");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 36916.0), "01/11/1421");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 36974.0), "29/12/1421");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 36975.0), "01/01/1422");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 37683.0), "30/12/1423");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 37684.0), "01/01/1424");
|
||||
}
|
||||
|
||||
/// Cycle years 15 (1425, leap) and 16 (1426, not leap): Excel's leap year is 15, not 16. The
|
||||
/// 2 March 2005 and 29 January 2006 values were read from Excel by a reviewer; the year
|
||||
/// boundaries follow from them and the 354/355-day years. The same comparison covered every
|
||||
/// serial from 36,000 to 49,000 with no disagreement.
|
||||
#[test]
|
||||
fn b2_leap_year_is_cycle_year_15() {
|
||||
assert_eq!(f("B2yyyy-mm-dd", 38413.0), "1426-01-21"); // 2005-03-02, from Excel
|
||||
assert_eq!(f("B2yyyy-mm-dd", 38746.0), "1426-12-29"); // 2006-01-29, from Excel
|
||||
assert_eq!(f("B2dd/mm/yyyy", 38037.0), "29/12/1424"); // 2004-02-20
|
||||
assert_eq!(f("B2dd/mm/yyyy", 38038.0), "01/01/1425"); // 2004-02-21
|
||||
assert_eq!(f("B2dd/mm/yyyy", 38392.0), "30/12/1425"); // 2005-02-09
|
||||
assert_eq!(f("B2dd/mm/yyyy", 38393.0), "01/01/1426"); // 2005-02-10
|
||||
assert_eq!(f("B2dd/mm/yyyy", 38747.0), "01/01/1427"); // 2006-01-30
|
||||
// Anchors the reviewer read from Excel.
|
||||
assert_eq!(f("B2dd/mm/yyyy", 46314.0), "08/05/1448"); // 2026-10-19
|
||||
assert_eq!(f("B2dd/mm/yyyy", 46313.0), "07/05/1448"); // 2026-10-18
|
||||
assert_eq!(f("B2dd/mm/yyyy", JAN_1_2001), "06/10/1421");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 61.0), "29/10/1317"); // 1900-03-01
|
||||
assert_eq!(f("B2dd/mm/yyyy", 64970.0), "01/01/1501"); // 2077-11-16
|
||||
assert_eq!(f("B2dd/mm/yyyy", 64971.0), "02/01/1501"); // 2077-11-17
|
||||
assert_eq!(f("B2dd/mm/yyyy", 0.0), "29/08/1317");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 2958465.0), "03/04/9666"); // 9999-12-31
|
||||
}
|
||||
|
||||
/// Excel honours `B2` after a colour tag, but a `[$-...]` locale tag before it keeps the date
|
||||
/// Gregorian (the prefix is still hidden); `B2[$-401]` is Hijri.
|
||||
#[test]
|
||||
fn calendar_prefix_and_bracket_tags() {
|
||||
assert_eq!(f("[Red]B2dd/mm/yyyy", 46314.0), "08/05/1448");
|
||||
assert_eq!(f("[<0]0;B2dd/mm/yyyy", JAN_1_2001), "06/10/1421");
|
||||
assert_eq!(f("B2[$-401]dd/mm/yyyy", 46314.0), "08/05/1448");
|
||||
for code in ["[$-401]B2dd/mm/yyyy", "[$-409]B2dd/mm/yyyy", "[$-1010401]B2dd/mm/yyyy", "[$-ar-SA]B2dd/mm/yyyy"] {
|
||||
assert_eq!(f(code, 46314.0), "19/10/2026", "{code}");
|
||||
assert_eq!(text_function(&Value::Number(46314.0), code, S).as_deref(), Ok("19/10/2026"), "{code}");
|
||||
}
|
||||
assert_eq!(f("[Red][$-401]B1dd/mm/yyyy", JAN_1_2001), "01/01/2001");
|
||||
// Checked in Excel 16.0 (build 20527, en-US): a currency tag cancels `B2` like a locale tag,
|
||||
// a tag after the prefix does not, and a repeated prefix is accepted.
|
||||
assert_eq!(f("[$€-407]B2dd/mm/yyyy", 46314.0), "€19/10/2026");
|
||||
assert_eq!(f("[$$-409]B2dd/mm/yyyy", 46314.0), "$19/10/2026");
|
||||
assert_eq!(f("[$-401][Red]B2dd/mm/yyyy", 46314.0), "19/10/2026");
|
||||
assert_eq!(f("B2[Red]dd/mm/yyyy", 46314.0), "08/05/1448");
|
||||
for code in ["B2B2dd/mm/yyyy", "B2B2B2dd/mm/yyyy", "B2b2dd/mm/yyyy", "[Red]B2B2dd/mm/yyyy"] {
|
||||
assert_eq!(f(code, 46314.0), "08/05/1448", "{code}");
|
||||
}
|
||||
assert_eq!(f("B1B1dd/mm/yyyy", 46314.0), "19/10/2026");
|
||||
assert_eq!(f("B2 dd/mm/yyyy", 46314.0), " 08/05/1448");
|
||||
assert_eq!(f("B2\"x\"dd", 46314.0), "x08");
|
||||
for code in ["B1\"x\"", "B2\"x\""] {
|
||||
assert_eq!(f(code, 46314.0), "x", "{code}");
|
||||
}
|
||||
for code in ["B1General", "B1@", "0;0;0;B2@", "0;B2dd/mm/yyyy"] {
|
||||
assert_eq!(f(code, 46314.0), "46314", "{code}");
|
||||
}
|
||||
}
|
||||
|
||||
/// A quoted `B2` is literal text; any other `B` and digit makes Excel reject the code (`TEXT` is
|
||||
/// `#VALUE!`), as does a prefix before a number format. A rejected code shows as General. All
|
||||
/// checked in Excel 16.0 (build 20527, en-US), with TEXT and with a cell format.
|
||||
#[test]
|
||||
fn misplaced_calendar_prefix_is_rejected() {
|
||||
assert_eq!(f("\"B2\"dd/mm/yyyy", 46314.0), "B219/10/2026");
|
||||
assert_eq!(f("dd/\"B2\"/yyyy", 46314.0), "19/B2/2026");
|
||||
for code in [
|
||||
"B0dd/mm/yyyy",
|
||||
"B3dd/mm/yyyy",
|
||||
"B4dd/mm/yyyy",
|
||||
"B9dd/mm/yyyy",
|
||||
"B02dd",
|
||||
"B12dd",
|
||||
"B21dd",
|
||||
"B1B2dd/mm/yyyy",
|
||||
"B2B1dd/mm/yyyy",
|
||||
"B2[Red]B2dd/mm/yyyy",
|
||||
"B2 B2dd",
|
||||
"dd/mm/yyyyB2",
|
||||
"dd/mm/yyyyB1",
|
||||
"dd B2 mm",
|
||||
"dd b2 mm",
|
||||
"dd \\B2 yyyy",
|
||||
"B2\\B2dd",
|
||||
"B20.00",
|
||||
"B2#,##0",
|
||||
"B10.00",
|
||||
"B1#,##0",
|
||||
"0;B3dd",
|
||||
] {
|
||||
assert_eq!(text_function(&Value::Number(46314.0), code, S), Err(CellError::Value), "{code}");
|
||||
assert_eq!(f(code, 46314.0), "46314", "{code}");
|
||||
assert!(NumberFormat::parse(code).is_general(), "{code}");
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn b2_edges_of_the_date_range() {
|
||||
// Serial 60 is the 1900 system's fictitious 29 February: Excel has no Hijri date for it.
|
||||
assert_eq!(text_function(&Value::Number(60.0), "B2dd/mm/yyyy", S), Err(CellError::Value));
|
||||
assert_eq!(f("B2dd/mm/yyyy", 60.0), "########");
|
||||
assert_eq!(f("B1dd/mm/yyyy", 60.0), "29/02/1900");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 59.0), "28/10/1317");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 61.0), "29/10/1317");
|
||||
// So does any clock part (Excel 16.0, build 20527, en-US), but not an elapsed `[h]`, a literal,
|
||||
// or a locale tag that cancels `B2`.
|
||||
for code in ["B2hh:mm", "B2hh", "B2ss", "B2am/pm", "B2hh:mm AM/PM", "B2[$-401]hh:mm", "B2dd/mm/yyyy;\"x\""] {
|
||||
assert_eq!(text_function(&Value::Number(60.0), code, S), Err(CellError::Value), "{code}");
|
||||
assert_eq!(text_function(&Value::Number(60.5), code, S), Err(CellError::Value), "{code}");
|
||||
}
|
||||
assert_eq!(f("B2[h]", 60.0), "1440");
|
||||
assert_eq!(f("B2\"x\"", 60.0), "x");
|
||||
assert_eq!(f("[$-401]B2hh:mm", 60.0), "00:00");
|
||||
assert_eq!(f("B1hh:mm", 60.0), "00:00");
|
||||
assert_eq!(f("B2hh:mm", 61.0), "00:00");
|
||||
// Serial 0 shows as serial 1.
|
||||
assert_eq!(f("B2dd/mm/yyyy", 0.0), "29/08/1317");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 1.0), "29/08/1317");
|
||||
// No upper limit short of 31 December 9999: 16 and 17 November 2077, 31 December 9999.
|
||||
assert_eq!(f("B2dd/mm/yyyy", 64970.0), "01/01/1501");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 64971.0), "02/01/1501");
|
||||
assert_eq!(f("B2dd/mm/yyyy", 2958465.0), "03/04/9666");
|
||||
// In the 1904 system, serial 60 is a real day.
|
||||
assert_ne!(format_value(&Value::Number(60.0), &NumberFormat::parse("B2dd/mm/yyyy"), DateSystem::D1904).text, "########");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn b2_out_of_range_does_not_panic() {
|
||||
for v in [-1.0, 0.0, 60.0, 60.5, 2958465.0, 2958466.0, 1e300, f64::NAN] {
|
||||
let _ = f("B2dd/mm/yyyy mmmm dddd", v);
|
||||
let _ = text_function(&Value::Number(v), "B2dd/mm/yyyy mmmm dddd", S);
|
||||
}
|
||||
assert_eq!(f("B2dd/mm/yyyy", -1.0), f("dd/mm/yyyy", -1.0));
|
||||
assert_eq!(f("B2dd/mm/yyyy", 2958466.0), f("dd/mm/yyyy", 2958466.0));
|
||||
// Before 1 Muharram 1 AH there is no Hijri date (no serial reaches it).
|
||||
let before_epoch = gridcraft_core::date::DateTime { year: 600, month: 3, day: 9, ..Default::default() };
|
||||
assert!(calendar::Calendar::Hijri.date(&before_epoch).is_none());
|
||||
}
|
||||
|
||||
/// "Set precision as displayed": the value each format stores. Expected values observed in Excel
|
||||
/// (COM, `Workbook.PrecisionAsDisplayed = True`) with synthetic values.
|
||||
#[test]
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user