auto_row_height grew a row for every line break in a cell's text whether
or not the cell wrapped. Excel counts breaks only when Wrap Text is on:
with it off a multi-line value displays on one line and AutoFit gives a
one-line row. A file whose unwrapped cells hold pasted multi-line text
(the shape generating libraries write) opened with tall rows and nothing
visibly tall in them.
The fast path no longer treats a line break as a reason to measure, and
the measuring path lays text out over lines only for a wrapping cell.
Typing a break still turns wrap on (the Alt+Enter rule in cell entry) and
grows the row, as before.
Test: an unwrapped multi-line cell from a file keeps the default height
on open; wrap on grows the row; wrap off returns it; a typed break wraps
and grows. The open-time fit tests from #833 are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`=1+` reported "unexpected `)`", although no `)` was typed: entry closes
missing parentheses by appending up to eight `)`, and when every retry
failed it reported the last retry's error. It now reports the error of
the formula as typed ("unexpected end of formula").
`gridcraft_formula::parse_closing_parens` holds the convenience and takes
the parse function, so cell entry and the UI's Spanish formula path share
it instead of each having its own loop.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The theme watcher (DesktopTheme) reads the desktop's light/dark choice
from the XDG Desktop Portal over a blocking D-Bus call with no timeout.
zbus waits forever by default, so a portal that accepts the call and
never answers stalled the watcher's thread on its first read, and "Use
system setting" stopped following the desktop for the rest of the
session, even after the portal recovered.
The session connection is now built with a 2 s method timeout (REPLY,
shorter than the 4 s POLL). A timed-out call returns InputOutput, which
read() already handles by dropping the connection, so the next poll
asks again on a fresh one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0115HyvML9WnuFRdmtMcLP5t
The Japanese catalog still lived inline in i18n.rs with its ribbon and widget
labels, so with the interface set to 日本語 the dialog titles and labels, message
boxes, Backstage/File-view items, template names and native menus showed in
English. This moves it to i18n/ja.rs and adds the 545 keys the Brazilian
Portuguese catalog translates beyond it: 943 keys, the coverage pt-BR has
(~43% of ~2,200 strings), in Excel for Mac 日本語版 wording.
CORE keeps the shared reference set every catalog covers, verbatim and in the
same order (the cross-catalog coverage test still points at it); DIALOGS holds
the new strings, and the lookup checks it first. One test parses pt.rs's source
so the two catalogs cannot drift apart again, one checks that every key is
unique across the two lists and every value is translated, and one spot-checks
the high-value strings.
Docs: the Japanese row and the "Where we are" bullet in
docs/localization-parity.md (943, ~43%), a revision row, the ROADMAP Languages
paragraph and progress log, and gaps.md §11.
xl/sharedStrings.xml can carry phonetic runs (<rPh sb="0" eb="2"><t>ふりがな</t></rPh>)
and a <phoneticPr> beside a string item. GridCraft read only the text and wrote every
string back as <si><t>…</t></si>, so a plain open + save deleted a Japanese workbook's
furigana.
Both are now read — offsets validated against the base text, runs with bad or missing
offsets, sb >= eb or no reading dropped with a warning, and never used to index — and
kept on the workbook as Workbook::phonetics, one entry per string text. The writer
emits them next to the <si> of the same text. The saved table holds each text once, so
two identical texts cannot keep two different readings: the reader keeps the first and
warns. Strings without phonetic data are written byte for byte as before.
Behaviour reference: ISO/IEC 29500-1 §18.4.6 rPh and §18.4.7 phoneticPr (Microsoft
Learn PhoneticRun/PhoneticProperties). Tests: crates/xlsx/src/tests/phonetic_tests.rs
(round trip, plain string, hostile offsets, shared text, unused data, comments and
inline strings). Docs: file-format-parity.md, localization-parity.md.
A theme's <a:fontScheme> carries latin, ea (East Asian) and cs (complex
script) typefaces for each of the major and minor fonts. The model held
only the latin ones and the writer emitted <a:ea typeface=""/> and
<a:cs typeface=""/> unconditionally, so a Japanese workbook's theme fonts
(e.g. 游ゴシック, ＭＳ Ｐゴシック) were silently deleted on open + save.
Theme gains major_font_ea, minor_font_ea, major_font_cs and
minor_font_cs (#[serde(default)], so older GridCraft JSON still loads;
empty means the file did not set one). The XLSX reader and writer
round-trip them escaped, and a theme without them writes exactly the
empty attributes it wrote before.
Rendering and font resolution are unchanged: the UI does not consume ea
yet, and wiring it in is a follow-up.
The French catalog once shipped with over a hundred translations sitting
on their neighbours' keys ("Close" read "Société"); main has since fixed
them. A catalog shifted that way still covers every key, so the existing
coverage tests don't catch it.
A new test checks every language's translations keep the shape of their
English: the same {placeholders}, and no line break in a one-line label
(the two big ribbon buttons that wrap are exempt). Plus spot checks of
French labels that were once misplaced.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The left end of a flat operator chain is not nesting, so (((...1)))+1 is
accepted whenever (((...1))) is, in Spanish entry and display.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
along: =1+1+... with 4,096 operands (8,192 characters, exactly Excel's
limit, which Excel accepts and calculates) overflowed the stack when
entered, because the derived Clone of Expr recursed once per operand
down the chain's left spine. Saving hit the same in the XLSX writer.
- Expr: Clone, PartialEq, walk and map follow the left spine in a loop
and recurse only into right operands and real nesting.
- XLSX: _xlpm. prefixing and the implicit-intersection scan follow the
spine too. The scan's depth limit of 64 wrote any chain of more than
about 64 operands as a dynamic-array formula; such sums are plain
formulas again (ranges in the chain still get array evaluation).
- Evaluate Formula borrows sub-expressions instead of cloning each
operator's subtree (quadratic memory on a long chain).
- Spanish formula entry and display no longer refuse chains of more
than 256 operands; the left spine does not count as nesting.
Tests run on a 1 MiB stack (the Windows main thread) with Excel's
observed results: formula (clone/eq/walk/map/adjust at 4,096), calc
(401 and 4,096 operands, -, *, /, ^, &, %, comparisons, 1,500 refs),
xlsx (long chains written as scalar formulas), engine (cell.set, fill,
copy/paste, insert rows/columns, undo/redo, Evaluate Formula, XLSX save
and reopen), Spanish entry. Each fails without the fix.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sheet names starting with '.' or reading as R1C1 (R, C, RC, R1C1...) were saved
unquoted, so Excel could not open the file; PivotTable sources and ADDRESS left
cell-like names (A1) bare. The quoting rule, unquoting and sheet splitting now
live once in gridcraft_core::addr and replace the copies in the formula printer,
pivot, ADDRESS, chart, xlsx, Solver, Go To and the Name Box (which now reads ''
as ').
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Lower-casing then upper-casing made dotless ı equal to i (both become I),
which Unicode's default case folding keeps apart; LET(i,1,ı,2,i) could
return the ı binding. same_name now lower-cases and adds the folds lower
case misses (ß/ẞ -> ss, ſ -> s, ς -> σ). Tests cover ı/i and the Greek
final sigma.
app.language.set's engine.commands docs and its unknown-language error
were written by hand and had fallen behind: the docs missed es, fr and
sr, the error missed de, fr and sr. Both are now built from
Language::ALL (Language::code_list), and tests check every language is
listed in each.
iter_range walks the sparse store's bands for the area's row range
only, so a chain of N formulas no longer rescans every stored cell of
the sheet per area.
Co-Authored-By: Claude <noreply@anthropic.com>
- gridcraft_formula::same_name lower-cases then upper-cases each
character, so ß matches ẞ (to_uppercase alone maps ß to SS).
- Formula-bar input uses it too, so a binding such as ſi is not
translated to IF while the evaluator binds it.
- Tests: ISOMITTED with a case variant, ß/ẞ, and the Spanish input case.
- The region already starts at the header row when headers are on, so
the dropdown label reads start.row for both cases (was showing the
first data value, e.g. B: 30 instead of B: Revenue).
- + Add Level picks the first region column no level uses yet instead
of columns.get(levels.len()), which re-seeded an already-used column.
Co-Authored-By: Claude <noreply@anthropic.com>
Excel matches LET and LAMBDA names case-insensitively in every script,
so =LET(dług,100,DŁUG*2) is 200. The evaluator compared them ASCII-only
and returned #NAME?. The three lookups (name evaluation, ISOMITTED, a
LET variable shadowing a function name) now fold Unicode case.
The outline glides 0.12 s toward each new selection. While selecting with the mouse that makes
it trail the pointer, which reads as lag (the grid itself paints in under a millisecond per
frame). With a mouse button held it is now drawn on its cells at once, as in Excel; keyboard
moves still glide.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: MH <q6rrzz@gmail.com>
A formula at the nesting limit reading many uncalculated formulas (for
example SUM over a wide range) suspended the abandoned evaluations once per
cell read, restarting them each time. The other formulas read at the limit
are now collected and calculated together with the suspended one, before the
abandoned evaluations restart. Results, circular-reference and iteration
behaviour are unchanged. Adds a regression test (3000 suspensions before, 1
after).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The deref doc comment still said blocks are trimmed to the used range. Describe
the actual rule: blocks up to FULL_SIZE_CELLS, or up to FULL_SIZE_BOUNDED_CELLS
when they do not run to the sheet's edge, keep their full shape; the rest stop
at the used range. Docs only.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
`cargo xtask ci` has not passed on macOS or Windows since cc21643:
error: this let-binding has unit value
--> crates/ui-egui/src/desktop_theme.rs:36:13
36 | let mut link = Link::default();
`Link` is the session-bus connection the watcher keeps between reads. Off Linux
there is no bus to keep, so it was aliased to `()`, which makes that binding a
unit binding: `clippy::let_unit_value` reports it and the gate's `-D warnings`
turns it into an error. CI runs Linux only, where `Link` is a real
`Option<Connection>`, so nothing catches it there.
Off Linux `Link` is now a unit struct rather than `()`. The binding has a type,
the lint does not apply, and nothing else changes: `read` still ignores it on
those platforms. An `allow` on the line would also work, but a named type says
what the thing is rather than silencing a lint that only fires on some
platforms.
`SheetProtection` carries eleven permissions. `format_cells` is checked in
`apply_style`, and an edit to a locked cell is caught after the fact by
`check_protection`, but `sort` and `autofilter` had no reader anywhere outside
`sheet_read.rs`, `sheet_write.rs` and a round-trip test: they only ever
travelled through the file. So a protected sheet could still be sorted, and
Ctrl+Shift+L still put an AutoFilter on it.
`refuse_unless_allowed` sits beside `locked_cell_error`, with three call sites:
`do_sort` (which both Sort… and the A-to-Z buttons come through), `toggle_filter`
and `filter_by`.
Where the box decides is worth knowing. Sorting a range of locked cells was
already refused, by the locked-cell guard rather than the permission, because a
sort moves values between locked cells — which is Excel's rule too ("users
cannot sort ranges that contain locked cells on a protected worksheet,
regardless of this setting"). So the Sort box decides for a range that has been
unlocked, while an AutoFilter changes no cell's contents and the box is the only
thing that ever decided it. The test asserts both, and that an unprotected sheet
still sorts and filters.
Not fixed here: Delete Rows and Delete Columns are refused even when their boxes
are ticked, because deleting removes locked cells and trips the same post-hoc
guard, while Insert Rows and Insert Columns work. Letting the guard know which
operation ran is a change to code every command passes through and wants its own
PR.
It returned an empty string for every workbook, saved or not, so the classic
sheet-name formula `MID(CELL("filename",A1),FIND("]",…)+1,31)` gave #VALUE!:
with no `]` in the string there is nothing for FIND to find. Excel returns
`<folder>\[<file>]<sheet>`, and empty only before the first save.
The model had no file to report. `source_dir` could not serve: it names the
folder relative links to other workbooks are relative to, and must keep naming
it after a Save As into another folder, because the links held in memory are
still written for the old one — only the saved copy is rebased. `source_path`
is therefore its own field, with that distinction written down.
It is set when a workbook is opened, from the path made absolute so the folder
is the one the user sees, and when a full-fidelity save makes a file the
document's own — before the `saved` snapshot, or the workbook would read as
dirty the moment it was saved, and only when it changes, so an ordinary Save
does not clone the workbook for nothing. A workbook that arrived as bytes knows
its name but never a folder and reports `[name]Sheet`, which still splits on the
`]`; an import reports the file it came from, as Excel does. The sheet is the one
the reference is on, not the active one.
CELL is volatile and already in the volatile sets, so a cell holding it answers
from the next recalculation; after a Save As that means F9 rather than the save
itself, which the engine test asserts rather than papers over. Making a save
recalculate is a larger decision than this issue.
`check` still skips CELL("filename") as volatile, as docs/cli.md says: the
cached value in a file is the folder of the machine that saved it.
The Sort dialog's 'Sort by' box was a free-text field hard-coded to
'A', so sorting a region in column B silently sorted column A unless
the user noticed and retyped the letter. It now seeds from the column
the cursor is in and renders a dropdown of the current region's
columns, labelled with their header text. '+ Add Level' seeds the
next region column instead of a hard-coded B.
Closes#1906
Co-Authored-By: Claude <noreply@anthropic.com>
desired_rows(1) only sets the field's minimum height; it grows with its
text. Pins that an Alt+Enter edit in a cell without Wrap Text shows all of
its lines and reserves no more. Review feedback on #1974.
quantmind-br