* CI: dev builds keep line tables only
* CI: one integration-test binary per crate
* CI: xtask runs tests with nextest, doctests apart, --keep-going on clippy and wasm
cargo xtask test runs the workspace tests (--lib --bins --tests) with cargo-nextest when it is installed, else cargo test, then the doctests. cargo xtask ci uses the same two steps.
* CI: docs name cargo xtask test
* CI: agent guides name the moved fuzz test
* CI: tests added on main join their crate's test binary
* Use chrono for UTC/calendar conversion instead of hand-rolled copies
Replaces the hand-rolled civil_from_days / days_from_civil copies and the
is_leap / days_in_month calendar helpers with chrono as a workspace dependency
(default-features = false, std only, so wasm is unaffected). The spreadsheet
serial-date semantics are untouched: Excel's 1900 fictitious leap day (serial
60 = 1900-02-29), the 1899-12-30 / 1904-01-01 serial offsets, Excel's weekday
numbering and the 1900/1904 systems keep their explicit special cases in
serial_from_ymd / datetime_from_serial / days_in_month_in. chrono only backs
the plain proleptic-Gregorian day-number <-> date conversion and the
leap-year / month-length helpers those build on; a small fallback keeps the old
answer for dates outside chrono's NaiveDate range. The duplicate local is_leap
in the financial functions now uses the core one. Output is byte-identical and
pinned by the existing date, parse, financial and logging tests. AGENTS.md now
says time goes through chrono.
Follow-up to storytold/cadcraft#1785
* Never panic on out-of-range clocks or years; keep calendar results outside chrono's range
- logging::timestamp: chrono's From<SystemTime> unwraps, so a clock past year
262143 panicked inside the logger. Use the fallible DateTime::from_timestamp
and read such a clock as the epoch, like the old saturating code. Years past
9999 are written with {:04} (not %Y, which prefixes a '+').
- date::is_leap: restore the Gregorian leap-year rule for years outside chrono's
NaiveDate range (e.g. is_leap(400000) is true again) via direct arithmetic.
- date::days_in_month: compute the December boundary with checked_add so
days_in_month(i32::MAX, 12) no longer overflows, and compute month lengths
through days_from_civil's fallback so out-of-range years keep their real
lengths (e.g. days_in_month(400000, 2) is 29). Invalid months still return 0.
- civil_from_days: add missing docstring.
- AGENTS.md: note the narrow out-of-range fallback exception to the chrono rule.
* Don't build an unrepresentable SystemTime in the far-future test
UNIX_EPOCH + Duration::from_secs(u64::MAX / 2) overflows Windows'
SystemTime and panicked before timestamp() ran, failing the Windows
test job. Build it with checked_add and only assert where it exists.
* Route out-of-range dates through chrono's 400-year cycle, not hand-rolled copies
The PR kept hand-rolled Hinnant `days_from_civil_fallback`/`civil_from_days_fallback`
copies for dates outside chrono's ±262143-year range, plus a hand-rolled
Gregorian leap rule in `is_leap` and an i32::MAX special-case in `days_in_month`.
AGENTS.md carved a "narrow fallback" exception for them, which left calendar
arithmetic outside chrono after all.
Replace all of it with the Gregorian 400-year cycle (146_097 days): chrono
converts the day/year within one cycle and whole cycles are added back via i128
(so the i32/i64 extremes can't overflow). `is_leap` and `days_in_month` reduce
the year with `rem_euclid(400)` (leap years and month lengths repeat every 400
years), which also removes the December `i32::MAX` overflow special-case.
chrono is now the only calendar logic, and every output is unchanged: verified
the cycle result equals the old formula over wide ranges and at the i32/i64
extremes with a scratch chrono program. The Excel 1900 leap bug (serial 60 =
fictitious 1900-02-29) and 1904 semantics are untouched -- `serial_from_ymd`/
`datetime_from_serial` only ever hand the converters in-range years, so their
serials are byte-for-byte the pre-chrono ones (pinned by a new test).
New tests pin `days_from_civil`/`civil_from_days` past chrono's range on both
sides of the epoch, round-tripping, with values from the old formula; the
existing leap/month-length-beyond-range test still passes unchanged. Removed the
AGENTS.md "narrow fallback" exception.
Tested: cargo test -p gridcraft-core -p gridcraft-functions; fmt --check;
clippy --all-targets -D warnings.
* time: preserve out-of-range years in the log timestamp instead of the epoch
`DateTime::from_timestamp(..).unwrap_or_default()` changed the existing output: a
clock set far ahead (an input the old hand-rolled civil-from-days converted)
collapsed to `1970-01-01`. Restore the old conversion domain with a cycle-based
helper: reduce by whole 400-year Gregorian cycles (146_097 days), convert the
in-range remainder with chrono, and carry the year back as i128. Only the year
leaves chrono's range; month/day/time stay exact. Pre-1970 still clamps to the
epoch as before. chrono stays the only calendar logic (no new features).
apps/gridcraft logging.rs `timestamp` (SystemTime, u64 secs), pinned via a
platform-safe `UNIX_EPOCH.checked_add`:
u64::MAX/2 s → "292277026596-12-04T15:30:07.000Z" (was "1970-01-01..."),
10e12 s → "318857-05-20T17:46:40.000Z", 253_402_300_800 s → "10000-01-01...".
* Name the 400-year range extension as the one exception to the time rule
The rule said never to hand-roll calendar math, but the range fix
(shifting by whole 400-year Gregorian cycles so dates past chrono's year
range keep their old output) is cycle arithmetic, so the rule and the code
disagreed. The rule now names that exception: the library still computes
month, day and time, and only whole cycles are carried in the year.
* Say the library computes month and day; time-of-day is plain arithmetic
Narrows the range-extension wording: the date library computes the
calendar fields, and splitting seconds into hours, minutes and seconds
is ordinary arithmetic, which is what the code does.
- Shell: [[ ]] for conditional tests, positional parameters copied into
named locals, explicit return at the end of the env.sh helpers.
- PowerShell: the desktop-shortcut smoke test writes its progress with
Write-Output instead of Write-Host.
- Web: the viewport no longer sets user-scalable=no, so people can zoom
the page.
Home > Format as Table on a cell or range inside an existing table
failed with "A table can't overlap another table." Excel applies the
picked style to that table instead. The command now returns that table
and its range, as it does for a new table. A range crossing a table's
edge is still rejected.
Without a range, the active area of the selection decides, as it does
for Insert > Table. The native Format menu's Format as Table now runs
home.formatAsTable instead of insert.table.
A script line whose parameters sat under a key the runner does not read ran the
command with no parameters at all, and the run still exited 0:
{"command":"home.insertSheet","args":{"name":"Report"}}
{"command":"home.insertSheet","name":"Report2"}
made Sheet2 and Sheet3 instead of Report and Report2, and said nothing. A typo
in a script produced a valid, wrong workbook.
A line is read for `command` (or `id`) and `params`; any other key is a usage
error now, naming the keys and the line, and the run stops before the workbook
is written:
gridcraft-cli: s.jsonl:2: unknown key `args`: a line is {"command": "id", "params": {…}}
The usage text says what a line looks like, which is the other half of the
report — the format was not written down anywhere, so the reporter found it by
trying `args`, a flat object and `id` in turn.
Tested at both levels: `parse_command` for the keys, their sorted list when
there are several, and the forms that are read (`id`, `params`, `id={json}`,
comments); and a run of a four-line script that stops at line 2 without writing
the file, then the same script corrected, whose sheets come out named. Without
the check the mistaken line parses again.
An unknown key *inside* `params` is still ignored — the rest of #2256 — and
needs the command's parameter list, which the engine does not expose yet. Noted
on the issue.
A request that waited more than 60 s was answered with a timeout and left in
the queue, so `control::poll` ran it on the next UI pass and dropped the reply.
The client was told its command had failed while the command was applied: an
agent that retried after the error ran it twice, and one that gave up left an
unexpected change in the workbook.
`ControlRequest` now carries who has it -- pending, taken by the app, or
abandoned by its sender -- as the report suggests. `poll` claims each request
and skips one it cannot claim; the connection thread abandons the request when
its reply times out, and says which happened:
still queued timed out waiting for the UI (is the window minimised?)
already started timed out waiting for the UI, which had already started the
request, so it may still finish; its outcome is unknown
The second case cannot be called a failure without guessing. This is the shape
DeckCraft settled on for the same bug (its #719), so the two apps now agree.
Fix: #1914
A line that is not a JSON object with a string method, including an HTTP
request line, bytes that are not UTF-8 or a line over 64 MiB, gets one
error reply and the connection closes before anything after it is read.
Request lines are read with a bound instead of buffering them whole.
Business Central's "Open in Excel" exports write booleans as
<c><f>TRUE()</f><v>1</v></c> with no t="b", so the cached value reads back
as the number 1 and the calc oracle reported every such cell as a
`number → bool` mismatch. A cached 1 or 0 now matches GridCraft's TRUE or
FALSE, and the cells are counted apart (`boolAsNumber` in --json, a note in
the text report) so a real number/logical disagreement stays visible.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GridCraft imports XLSB (#101) and ODS (#100) files, and #321 added both
to the Linux and Windows "Open with" lists, but the macOS Info.plist
still declared only .xlsx, .xlsm, .csv and .tsv, so Finder's Open With
didn't offer GridCraft for them. Add two CFBundleDocumentTypes entries
(Editor, rank Alternate) with the UTIs macOS reports for these files,
plus an imported UTI for .xlsb like the .xlsm one. A test now checks
that the macOS and Windows packaging offer the same file types.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
In 0.5.0 a click anywhere on the Home ribbon's Paste button (label,
icon or dropdown arrow) froze the app for good. home() called
paste.secondary_clicked() inside ctx.data_mut(); after any click egui
answers that with ctx.input(), which takes the context lock the closure
already holds, so the UI thread waited on itself. Sampling the hung
0.5.0 process on macOS shows the main thread parked in
RawRwLock::lock_exclusive_slow under Context::write <- data_mut <-
ribbon::home. #272 moved the query out of the closure; it landed after
the 0.5.0 release, which is the build both reporters used.
No test ever clicked the Paste button itself, which is how this
shipped. The new test clicks it, then opens its menu with a secondary
click and picks Paste Values, on a worker thread with a timeout so a
regression fails instead of hanging the run. With #272 reverted it
fails (egui's debug lock check panics after 10 s); with it, it passes.
The Paste button gets an accessible "Paste" name so the test (and
screen readers) can find it, the same way Pictures and Border presets
are named.
With the freeze gone, the button still did not do what it says: it
pasted GridCraft's own copy only (GridState::system_clipboard was never
assigned), so text copied in another app gave "Nothing to paste." The
ribbon button, its options menu's Paste, the cell and header context
menus and the macOS Edit menu now read the system clipboard through a
new paste_text host service and paste it like Cmd/Ctrl+V does, falling
back to GridCraft's own copy when there is no text (or no service, as
on the web). Asking egui for the clipboard instead (RequestPaste) was
tried and stalled the macOS event loop until the next input event.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Excel stores the path of a linked workbook relative to the file that uses it, and rewrites it when the file is saved into another folder. The path was kept as text, so after Save As it named a file next to the new copy: a missing link, or another workbook with the same name. Relative externalLinkPath targets are now resolved against the folder the workbook was read from and written relative to the folder it is saved into (absolute when there is no relative path, on another drive).
A page that frames GridCraft on the same origin can now open workbooks
in it and store what it saves, through postMessage, instead of Save and
Export becoming browser downloads. Built for a Nextcloud integration,
but nothing in it is host-specific.
Only the web shell changes:
- gridcraft:open from the parent goes through the inbox, like a picked
or dropped file
- Services::download sends the bytes to the parent (gridcraft:save); an
.xlsx workbook handed over counts as saved under that name
- gridcraft:dirty reports unsaved changes; gridcraft:ready and
gridcraft:failed report start-up
- in host mode no sample opens, and ?author= names the user for notes
and comments
Enabled only by ?host=parent inside a frame; only the parent window on
the page's own origin is heard or answered. Without it nothing changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01G6GtUCaPiCMEQCN4QcdQnp
Brandon Thomas