Files
wordcraft/crates/layout/src/table.rs
T
+17 cc7316e871 Tables: a row fits on a page only with room for the border below it (#3814) (#3815)
* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Tables: a row fits on a page only with room for the border below it

Word places a table row on a page only when the border under it fits too:
that border belongs to the next row's top band, but the row is moved on when
the page has its height and not the border. WordCraft checked the row's height
alone, so a page could hold one row more than Word's.

RowLayout::edge is the width of the bottom border of a row that another row
follows; place_table counts it when it checks whether a row fits, and so does
the check that keeps repeated header rows from staying alone at the foot of a
page. The part of a split row left on the page has none (it ends the page).

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* #3944: drop the wall-clock assert from the long-root-tag test

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>
Co-authored-by: dfgodvan <dfgodvan@gmail.com>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>

* Write a text-box story shared by many shapes only once on docx save (#3983)

Fixes #3653

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Keep pending formatting revisions when Style Inspector clears formatting (#3971)

Fixes #3671

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the headless ui_requests queue and keep one Paste Special request (#3997)

Fixes #3641

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Validate PNG extension in file.exportPng (#3837)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* changes to filter dialogue

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-au…

* Khmer: ignore w14:ligatures when shaping Khmer text (#3921)

Documents that Word saves with w14:ligatures="historicalDiscretional" (every Khmer run
in typical Khmer OS files) turned liga and clig off. Khmer OS fonts build their
syllables there, so pre-base vowels and stacked signs came out in the wrong order
(ស៊ើបអង្កេត drawn as សីបអង្កតេ). Khmer text is now shaped with the font's defaults,
as Word does.

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Skip quoted attribute values when scanning settings.xml tags (#3954)

Fixes #3682

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Rename kept docx parts whose names are not valid OPC part names (#3947)

* Rename kept docx parts whose names are not valid OPC part names

Fixes #3691

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its re…

* #3637: A table cell span from a .json file no longer overflows (#3927)

Insert Column, Merge Cells and Delete Column did arithmetic on the raw
CellProps::span, which a .json document can set to any u32, so a span
near u32::MAX overflowed (a panic in debug, a wrapped span in release).
They now work from the clamped span() and cap the result at MAX_COLS.

* #3663: "NaN" in the Font dialog's Size box no longer closes the app (#3928)

* #3663: "NaN" in the Font dialog's Size box no longer closes the app

"NaN" parses as an f32 and clamp keeps NaN, so the preview font size was
NaN and epaint's glyph assertion panicked while painting the dialog. The
preview now drops non-finite sizes and shows 12 pt.

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrel…

* Load ui.json with many unknown keys in linear time (#3968)

* Load ui.json with many unknown keys in linear time

Fixes #3659

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so…

* docbin: mark the format-mandated SHA-1 as reviewed (SonarCloud S4790) (#4020)

MS-OFFCRYPTO §2.3.5 derives the RC4 CryptoAPI key with SHA-1, so the
digest cannot be strengthened without breaking Word-written .doc files.
The NOSONAR marker records that this use is required by the format.

* Cap and deduplicate the .aff TRY alphabet at parse time (#3981)

Fixes #3654

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Render: a dimmed header's borders keep one tone where they meet (#2946) (#3725)

#2947 fixed the fade by drawing a dimmed layer inside a single opacity layer,
but left no regression test for the tone: where a header table's borders cross,
a crossed or shared rule would quietly come out darker than a rule that is not
drawn on top of anything, and the undimmed view and the PDF export would not
agree with the screen.

The test renders the same page with the header dimmed and undimmed and checks
three rules: the darkest pixel at a crossing of a cell's top and left rules,
the darkest pixel on a border two neighbouring cells share, and one on an
outer border well clear of both. With the fade done as one layer, the
crossing and the shared border keep the tone of an isolated rule; measured
against a clean full-strength render, the dimmed tone is 0.45 of the
background, not 0.45 squared.

* Refuse arrange.group when members exceed the group child cap (#3958)

Fixes #3679

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Ignore docx chart number-format codes longer than 255 bytes (#3963)

Fixes #3729

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Format Painter: show the copy pointer over the text while armed (#3453) (#3560)

* Format Painter: show the copy pointer over the text while armed (#3453)

The canvas chose the I-beam over text whether or not the Format Painter
held copied formatting, so nothing showed the tool was waiting for a
click. While it is armed, the pointer over the text is now the copy
pointer (egui has no paintbrush), the same choice as the open DeckCraft
fix for the same report (storytold/deckcraft#846). Pictures, table
handles and resize edges keep their own pointers.

* Format Painter: the selection bar keeps its arrow while the painter is armed

Review follow-up: the bar's arrow shows that a click there selects whole
lines, so it outranks the painter's copy pointer (#3453).

* Format Painter pointer test: the selection bar keeps its arrow while armed

Review follow-up (#3453): the test now also hovers the selection bar with
the painter armed, so a precedence regression there fails it.

* Canvas: document show()

Review follow-up (#3453): CodeRabbit's docstring check counts the
function the pointer change sits in.

* #3453: Format Painter pointer test requires the I-beam once the painter is off

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* #3453: Format Painter pointer test expects the hidden pointer where the I-beam is drawn

On Windows the canvas hides the system I-beam and draws its own (#1746), so
the pointer over the text reads None there, not Text.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Ruler: dragging a margin edge leaves the view where it is (#3900)

Each step of a margin drag on a ruler runs layout.margins, and like any
command that brought the caret back into view. With the caret off screen
(dragging the bottom margin while the text being edited is near the top
of the page) the view scrolled mid-drag: the page slid under the pointer,
the pointer's distance from the page's top changed with it, and the margin
ran far past where it was dropped. A drag of about 1 cm on the vertical
ruler set a 2 cm bottom margin to 8.1 cm.

The margin handles now run the command with run_keep_view, as table border
drags do (#1738).

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Save, export and print keep the view where it is (#3895)

* Save, export and print keep the view where it is (fixes #3864)

* File › Options keeps the view where it is too

Changing a setting such as spell checking moves neither the caret nor
the text, so it no longer scrolls back to the caret.

* Layout: an empty paragraph with a hidden mark shows nothing, not even its list number (#3982) (#3992)

* Layout: an empty paragraph with a hidden mark shows nothing, not even its list number (#3982)

A paragraph whose mark is hidden and whose text is empty or hidden too is
now laid out as nothing, like a paragraph deleted in the final text: no
room, no list number and no shading. Its number is still counted, as in
Word ("1. Alpha", hidden item, "3. Beta"). Before, the body gave it no
room but drew its number under the next item's, and a table cell gave it
a line of its own with the number.

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* #3816: also assert the next paragraph's first line stays on page 0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>

* Parse the Flat OPC part root tag once when inheriting namespaces (#3944)

* Parse the Flat OPC part root tag once when inheriting namespaces

Fixes #3727

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* Engine: no dead-code warnings on wasm (#3542)

* fix(engine): do not treat decimal points as sentence boundaries in Sentence case

In crates/engine/src/cmd/format.rs, convert_chars treated any period as a sentence boundary, causing numbers with decimal points (e.g. 3.14, 2.5) to capitalize the subsequent word in Sentence case.

Check if a period is surrounded by ASCII digits, and do not trigger uppercase capitalization when encountering a decimal point.

Closes #3485

* proof: drop U+FE0F from FORBIDDENWORD and NOSUGGEST flags too

Review follow-up: the directives that name one flag read the same emoji
form, so they now skip the variation selector like PFX/SFX and stem flags.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fmt: rustfmt equation array centring

* #3108 review: behind-text and in-front-of-text floats don't grow the row

FredApps' review: the row-growth loop counted every non-inline object, so a
page-anchored logo or watermark behind a cell's text would stretch its row to
the object's bottom edge - contradicting table.rs (behind-text frames count
for vertical alignment but don't increase the row's flow height) and Word.
Only floats that take part in the text flow (Square, Tight, Through,
TopAndBottom) grow the row now, the same classification layout already uses
elsewhere for "takes part in wrapping".

New test: a 300pt behind-text and a 300pt in-front-of-text picture each leave
the row at its text height; it fails with the old condition. The existing
Square-wrap tests and cell_frame_centres_empty still pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Engine: no dead-code warnings on wasm

`cargo check -p wordcraft-engine --target wasm32-unknown-unknown` warned
about five items whose only callers are platform-gated:
- citations `by_author`, user_templates `MAX_ENTRIES` and `label_of`: used
  only by the desktop (non-wasm) source import and template listing.
- speech `voice_language`: used by espeak (Unix but macOS) and the macOS
  voice picker, so `any(unix, test)`, not `any(not(windows), test)` (which
  also took in wasm).
- speech `primary_tag`: used by those plus the Windows synthesizer, so
  `any(unix, windows, test)`.

Each item now carries exactly its callers' gate rather than a blanket
allow(dead_code). Checked with cargo check on Linux, Windows (gnu), macOS
and wasm32 (zero warnings each), clippy -D warnings, and the engine tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Integration: #1847 reject comment IDs above u32::MAX

* Integration: #3267 leave footnote definitions out of link definitions

* Integration: #3471 drop duplicate egui options block; install_ui_fonts sets them

* Integration: #2153 drop trailing blank line and use assert! for bool checks in tab_separated_list_markers

* Integration: #2268 ODT lengths accept only the standard pc unit for picas

* Integration: #2269 LaTeX lengths keep macro names case-sensitive

* Integration: #2673 Khmer substitutes return the (list, flag) pair after #3245

* Integration: #2273 Latin weight/style and line styles take precedence over Asian, complex and line-type attributes

* Integration: #1851 DOCPROPERTY accepts the reference app's LastSavedBy name

* Integration: #1985 only ^0nnn codes read as Windows-1252

* Integration: #2416 collapse the nested vertical-align if (clippy::collapsible_if)

* Integration: #2185 hyperlink_target slices with get() and needs the whole field name

* Integration: #2034 objects without a stacking order draw above the ones before them

* Integration: #3434 drop review-bot notes from test doc comments

* Integration: #3419 doc comments back on their own items (EquationFrame and its test above them)

* Integration: #2334 test a formula string result continued in a CONTINUE record

* Integration: #3429 the based-on-chain test's doc comment back above its test

* Integration: #2453 drawing_payload test builds the shape fill as a Fill

* Integration: #3257 separate its Find-box test from #2407's recent-files test (auto-merge interleaved them)

* Integration: cargo fmt

* Integration: #1728 close the layout test the merge resolution left open

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: cargo fmt

* Integration: #2722 Dutch registry comment on its own line

* Integration: #3203 English text for the keys the Romanian catalog lacks

Its complete-catalog test needs every key of every other catalog; the merged PRs added 72 since it was written. Rows carry the English text until translated.

* Integration: #1076 English text for the keys the Slovenian catalog lacks

Its complete-catalog test needs every key of every other catalog. Rows carry the English text until translated.

* Integration: status lines list each change once

Combining the tracks' status lines repeated a few entries (#2852, #1500, #2617).

* Integration: cargo xtask parity

* Integration: cargo fmt

* Integration: #1832 drop a useless .into() in the end-of-document citation test

clippy::useless_conversion failed cargo xtask ci.

* Integration: #2244 VML on/off values read case-insensitively by the shape fill reader

Main (#1500) moved VML fill reading into shape_style.rs, where filled="False" still counted as on; #2244's test vml_shape_wrap_side_and_anchor_case_insensitivity expects no fill.

* Integration: #2041 English text in the Polish catalog for the keys Italian gained

polish_covers_every_key_of_every_other_catalog expects Italian to have only "Style: APA" beyond Polish; #2041 translated 14 newer strings into Italian. Rows carry the English text until translated.

* Integration: #2722 English text for four labels the Dutch and Danish catalogs lack

every_tab_and_command_is_translated needs Format Shape Pane, Full Screen View, Personal Templates and Underline Color, added by other merged PRs. Rows carry the English text until translated.

* Integration: #3335 header Esc test allows the half-pixel return scroll that #3201 brings

#3201 makes entering an empty header scroll to it; Esc then scrolls back to the caret, landing 0.5 px from the start (17291.0 vs 17290.5). The test checks the view is back within a pixel.

* IME: the caret follows the composition text (#3556)

* Header and footer: a click in the other band edits it there (#3557)

* README: the app icon above the title

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouched, and reading the document — the inspector
without removals — keeps working. A document that does not track changes
is inspected as before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The inspector's tracking refusal doesn't cite an unrelated issue

#1344 is about the inspector missing hidden text, not about its removals
going untracked, so the reference only sent a reader to the wrong place.

Co-Authored-By: opencode <noreply@opencode.ai>

* A version snapshot belongs to the document it was taken from

`Session::set_document` resets everything about the open document — the
document id, the document, the selection, the undo history, the password,
the file name — but not the version snapshots. A snapshot taken of one
document therefore survived `file.new`, `file.open`, `file.newFromTemplate`
and `file.recover`, and `file.versions {"restore": index}` then put the
previous document's content into the open one.

The snapshots go with the document they were taken from: `set_document`
clears them, so a restore can only ever bring back a version of the
document that is open. Every caller replaces the document with another
one (a new or opened file, a recovered copy, a merge result, a sheet of
labels), so none of them loses anything it needs.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the inspector's tracking refusal and the version snapshot's scope

Co-Authored-By: opencode <noreply@opencode.ai>

* A comparison clears the version snapshots of the document it replaces

`review.compare` and `review.combine` open the original with `s.doc = doc`,
which bypasses `set_document` — so the version snapshots of the document the
comparison replaces survived it, and a restore could put the previous document
back over the comparison result.

The snapshots go with the document they were taken from, as everywhere else.
The document itself is still assigned directly: `set_document` resets the undo
history, and the comparison's own undo step would go with it.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3277: Test Inspector refusal without protection

* #1327: Keep snapshots with their document through Compare

* #1327: Preserve file identity after rejected document replacement

* test: mark the portal test's blocking sleep as intentional (SonarCloud) (#3989)

The future is polled by hand with no async runtime, so a blocking sleep
is how the test models a slow dialog.

🤖 Generated with Codebuff

Co-authored-by: Codebuff <noreply@codebuff.com>

* #3542: gate the label_of test off wasm32

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>

* Drop U+FFFE and U+FFFF from ODT export text (#3817)

Found while testing crates/formats/src/odt.rs:49-63 (fn x, the ODT text escaper; filter at line 58)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Cap the number of VML path commands parsed from a path attribute (#3951)

Fixes #3683

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* Return None from parse_number when the value overflows f64 (#3577)

Found while testing crates/doc/src/locale_number.rs:104

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Fix spurious 零 after a group's trailing zeros in Chinese numerals (#3745)

Found while testing crates/doc/src/cjk_numbers.rs:47-55 (Counting::format)

Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>

* picture.corrections: a tiny sharpen value no longer panics in the blur (#3646) (#3830)

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* Reject signs and non-digits in parse_date fields (#3575)

Found while testing crates/engine/src/cmd/controls.rs:113-123 (parse_date)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take (#3816)

* Restrict Editing: a file command that changes the document is checked too

`always_allowed` waved through every `file.*` id, so on a protected document
the file commands that do change it went unchecked: the privacy options and
the properties wrote document settings, and the inspector's removals deleted
content. The privacy options' own comment already said they were "refused on
a protected document"; that wasn't true.

The allow-list now names the file commands that don't change the open
document, and the ones that do are checked like any other command. Reading
the document stays allowed: the inspector without removals, the privacy
settings, the info, and saving a version snapshot.

Restoring a stored version replaces the whole document, and a snapshot's own
protection is stale — restoring one lifted the protection and dropped its
password. The live document's protection, password and formatting limit now
survive a restore.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the user name can't change under protection

`file.setAuthor` sets the name a protected document's named exception
ranges are checked against, and the author new revisions and comments are
recorded under. `always_allowed` waved through every `file.*` id, so on a
protected document the name could be changed: an agent could take the
identity of a named editor and unlock their exception range, and forge the
authorship of the revisions it leaves behind.

A protected document refuses the change now, with the rule Word uses for a
content control's properties: stop protection first. The check is in the
command itself — `file.setAuthor` is a pure command, so `check_before`
never sees it — and in `check_before` for the day it isn't.

Word lets the user name change under protection: it is a preference there.
This project doesn't, because that identity gates what may be edited.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: a protected document's protection needs an explicit stop

The guard in `restrict` refused a protected document only when its
protection had a password, so a password-less one could have its
protection changed — and `file.protect {"mode": "none"}` stopped it
outright. An agent could strip a user's protection silently and then edit
the document freely.

Two rules now, as Word has them (Stop Protection first):

- The protection of a protected document doesn't change, whatever the
  mode. The formatting-only protection doesn't count: it stores the mode
  "none", and it limits the styles, not the editing.
- Stopping the protection is explicit. The "none" mode is the shorthand
  for it, but on a protected document only `{"stop": true}` stops it — the
  way the pane's Stop Protection button does it — so a command that
  merely names a mode can't strip it.

The password still guards the stop, and the formatting-only path (the
mode "none" with the formatting restriction) still works.

`cmd::tools::tests::protection_blocks_edits` stops the protection with
`{"mode": "none"}`; it has to stop it explicitly now.

Co-Authored-By: opencode <noreply@opencode.ai>

* Restrict Editing: the post-check compares the settings and the metadata

`outside_unchanged` is the check that rolls a command back when it
changed anything outside the editable regions. It compared the text, the
formatting, the page setup, the document defaults and the styles — but
not the document's settings, its metadata, its custom properties, the
resolved flag of its comments, its list definitions or its bibliography
sources. A command that changed the metadata of a protected document was
accepted and never rolled back: `file.properties {"title": "Forged
title"}` on a protected document with an exception changed the title and
stayed.

The settings, the core properties, the custom properties, the comments'
resolved flags, the list definitions and the sources are compared now, so
a command that changes them is rolled back like any other change outside
the editable regions.

The revisions and the media store stay out of it on purpose: an edit
inside an exception legitimately leaves a revision, and a picture
inserted there adds to the media store. The list definitions and the
sources are compared like the styles: the ones the document had must be
unchanged, but a new one may be added — a list or a citation inserted
inside an editable region adds its definition, like applying a style adds
one.

Saving is unaffected: it writes the metadata, but `file.save` is a pure
command and isn't post-checked, and the protection itself is changed by
`review.restrict`, which is always allowed.

Co-Authored-By: opencode <noreply@opencode.ai>

* The protection is stopped explicitly, and the docs say so

The test that stopped the protection used the "none" shorthand, which a
protected document now refuses; it stops it the explicit way instead, and
pins that the shorthand doesn't stop it. `file.protect` says the same: it
starts enforcement, and "none" doesn't stop a protected document.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the Restrict Editing gaps that are closed

Co-Authored-By: opencode <noreply@opencode.ai>

* A password guards the protection whatever its mode

The guard that stops the protection of a protected document from changing
asked whether the editing was restricted, so a formatting-only protection —
the mode "none", which limits the styles rather than the editing — was
exempted. A formatting-only protection can carry a password, and that password
guarded it: it was exempt from the guard, so `{"mode": "readOnly", "password":
"other"}` replaced the protection and the password with ones of the caller's
choosing, and the protection could then be stopped with them.

A password guards the protection whatever its mode, as it did before.

Co-Authored-By: opencode <noreply@opencode.ai>

* The CLI's --author names the session, it doesn't change the document

`--author` dispatched `file.setAuthor`, which a protected document refuses
now that the name is checked as the identity its exceptions are matched
against — so a script that named itself on a protected file failed before it
ran a single command.

The flag names the session, not the open document, so it sets the name
directly, with the same normalisation and blank-name check the command uses.
`file.setAuthor` keeps its protection check for the callers that mean to
change the document's identity.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the progress docs' status lines carry today's date

Co-Authored-By: opencode <noreply@opencode.ai>

* The document's properties and a version restore need the protection stopped

Two gaps the review found, both of them a change to a protected document that
went through without the protection seeing it.

The properties (`file.properties`) and the privacy options
(`file.privacyOptions`) are document-wide state, and they aren't tracked: under
tracked changes they changed the document with no revision to show for it, and
under the other modes they were only rolled back by the post-check, after the
command had run. They are refused now, before the command runs, as the content
control's properties are.

A version restore (`file.versions`) replaces the whole document, including the
content the protection kept read-only, and it did that without the password:
`file.versions` is a pure command, so the protection never saw it. Keeping the
live protection and its password stopped them being dropped, but the content
underneath was still replaced. A restore needs the protection stopped first,
and the password with it when the document has one.

The live protection still wins over the snapshot's stale one, so a restore
can't lock the user out of their own document either.

Co-Authored-By: opencode <noreply@opencode.ai>

* Layout: keep with next takes the 0.01 pt of slack the other fit checks take

place_para_in_flow moved a paragraph kept with the next one to a new page
when the two needed more than the room left by any amount, while lines, rows
and floats fit with 0.01 pt to spare. A kept paragraph whose need came to a
rounding error over the page bottom moved on alone.

* Release 0.10.0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* The inspector's removals are refused in every protected mode

`file.inspect` with removals takes content out of the document — the
comments, the revisions, the properties, the hidden text, the headers.
Under readOnly and comments it was refused, or rolled back by the
post-check where the document had an exception; under trackedChanges
nothing stopped it, so a document that tracks changes could have its
comments and revisions taken out with no revision to show for it.

The removals are refused now, in every mode a protected document can be
in, next to the properties and the privacy options. An inspection without
removals only reads the document, and still works.

The protection test covers the modes now, not only readOnly.

Co-Authored-By: opencode <noreply@opencode.ai>

* Docs: the session author, not the user name

"the user name can't change while the document is protected" was the
wrong term twice over: it is the session author — the name new tracked
changes and comments are recorded under, and the one a named exception
is matched against — and it is `file.setAuthor` that refuses to change
it. The CLI's `--author` names the session directly, so it works on a
protected document.

The parity row, the two status lines and the two progress rows now say
that.

Co-Authored-By: opencode <noreply@opencode.ai>

* #3279: Keep protected session author fixed in CLI and Options

* The Document Inspector refuses its removals while the document tracks changes

The inspector's removals delete straight from the document: the hidden
text, the comments, the revisions and the headers go without ever reaching
the engine's tracked-delete path. On a document that tracks changes — and
`trackedChanges` protection turns tracking on — the content was gone with
no revision to show for it, and the revision trail was broken behind the
user's back.

The removals are not tracked, so they cannot be: rather than untrack them
silently, the inspector refuses the whole removal form while the document
tracks changes, and says why ("the document tracks changes: stop tracking
before removing content"). The check comes before anything is removed, so
the document is left untouch…

---------

Co-authored-by: Mat Strange <mat@matstrange.net>
Co-authored-by: opencode <noreply@opencode.ai>
Co-authored-by: dfgodvan <180286157+dfgodvan@users.noreply.github.com>
Co-authored-by: Brandon Thomas <brandon@storyteller.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Michael Mu <69072633+MichaelMuClaude@users.noreply.github.com>
Co-authored-by: Divyanshu Gupta <divyanshu2074@gmail.com>
Co-authored-by: dmim2502-commits <234869083+dmim2502-commits@users.noreply.github.com>
Co-authored-by: vangel76 <vangel1976@googlemail.com>
Co-authored-by: aiongg <aiongtaigi@gmail.com>
Co-authored-by: aiongg <67543387+aiongg@users.noreply.github.com>
Co-authored-by: bflatastic <bombayapirat@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Co-authored-by: Joseph Anglada <fluorine@users.noreply.github.com>
Co-authored-by: Michael Higgins <55243596+Higgs32584@users.noreply.github.com>
Co-authored-by: In Mean <6434741+inmean@users.noreply.github.com>
Co-authored-by: Jayesh Prajapati <156836276+Jayesh-Prajapati9@users.noreply.github.com>
Co-authored-by: Gabriel Matozo <69991354+GabrielMatozo@users.noreply.github.com>
Co-authored-by: Mathieu Rancourt <40959576+MatRanc@users.noreply.github.com>
Co-authored-by: pakiat <49967840+pakiat@users.noreply.github.com>
Co-authored-by: Marco Arnulfo <103679419+marcoarnulfo@users.noreply.github.com>
Co-authored-by: Menonro <nitayd2001@gmail.com>
2026-10-23 12:00:11 -04:00

1346 lines
62 KiB
Rust

//! Table layout: column widths, cell boxes, row heights, merged cells, borders and shading,
//! table-style conditional formatting. Rows are laid out as free-standing boxes; pagination
//! moves whole rows or splits them between lines (`split_row`).
use wordcraft_doc::numbering::Counters;
use wordcraft_doc::props::{Align, Border, Borders, HeightRule, Rgb, TextDirection, VAlign, VMerge};
use wordcraft_doc::styles::{PartSet, TablePart, TableStyleProps};
use wordcraft_doc::{Block, Blocks, Paragraph, Row, StoryRef, Table};
use wordcraft_geom::Rect;
use crate::para::{CellText, ClKind, ParaLayout};
use crate::{BoxContext, BoxLayout, Ctx, Placed, layout_box};
/// Largest preferred table width, as a share of the text column, that a file can ask for.
const MAX_WIDTH_SHARE: f32 = 10.0;
pub struct RowLayout {
pub height: f32,
/// Items relative to the table's top-left (x) and the row's top (y).
pub items: Vec<Placed>,
/// The room the row keeps below its text: the bottom cell margin and border band. Part of a
/// row split across pages keeps it too.
pub bottom: f32,
/// The width of the border below a row that another row follows (it belongs to the next
/// row's top band). Word wants it on the page too: the row fits only where its height and
/// this border do.
pub edge: f32,
/// The row's cells: the items each one owns and where its content ends.
pub cells: Vec<CellPart>,
}
/// A cell of a [`RowLayout`].
pub struct CellPart {
/// Top cell margin and visible border band retained on a continuation.
pub(crate) top: f32,
/// Its items in the row's `items` (its fill, text, borders and cell area).
pub items: std::ops::Range<usize>,
/// Where its content ends below the row's top, its last paragraph's space after included
/// (what the part of the row left on a page must hold); `None` for turned text and a cell
/// covered by a vertical merge.
pub end: Option<f32>,
}
pub struct TableLayout {
/// Table x offset within the column.
pub x: f32,
pub width: f32,
pub rows: Vec<RowLayout>,
}
const DEFAULT_MARGINS: [f32; 4] = [0.0, 5.4, 0.0, 5.4];
/// The longest line turned text gets when its row height isn't exact (a Letter page's body
/// height): longer text wraps.
const TURNED_MAX: f32 = 648.0;
/// The longest line turned text ever gets (Word's largest page, 22 inches).
const TURNED_LIMIT: f32 = 1584.0;
/// A cell whose text is turned, between the two passes of [`layout_table`].
struct Turned {
turn: TextDirection,
/// Line length it was laid out with, and how far its lines reach across the cell.
len: f32,
across: f32,
/// List counters and equation count before the cell, to lay it out again the same way.
snap: (Counters, u32),
/// Its table style parts and cell index in the row.
region: Region,
ci: usize,
/// Top margin plus top border band.
top: f32,
}
/// How long the lines laid out in `items` are without the room they didn't use: the length
/// turned text needs so that nothing wraps.
fn natural_length(items: &[Placed]) -> f32 {
let mut len = 0.0f32;
for it in items {
let Placed::Lines { para, l0, l1, x, .. } = it else { continue };
let indents = para.rp.indent_left.max(0.0) + para.rp.indent_right.max(0.0) + para.rp.indent_first.max(0.0);
for l in para.lines.get(*l0..*l1).unwrap_or(&[]) {
let (mut lo, mut hi) = (f32::MAX, f32::MIN);
for v in &l.vis {
lo = lo.min(v.x);
hi = hi.max(v.x + v.w);
}
if l.vis.is_empty()
&& let (Some(a), Some(b)) = (l.xs.first(), l.xs.last())
{
(lo, hi) = (*a, *b);
}
let hy = l.hyphen.map_or(0.0, |h| h.2);
if hi >= lo {
len = len.max(x.max(0.0) + hi - lo + hy + indents);
}
}
}
// A little slack, so laying the text out again at this length doesn't wrap it.
len + 0.5
}
/// The table's style with its based-on chain merged.
fn table_style(ctx: &Ctx, t: &Table) -> Option<TableStyleProps> {
ctx.doc.styles.table_style(t.props.style.as_deref()?)
}
/// The table's default cell margins: its own, else its style's, else Word's.
fn default_margins(t: &Table, style: Option<&TableStyleProps>) -> [f32; 4] {
t.props.cell_margins.or(style.and_then(|s| s.parts.cell_margins)).unwrap_or(DEFAULT_MARGINS)
}
/// The margins (top, left, bottom, right; points) of cell `cell` of row `row`: its own, else the
/// table's defaults. The layout uses the left and right ones as they are (the app's ruler
/// measures a cell's indents from them) but gives every cell in a row the row's largest top and
/// bottom margins.
pub fn cell_margins(doc: &wordcraft_doc::Document, t: &Table, row: usize, cell: usize) -> [f32; 4] {
let style = t.props.style.as_deref().and_then(|id| doc.styles.table_style(id));
let defaults = default_margins(t, style.as_ref());
t.rows.get(row).and_then(|r| r.cells.get(cell)).map_or(defaults, |cell| cell.props.resolved_margins(defaults))
}
/// The most paragraphs [`measure_table_columns`] lays out: a huge table is measured by its first
/// rows only, so AutoFit stays quick.
const MEASURE_BUDGET: usize = 4000;
/// The line length paragraphs are measured at: wide enough that nothing real wraps.
const MEASURE_WIDE: f32 = 15_840.0;
/// What an empty cell's text needs (about a paragraph mark), points.
const EMPTY_TEXT: f32 = 12.0;
/// What one paragraph needs: the longest run of text without a break opportunity (`min`) and
/// its longest line when nothing wraps (`max`), indents included.
fn measure_para(p: &wordcraft_doc::Paragraph, env: &crate::para::ParaEnv) -> (f32, f32) {
use crate::para::ClKind;
let pl = crate::para::layout_para(p, env);
let rp = &pl.rp;
let indents = rp.indent_left.max(0.0) + rp.indent_right.max(0.0) + rp.indent_first.max(0.0);
let (mut word, mut min) = (0.0f32, 0.0f32);
for c in &pl.clusters {
match c.kind {
ClKind::Text | ClKind::Object(_) => word += c.adv.max(0.0),
ClKind::Marker => {}
_ => {
min = min.max(word);
word = 0.0;
}
}
if c.break_after {
min = min.max(word);
word = 0.0;
}
}
min = min.max(word);
let mut max = 0.0f32;
for l in &pl.lines {
let (mut lo, mut hi) = (f32::MAX, f32::MIN);
for k in l.c0..l.c1 {
let (Some(c), Some(a), Some(b)) = (pl.clusters.get(k), l.cl_left(k), l.cl_right(k)) else { continue };
lo = lo.min(a.min(b));
if !matches!(c.kind, ClKind::Space | ClKind::Marker | ClKind::LineBreak | ClKind::PageBreak | ClKind::ColumnBreak) {
hi = hi.max(a.max(b));
}
}
if hi > lo {
max = max.max(hi - lo + l.hyphen.map_or(0.0, |h| h.2));
}
}
let min = if min.is_finite() { min } else { 0.0 };
let max = if max.is_finite() { max.max(min) } else { min };
(min + indents, max + indents)
}
/// What each grid column of `t` needs for AutoFit to Contents, as (narrowest, widest) in points
/// with cell margins included: the narrowest fits each cell's longest word (text that can't
/// wrap), the widest fits every paragraph on one line. A cell spanning several columns spreads
/// what its columns lack over them. Paragraphs are laid out with the table style's formatting;
/// after [`MEASURE_BUDGET`] paragraphs the remaining rows are skipped.
pub fn measure_table_columns(doc: &wordcraft_doc::Document, t: &Table) -> Vec<(f32, f32)> {
use wordcraft_doc::Block;
let style = t.props.style.as_deref().and_then(|id| doc.styles.table_style(id));
let margins_def = default_margins(t, style.as_ref());
let ncols = t.cols().clamp(1, wordcraft_doc::table::MAX_COLS);
let fields = crate::FieldCtx::default();
let mut cols = vec![(0.0f32, 0.0f32); ncols];
// Cells spanning several columns: (first column, span, min, max), applied after the others.
let mut spanned: Vec<(usize, usize, f32, f32)> = Vec::new();
let mut budget = MEASURE_BUDGET;
'rows: for (ri, row) in t.rows.iter().enumerate() {
let mut g = row.first_grid();
for cell in &row.cells {
let span = cell.span();
let margins = cell.props.resolved_margins(margins_def);
let side = margins[1].max(0.0) + margins[3].max(0.0);
let text = style.as_ref().map(|st| region_text(st, region_of(t, style.as_ref(), ri, g, span, ncols)));
let env = crate::para::ParaEnv {
doc,
width: MEASURE_WIDE,
label: None,
fields: &fields,
show_hidden: false,
field_codes: false,
hide_deleted: Default::default(),
table: text.as_ref(),
proofing: None,
typing_word: None,
// Measurement uses a fixed environment (like show_hidden/proofing above).
numeral: crate::NumeralMode::default(),
exclusions: &[],
eq_number: 0,
read_line_spacing: None,
grid: crate::para::Grid::default(),
fit_objects: false,
fitted: &[],
};
let (mut min, mut max) = (EMPTY_TEXT, EMPTY_TEXT);
if cell.props.vmerge != VMerge::Continue && !cell.props.text_direction.is_turned() {
for b in &cell.blocks {
match &**b {
Block::Para(p) => {
if budget == 0 {
break 'rows;
}
budget -= 1;
let (a, b) = measure_para(p, &env);
min = min.max(a);
max = max.max(b);
}
// A nested table keeps its own column widths.
Block::Table(inner) => {
let w: f32 = inner.grid.iter().filter(|w| w.is_finite()).map(|w| w.clamp(0.0, 1584.0)).sum();
min = min.max(w);
max = max.max(w);
}
}
}
}
let (min, max) = ((min + side).min(1584.0), (max + side).min(1584.0));
if span == 1 {
if let Some(c) = cols.get_mut(g) {
c.0 = c.0.max(min);
c.1 = c.1.max(max);
}
} else {
spanned.push((g, span, min, max));
}
g += span;
}
spanned.extend(row_gaps(row, g).map(|(g, span, w)| (g, span, w, w)));
}
for (g, span, min, max) in spanned {
let Some(cs) = cols.get_mut(g..(g + span).min(ncols)) else { continue };
if cs.is_empty() {
continue;
}
let n = cs.len() as f32;
let (have_min, have_max) = cs.iter().fold((0.0, 0.0), |(a, b), c| (a + c.0, b + c.1));
for c in cs.iter_mut() {
c.0 += ((min - have_min) / n).max(0.0);
c.1 += ((max - have_max) / n).max(0.0);
c.1 = c.1.max(c.0);
}
}
// A column no cell starts in still needs room for its margins.
let floor = margins_def[1].max(0.0) + margins_def[3].max(0.0) + EMPTY_TEXT;
cols.iter_mut().for_each(|c| {
c.0 = c.0.max(floor);
c.1 = c.1.max(c.0);
});
cols
}
/// Column widths for AutoFit to Contents, the way Word shares the room: every column gets its
/// widest content when that all fits in `avail`; otherwise each gets its narrowest plus a share
/// of the remaining room in proportion to how much wider its content would like to be. When
/// even the narrowest widths don't fit, the columns keep them (the table runs past the margin,
/// as in Word).
pub fn autofit_widths(cols: &[(f32, f32)], avail: f32) -> Vec<f32> {
let clean = |v: f32| if v.is_finite() { v.clamp(0.0, 1584.0) } else { 0.0 };
let cols: Vec<(f32, f32)> = cols.iter().map(|&(a, b)| (clean(a), clean(b).max(clean(a)))).collect();
let avail = clean(avail);
let (sum_min, sum_max) = cols.iter().fold((0.0f32, 0.0f32), |(a, b), c| (a + c.0, b + c.1));
if sum_max <= avail {
return cols.iter().map(|c| c.1).collect();
}
if sum_min >= avail || sum_max <= sum_min {
return cols.iter().map(|c| c.0).collect();
}
let k = (avail - sum_min) / (sum_max - sum_min);
cols.iter().map(|c| c.0 + (c.1 - c.0) * k).collect()
}
/// The first cell's left margin: how far its text sits inside the table's edge.
pub(crate) fn first_cell_left_margin(ctx: &Ctx, t: &Table) -> f32 {
first_cell_left_margin_in(t, default_margins(t, table_style(ctx, t).as_ref()))
}
fn first_cell_left_margin_in(t: &Table, def: [f32; 4]) -> f32 {
t.rows.first().and_then(|r| r.cells.first()).map_or(def, |cell| cell.props.resolved_margins(def))[1]
}
/// What a table style gives the cells in one set of parts: shading, vertical alignment and each
/// part's cell borders, lowest precedence first.
struct RegionLook {
fill: Option<Rgb>,
v_align: Option<VAlign>,
borders: Vec<(TablePart, Borders)>,
}
fn region_look(st: &TableStyleProps, region: Region) -> RegionLook {
let mut out = RegionLook { fill: None, v_align: None, borders: Vec::new() };
for part in region.iter() {
let f = st.part(part);
out.fill = f.fill.or(out.fill);
out.v_align = f.v_align.or(out.v_align);
// The whole table's borders are the table borders, applied on their own.
if part != TablePart::WholeTable
&& let Some(b) = f.borders
{
out.borders.push((part, b));
}
}
out
}
/// A cell's borders from the table style's parts it is in, later parts winning. Row parts give
/// their left/right edges at the table's sides and `inside_v` between cells; column parts their
/// top/bottom edges at the table's top and bottom and `between` between rows; corner cells their
/// four edges.
fn region_borders(parts: &[(TablePart, Borders)], (ri, nrows): (usize, usize), (first_g, last_g): (bool, bool)) -> Borders {
let mut out = Borders::default();
for (part, b) in parts {
let b = *b;
let mapped = if part.is_row() {
Borders {
top: b.top,
bottom: b.bottom,
left: if first_g { b.left } else { b.inside_v },
right: if last_g { b.right } else { b.inside_v },
between: None,
inside_v: None,
}
} else if part.is_column() {
Borders {
top: if ri == 0 { b.top } else { b.between },
bottom: if ri + 1 >= nrows { b.bottom } else { b.between },
left: b.left,
right: b.right,
between: None,
inside_v: None,
}
} else {
Borders { between: None, inside_v: None, ..b }
};
out.overlay(&mapped);
}
out
}
pub fn layout_table(ctx: &mut Ctx, story: StoryRef, t: &Table, path: &[u32], avail: f32, depth: usize) -> TableLayout {
let style = table_style(ctx, t);
// Cell text formatting and looks per set of style parts, built once per table.
let mut cell_text: Vec<(Region, CellText)> = Vec::new();
let mut looks: Vec<(Region, RegionLook)> = Vec::new();
let ncols = t.cols().max(1);
// Column widths.
let mut grid: Vec<f32> = if t.grid.len() == ncols {
// Grid units can be narrower than a text box, including omitted row-end columns.
t.grid
.iter()
.map(|&w| match w {
w if !w.is_finite() => 72.0,
w if w <= 0.0 => 4.0,
w => w,
})
.collect()
} else {
vec![avail / ncols as f32; ncols]
};
apply_cell_widths(t, &mut grid);
let mut total: f32 = grid.iter().sum();
// A top-level table in an inherited header or footer takes its share of the defining
// section's width (`Ctx::pct_base`).
let pct_base = if depth <= 1 { ctx.pct_base.unwrap_or(avail) } else { avail };
let target = match (t.props.width_pct, t.props.width) {
// Word keeps a percentage past 100%; the cap only bounds hostile input.
(Some(p), _) if p > 0.0 => Some(pct_base * (p / 100.0).min(MAX_WIDTH_SHARE)),
(_, Some(w)) if w > 0.0 => Some(w),
_ => None,
};
if let Some(tw) = target {
if total > 0.0 && total < 1.0 {
// Divide first: a tiny positive total can overflow the scale factor.
grid.iter_mut().for_each(|w| *w = (*w / total) * tw);
} else {
let k = tw / total.max(1.0);
grid.iter_mut().for_each(|w| *w *= k);
}
total = tw;
}
// A table with its own width keeps it, however far it runs past the margins (as in Word).
if target.is_none() && total > avail * 1.5 && !t.props.fixed {
let k = avail / total.max(1.0);
grid.iter_mut().for_each(|w| *w *= k);
total = avail;
}
if !t.props.fixed {
// Measuring lays paragraphs out once more; it mustn't count their equations twice.
let eq_count = ctx.eq_count;
let content = column_content(ctx, t, style.as_ref(), depth);
ctx.eq_count = eq_count;
if let Some(content) = content {
autofit(&mut grid, &content, t, avail);
// Never wider than the room it had, however long an unbreakable word.
let fitted: f32 = grid.iter().sum();
let cap = avail.max(total);
if fitted > cap {
grid.iter_mut().for_each(|w| *w *= cap / fitted);
}
total = grid.iter().sum();
}
}
let mut colx = Vec::with_capacity(ncols + 1);
let mut acc = 0.0;
for w in &grid {
colx.push(acc);
acc += w;
}
colx.push(acc);
// Word 2013 and later (compatibility mode 15) put the table's border at the margin (plus its
// indent); earlier modes line the first cell's text up with it instead.
let margins_def = default_margins(t, style.as_ref());
let indent = t.props.indent.unwrap_or(0.0);
// The style's borders, overlaid by the table's own side by side.
let mut tborders = style.as_ref().and_then(|s| s.parts.borders).unwrap_or_default();
if let Some(own) = t.props.borders {
tborders.overlay(&own);
}
let first_cell = t.rows.first().and_then(|r| r.cells.first());
// A right-to-left table (`w:bidiVisual`) mirrors: its first column stands at the right, its
// start alignment and indent are measured from the right margin, and each cell's start
// (left) margin and border are on its right.
let rtl = t.props.bidi_visual;
// How far in from the start margin the table's edge sits when it is start-aligned.
let start_offset = if ctx.doc.settings.compat_mode >= 15 {
// The border is centred on the edge, so half of it sits outside: Word moves the table in
// by that half.
let border = first_cell.and_then(|c| c.props.borders.and_then(|b| b.left)).or(tborders.left);
indent + border.filter(Border::is_visible).map_or(0.0, |b| b.width.clamp(0.0, 12.0) / 2.0)
} else {
indent - first_cell_left_margin_in(t, margins_def)
};
let x = match (t.props.align, rtl) {
(Some(Align::Center), _) => (avail - total) / 2.0,
// End alignment: the right margin, or the left one in a right-to-left table.
(Some(Align::Right), false) => avail - total,
(Some(Align::Right), true) => 0.0,
(_, true) => avail - total - start_offset,
(_, false) => start_offset,
};
let nrows = t.rows.len();
// First pass: lay out every cell's content.
struct CellBox {
items: Vec<Placed>,
h: f32,
x: f32,
w: f32,
valign: VAlign,
fill: Option<Rgb>,
borders: Borders,
vmerge: VMerge,
margins: [f32; 4],
g: usize,
span: usize,
/// Turned text: `items` are still in the turned frame (placed in the second pass).
turned: Option<Turned>,
/// Where the content ends below the cell's top (before vertical alignment).
end: f32,
}
let mut rows: Vec<Vec<CellBox>> = Vec::with_capacity(nrows);
let mut heights: Vec<f32> = Vec::with_capacity(nrows);
let mut bottoms: Vec<f32> = Vec::with_capacity(nrows);
let mut edges: Vec<f32> = Vec::with_capacity(nrows);
for (ri, row) in t.rows.iter().enumerate() {
let mut g = row.first_grid();
let mut cells = Vec::with_capacity(row.cells.len());
let mut rh = 0.0f32;
// The tallest cell margins plus border bands in the row.
let mut row_insets = 0.0f32;
let mut row_below = 0.0f32;
let mut row_edge = 0.0f32;
// Every cell an empty one with a hidden end-of-cell mark: the row is only its borders.
let mut collapsed = true;
// Word gives every cell in a row the row's largest top and bottom margins: a cell without
// its own lines up with a neighbour that has 5.65pt ones.
let (row_top, row_bottom) = row
.cells
.iter()
.map(|c| c.props.resolved_margins(margins_def))
.fold((0.0f32, 0.0f32), |(top, bottom), m| (top.max(m[0]), bottom.max(m[2])));
for (ci, cell) in row.cells.iter().enumerate() {
let span = cell.span();
let x0 = colx.get(g).copied().unwrap_or(acc);
let x1 = colx.get((g + span).min(ncols)).copied().unwrap_or(acc);
// Right to left: grid columns run from the right edge, and the start (left) margin
// is on the cell's right.
let (x0, x1) = if rtl { ((acc - x1).max(0.0), (acc - x0).max(0.0)) } else { (x0, x1) };
let m = cell.props.resolved_margins(margins_def);
let margins = if rtl { [row_top, m[3], row_bottom, m[1]] } else { [row_top, m[1], row_bottom, m[3]] };
let cw = (x1 - x0 - margins[1] - margins[3]).max(4.0);
let region = region_of(t, style.as_ref(), ri, g, span, ncols);
// Conditional formatting, later parts winning (ECMA-376 §17.7.6): whole table, column
// bands, row bands, first/last column, header/total row, corner cells.
if let Some(st) = &style
&& !cell_text.iter().any(|(r, _)| *r == region)
{
cell_text.push((region, region_text(st, region)));
looks.push((region, region_look(st, region)));
}
let look = looks.iter().find(|(r, _)| *r == region).map(|(_, l)| l);
let fill = cell.props.shading.or(look.and_then(|l| l.fill));
let text = cell_text.iter().find(|(r, _)| *r == region).map(|(_, c)| c);
let mut cpath = path.to_vec();
cpath.push(ri as u32);
cpath.push(ci as u32);
// Effective borders: cell > table (outer vs inside).
let tb = tborders;
let edge =
|own: Option<Border>, outer: bool, outer_b: Option<Border>, inner_b: Option<Border>| own.or(if outer { outer_b } else { inner_b });
let cb = cell.props.borders.unwrap_or_default();
// A row's table property exceptions (`w:tblPrEx`) override the table's borders for its cells.
let row_borders = row.props.exceptions.as_ref().and_then(|e| e.borders).unwrap_or_default();
// A row that skips grid columns (`w:gridBefore`/`w:gridAfter`) has its edge at its own
// first and last cell.
let (first_g, last_g) = (g == row.first_grid(), g + span >= ncols.saturating_sub(row.props.grid_after as usize));
let rb = look.map(|l| region_borders(&l.borders, (ri, nrows), (first_g, last_g))).unwrap_or_default();
let mut borders = Borders {
top: edge(cb.top.or(if ri == 0 { row_borders.top } else { row_borders.between }).or(rb.top), ri == 0, tb.top, tb.between),
bottom: edge(
cb.bottom.or(if ri + 1 == nrows { row_borders.bottom } else { row_borders.between }).or(rb.bottom),
ri + 1 == nrows,
tb.bottom,
tb.between,
),
left: edge(
cb.left.or(if first_g { row_borders.left } else { row_borders.inside_v }).or(if first_g { rb.left } else { rb.inside_v }),
first_g,
tb.left,
tb.inside_v,
),
right: edge(
cb.right.or(if last_g { row_borders.right } else { row_borders.inside_v }).or(if last_g { rb.right } else { rb.inside_v }),
last_g,
tb.right,
tb.inside_v,
),
between: None,
inside_v: None,
};
// The logical start (left) and end (right) borders, drawn mirrored in an RTL table.
if rtl {
std::mem::swap(&mut borders.left, &mut borders.right);
}
// Word keeps the cell's text clear of its top border (and the last row's of its bottom
// border): the border's width sits on top of the cell margin.
let band = |b: Option<Border>| b.filter(Border::is_visible).map_or(0.0, |b| b.width.clamp(0.0, 12.0));
let (band_t, band_b) = (band(borders.top), if ri + 1 == nrows { band(borders.bottom) } else { 0.0 });
// A shared edge painted by the row above still clears this cell's text and minimum.
let band_t = if band_t > 0.0 || cell.props.vmerge == VMerge::Continue {
band_t
} else {
rows.last()
.into_iter()
.flatten()
.filter(|above| above.g < g + span && g < above.g + above.span)
.filter(|above| {
above.vmerge == VMerge::None
|| !t.cell_at_grid(ri, above.g).is_some_and(|ci| {
t.grid_col(ri, ci) == above.g && row.cells.get(ci).is_some_and(|c| c.props.vmerge == VMerge::Continue)
})
})
.map(|above| band(above.borders.bottom))
.fold(0.0, f32::max)
};
let insets = margins[0] + margins[2] + band_t + band_b;
let turn = cell.props.text_direction;
let mut turned = None;
let (items, h) = if cell.props.vmerge == VMerge::Continue {
(Vec::new(), 0.0)
} else if turn.is_turned() {
collapsed = false;
// Turned text: its lines run along the cell's height. An exact row height sets
// their length; otherwise they get the length the text needs unwrapped and the row
// grows to fit (the second pass lays them out again if the row ends up taller).
let snap = (ctx.counters.clone(), ctx.eq_count);
let len = match (row.props.height, row.props.height_rule) {
(Some(h), HeightRule::Exact) if h > 0.0 => (h + row_bottom - insets).clamp(4.0, TURNED_LIMIT),
_ => {
let probe = layout_box(ctx, story, &cell.blocks, &cpath, TURNED_MAX, BoxContext::Cell(text), depth, None).items;
(ctx.counters, ctx.eq_count) = (snap.0.clone(), snap.1);
natural_length(&probe).clamp(4.0, TURNED_MAX)
}
};
let BoxLayout { items, height, floats_bottom, .. } =
layout_box(ctx, story, &cell.blocks, &cpath, len, BoxContext::Cell(text), depth, None);
let across = height.max(floats_bottom);
turned = Some(Turned { turn, len, across, snap, region, ci, top: margins[0] + band_t });
(items, len)
} else {
let BoxLayout { mut items, mut height, floats_bottom, last_top, .. } =
layout_box(ctx, story, &cell.blocks, &cpath, cw, BoxContext::Cell(text), depth, None);
// A hidden end-of-cell mark: an empty last paragraph takes no room (Word for
// Windows: a row of empty such cells is only its borders).
if cell.props.hide_mark && cell.blocks.last().and_then(|b| b.as_para()).is_some_and(|p| p.text.is_empty()) {
height = last_top;
} else {
collapsed = false;
}
for it in &mut items {
it.translate(x0 + margins[1], margins[0] + band_t);
}
// The cell grows to hold its floating tables.
(items, height.max(floats_bottom))
};
let h = h + insets;
let end = h - margins[2] - band_b;
if cell.props.vmerge != VMerge::Restart {
rh = rh.max(h);
}
row_insets = row_insets.max(insets);
row_below = row_below.max(margins[2] + band_b);
if ri + 1 < nrows {
row_edge = row_edge.max(band(borders.bottom));
}
cells.push(CellBox {
items,
h,
x: x0,
w: x1 - x0,
// The style's vertical alignment unless the cell has its own.
// The cell's own alignment, even Top, beats the table style's.
valign: cell.props.valign.or_else(|| look.and_then(|l| l.v_align)).unwrap_or_default(),
fill,
borders,
vmerge: cell.props.vmerge,
margins,
g,
span,
turned,
end,
});
g += span;
}
// Word adds the cells' top and bottom margins and border bands to an at-least row height:
// a 30pt row with 5pt margins and 0.5pt borders is 40.5pt tall even when its text needs less.
// An exact height gains only the bottom margin: it runs from the top border to the bottom
// of the text area (a 35pt row with 4pt margins steps 39pt).
let rh = match (row.props.height, row.props.height_rule) {
(Some(h), HeightRule::Exact) if h > 0.0 => h + row_bottom,
(Some(h), _) if h > 0.0 => rh.max(h + row_insets),
_ => rh,
};
bottoms.push(row_below);
edges.push(row_edge);
heights.push(if collapsed && !cells.is_empty() { rh.max(0.0) } else { rh.max(4.0) });
rows.push(cells);
}
// Vertically merged cells: grow the last row of the merge if the content is taller.
for ri in 0..rows.len() {
let restarts: Vec<(usize, f32, usize)> =
rows.get(ri).map(|r| r.iter().filter(|c| c.vmerge == VMerge::Restart).map(|c| (c.g, c.h, c.span)).collect()).unwrap_or_default();
for (g, h, _) in restarts {
let mut end = ri;
while rows.get(end + 1).is_some_and(|r| r.iter().any(|c| c.g == g && c.vmerge == VMerge::Continue)) {
end += 1;
}
let have: f32 = heights.get(ri..=end).map(|s| s.iter().sum()).unwrap_or(0.0);
if h > have
&& let Some(last) = heights.get_mut(end)
{
*last += h - have;
}
}
}
// Grid columns holding a vertical-merge continuation, per row.
let cont: Vec<Vec<usize>> = t
.rows
.iter()
.map(|r| {
let mut gx = r.first_grid();
r.cells
.iter()
.filter_map(|x| {
let g = gx;
gx += x.span();
(x.props.vmerge == VMerge::Continue).then_some(g)
})
.collect()
})
.collect();
// Second pass: place cells with their row height, fills and borders.
let mut out = Vec::with_capacity(rows.len());
for (ri, cells) in rows.into_iter().enumerate() {
let rh = heights.get(ri).copied().unwrap_or(0.0);
let exact = t.rows.get(ri).is_some_and(|r| r.props.height_rule == HeightRule::Exact && r.props.height.is_some_and(|h| h > 0.0));
let mut items = Vec::new();
let mut parts = Vec::with_capacity(cells.len());
for (ci, c) in cells.into_iter().enumerate() {
let first = items.len();
// A restart cell spans several rows: its height is their sum.
let mut span_h = rh;
if c.vmerge == VMerge::Restart {
let mut k = ri + 1;
while cont.get(k).is_some_and(|v| v.contains(&c.g)) {
span_h += heights.get(k).copied().unwrap_or(0.0);
k += 1;
}
}
let rect = Rect::new(c.x, 0.0, c.w, span_h);
if c.vmerge != VMerge::Continue
&& let Some(f) = c.fill
{
items.push(Placed::Fill { rect, color: f });
}
// Behind-text frames count towards vertical alignment even without inline content.
// They do not increase the row's flow height.
let align_h = c
.items
.iter()
.filter_map(|item| match item {
Placed::Object { rect, wrap: wordcraft_doc::para::Wrap::BehindText, .. } => Some(rect.bottom()),
_ => None,
})
.fold(c.h, f32::max);
let mut dy = match c.valign {
VAlign::Top => 0.0,
VAlign::Center => ((span_h - align_h) / 2.0).max(0.0),
VAlign::Bottom => (span_h - align_h).max(0.0),
};
let end = (c.vmerge != VMerge::Continue && c.turned.is_none()).then_some(c.end + dy);
let mut cell_items = c.items;
if let Some(tn) = c.turned {
dy = 0.0;
let m = c.margins;
// The lines run the cell's whole height (laid out again if the row grew).
let len = (span_h - (c.h - tn.len)).clamp(4.0, TURNED_LIMIT);
let mut across = tn.across;
if (len - tn.len).abs() > 0.01
&& let Some(cell) = t.rows.get(ri).and_then(|r| r.cells.get(tn.ci))
{
let mut cpath = path.to_vec();
cpath.extend([ri as u32, tn.ci as u32]);
let text = cell_text.iter().find(|(r, _)| *r == tn.region).map(|(_, c)| c);
// Same list numbers as the first layout: lay it out from the same counters.
let now = (std::mem::replace(&mut ctx.counters, tn.snap.0), std::mem::replace(&mut ctx.eq_count, tn.snap.1));
let b = layout_box(ctx, story, &cell.blocks, &cpath, len, BoxContext::Cell(text), depth, None);
(cell_items, across) = (b.items, b.height.max(b.floats_bottom));
(ctx.counters, ctx.eq_count) = now;
}
// Lines stack across the cell from its start edge (right for top-to-bottom text,
// left for bottom-to-top); the vertical alignment moves them across it.
let cw = (c.w - m[1] - m[3]).max(0.0);
let off = match c.valign {
VAlign::Top => 0.0,
VAlign::Center => ((cw - across) / 2.0).max(0.0),
VAlign::Bottom => (cw - across).max(0.0),
};
let (ox, oy) = match tn.turn {
TextDirection::Up => (c.x + m[1] + off, tn.top + len),
_ => (c.x + m[1] + cw - off, tn.top),
};
for it in &mut cell_items {
it.turn(tn.turn, ox, oy);
}
}
for mut it in cell_items {
it.translate(0.0, dy);
items.push(it);
}
let b = c.borders;
let (x0, x1, y0, y1) = (c.x, c.x + c.w, 0.0, rh);
// Horizontal edges run on over half of each visible vertical edge, so the corners
// close the way Word draws them instead of leaving a notch at thick borders.
let half = |o: Option<Border>| o.filter(Border::is_visible).map_or(0.0, |e| e.width / 2.0);
let (hx0, hx1) = (x0 - half(b.left), x1 + half(b.right));
let top_vis = c.vmerge != VMerge::Continue;
if top_vis && let Some(e) = b.top.filter(Border::is_visible) {
items.push(Placed::Rule { x0: hx0, y0, x1: hx1, y1: y0, border: e });
}
let last_of_merge = c.vmerge == VMerge::None || !cont.get(ri + 1).is_some_and(|v| v.contains(&c.g));
if last_of_merge && let Some(e) = b.bottom.filter(Border::is_visible) {
items.push(Placed::Rule { x0: hx0, y0: y1, x1: hx1, y1, border: e });
}
if let Some(e) = b.left.filter(Border::is_visible) {
items.push(Placed::Rule { x0, y0, x1: x0, y1, border: e });
}
if let Some(e) = b.right.filter(Border::is_visible) {
items.push(Placed::Rule { x0: x1, y0, x1, y1, border: e });
}
items.push(Placed::Cell {
rect: Rect::new(c.x, 0.0, c.w, if c.vmerge == VMerge::Continue { rh } else { span_h }),
covered: c.vmerge == VMerge::Continue,
table: wordcraft_doc::Path(path.to_vec()),
row: ri,
cell: ci,
story,
clip: exact && c.h > span_h + 0.01,
});
let top = c.margins[0] + c.borders.top.filter(Border::is_visible).map_or(0.0, |b| b.width.clamp(0.0, 12.0));
parts.push(CellPart { top, items: first..items.len(), end });
}
out.push(RowLayout {
height: rh,
items,
bottom: bottoms.get(ri).copied().unwrap_or(0.0),
edge: edges.get(ri).copied().unwrap_or(0.0),
cells: parts,
});
}
TableLayout { x, width: total, rows: out }
}
/// Split a row at `cut` (points below the row's top) so the part above fits on this page.
/// Paragraph lines break between lines; fills, vertical rules and cell areas are cut; the
/// rest moves up to the top of the continuation. `None` when no line fits above the cut.
///
/// Like Word, the part left on the page keeps the cells' bottom margin and border below its
/// text and a cell's last paragraph its space after (a row only that much taller than the room
/// left splits rather than moving on whole), and widow control keeps two lines of a paragraph on
/// either side of the break. Each cell's text goes on at the top of the continuation, and the
/// continuation is as tall as the cell needing the most room, and at least the rest of the row
/// (a minimum height holds over the pages).
pub fn split_row(row: &RowLayout, cut: f32) -> Option<(RowLayout, RowLayout)> {
// Also rejects a NaN cut.
if !(cut > 0.0 && cut < row.height) {
return None;
}
let limit = cut - row.bottom.clamp(0.0, cut);
let fits = |bottom: f32| bottom <= limit + 0.01;
// Each cell's items: every item of a row is in one of its cells.
let cells: Vec<&[Placed]> = row.cells.iter().map(|c| row.items.get(c.items.clone()).unwrap_or(&[])).collect();
// A nested table's cell is deeper than the row's own.
let own_depth = row.items.iter().filter_map(|it| if let Placed::Cell { table, .. } = it { Some(table.0.len()) } else { None }).min();
// The first line of each paragraph that moves on (`l1` when all of them stay), by cell and
// item.
let splits: Vec<Vec<usize>> = cells
.iter()
.zip(&row.cells)
.map(|(items, c)| {
// Where the cell's text ends: the paragraph ending lowest is the cell's last.
let lowest =
items.iter().filter_map(|it| text_lines(it).and_then(|(p, l0, l1, y)| crate::item_bottom(y, p, l0, l1))).fold(f32::MIN, f32::max);
items
.iter()
.map(|it| {
let Some((para, l0, l1, y)) = text_lines(it) else { return 0 };
// The cell's last paragraph ends with the cell's content (its space after).
let last = l1 == para.lines.len() && crate::item_bottom(y, para, l0, l1).is_some_and(|b| b >= lowest - 0.01);
line_split(para, (l0, l1), y, limit, c.end.filter(|_| last))
})
.collect()
})
.collect();
// A cell that widow/orphan control leaves with nothing on this page, though its first line
// would fit there, takes the whole row on with it: Word doesn't leave the other cells behind
// beside an empty one (observed with synthetic documents). Trailing space alone can move
// the last paragraph without deferring the whole row.
for (items, splits) in cells.iter().zip(&splits) {
let first = items.iter().zip(splits).find_map(|(it, &split)| text_lines(it).map(|t| (t, split)));
if let Some(((para, l0, l1, y), split)) = first
&& split == l0
&& para.rp.widow_control
&& l1.saturating_sub(l0) > 1
&& crate::item_bottom(y, para, l0, l1).is_some_and(|bottom| !fits(bottom))
&& para.lines.get(l0).is_some_and(|l| fits(y + l.top - para.lines.get(l0).map_or(0.0, |f| f.top) + l.height))
{
return None;
}
}
// The top of the first thing that moves on in each cell.
let mut kept_any = false;
let mut moved = vec![f32::MAX; cells.len()];
for ((items, splits), m) in cells.iter().zip(&splits).zip(&mut moved) {
for (it, &split) in items.iter().zip(splits) {
let top = if let Some(r) = it.turned_bounds() {
// Turned text moves as a whole, like a picture.
if fits(r.bottom()) {
kept_any = true;
None
} else {
Some(r.y)
}
} else {
match it {
Placed::Lines { para, l0, l1, y, .. } => {
kept_any |= split > *l0;
let base = para.lines.get(*l0).map(|l| l.top).unwrap_or(0.0);
(split < *l1).then(|| para.lines.get(split).map_or(*y, |l| y + l.top - base))
}
Placed::Image { rect, .. } | Placed::Shape { rect, .. } | Placed::Graphic { rect, .. } | Placed::Clip { rect, .. } => {
if fits(rect.bottom()) {
kept_any = true;
None
} else {
Some(rect.y)
}
}
// A nested table's cell the cut runs through goes on from the cut, so what is
// below the cut keeps its size and the text after it stays below it (a
// floating table's fixed-height row in an otherwise fitting cell splits the
// row there). A row that is just taller than its text (a minimum height)
// moves on whole, as in Word.
Placed::Cell { rect, table, .. } if Some(table.0.len()) > own_depth && rect.y < cut && rect.bottom() > cut + 0.01 => Some(cut),
_ => None,
}
};
if let Some(top) = top {
*m = m.min(top);
}
}
}
let first_moved = moved.iter().copied().fold(f32::MAX, f32::min);
if !kept_any || first_moved == f32::MAX {
return None;
}
// Each cell resumes below its own top margin and border band.
let shifts: Vec<f32> = moved.iter().zip(&row.cells).map(|(&m, c)| ((if m < f32::MAX { m } else { first_moved }) - c.top).max(0.0)).collect();
// The continuation is as tall as the cell needing the most room (what it moves on and the
// bottom margin, the rest of the row for turned text), and at least the rest of the row.
let full = (row.height - shifts.iter().copied().fold(f32::MAX, f32::min)).max(4.0);
let mut need = row.height - cut;
for ((c, &m), &s) in row.cells.iter().zip(&moved).zip(&shifts) {
if m < f32::MAX {
need = need.max(c.end.map_or(row.height, |e| e + row.bottom) - s);
}
}
let height = need.clamp(4.0, full);
// What runs to the row's bottom (cell areas, shading, side borders) ends at the
// continuation's.
let rest = (row.height - height).max(0.0);
let geo_shift = |bottom: f32, shift: f32| if bottom >= row.height - 0.01 { rest } else { shift };
let (mut a, mut b) = (Vec::new(), Vec::new());
let mut parts = Vec::with_capacity(cells.len());
for (((c, items), splits), &shift) in row.cells.iter().zip(&cells).zip(&splits).zip(&shifts) {
let start = b.len();
for (it, &split) in items.iter().zip(splits) {
if let Some(r) = it.turned_bounds() {
let mut it = it.clone();
if fits(r.bottom()) {
a.push(it);
} else {
it.translate(0.0, -shift);
b.push(it);
}
continue;
}
match it {
Placed::Lines { story, path, para, l0, l1, x, y, turn } => {
let base = para.lines.get(*l0).map(|l| l.top).unwrap_or(0.0);
if split > *l0 {
a.push(Placed::Lines {
story: *story,
path: path.clone(),
para: para.clone(),
l0: *l0,
l1: split,
x: *x,
y: *y,
turn: *turn,
});
}
if split < *l1 {
let top = para.lines.get(split).map(|l| y + l.top - base).unwrap_or(*y);
b.push(Placed::Lines {
story: *story,
path: path.clone(),
para: para.clone(),
l0: split,
l1: *l1,
x: *x,
y: top - shift,
turn: *turn,
});
}
}
Placed::Image { rect, .. }
| Placed::Shape { rect, .. }
| Placed::Graphic { rect, .. }
| Placed::Clip { rect, .. }
| Placed::Object { rect, .. } => {
if fits(rect.bottom()) {
a.push(it.clone());
} else {
let mut it = it.clone();
it.translate(0.0, -shift);
b.push(it);
}
}
// Shading and cell areas: the part above the cut stays, the rest moves up with the
// continuation (a nested table's cell below the cut moves whole).
Placed::Fill { rect, color } => {
let (above, below) = split_rect(*rect, cut, geo_shift(rect.bottom(), shift));
a.extend(above.map(|rect| Placed::Fill { rect, color: *color }));
b.extend(below.map(|rect| Placed::Fill { rect, color: *color }));
}
Placed::Cell { rect, covered, table, row: r, cell, story, clip } => {
let (above, below) = split_rect(*rect, cut, geo_shift(rect.bottom(), shift));
let cell_at =
|rect| Placed::Cell { rect, covered: *covered, table: table.clone(), row: *r, cell: *cell, story: *story, clip: *clip };
a.extend(above.map(cell_at));
b.extend(below.map(cell_at));
}
Placed::Rule { x0, y0, x1, y1, border } => {
if (y0 - y1).abs() < 0.01 {
// Horizontal: the top edge stays, the bottom edge moves; both fragments close.
if *y0 <= 0.01 {
a.push(it.clone());
b.push(Placed::Rule { x0: *x0, y0: 0.0, x1: *x1, y1: 0.0, border: *border });
} else if *y0 >= row.height - 0.01 {
a.push(Placed::Rule { x0: *x0, y0: cut, x1: *x1, y1: cut, border: *border });
b.push(Placed::Rule { x0: *x0, y0: y0 - rest, x1: *x1, y1: y1 - rest, border: *border });
} else if *y0 <= cut {
a.push(it.clone());
} else {
b.push(Placed::Rule { x0: *x0, y0: y0 - shift, x1: *x1, y1: y1 - shift, border: *border });
}
} else {
let (t, bt) = (y0.min(*y1), y0.max(*y1));
if t < cut {
a.push(Placed::Rule { x0: *x0, y0: t, x1: *x1, y1: bt.min(cut), border: *border });
}
if bt > cut {
let shift = geo_shift(bt, shift);
b.push(Placed::Rule { x0: *x0, y0: (t - shift).max(0.0), x1: *x1, y1: bt - shift, border: *border });
}
}
}
}
}
parts.push(CellPart { top: c.top, items: start..b.len(), end: c.end.map(|e| e - shift) });
}
// The part left on the page is placed and never split again.
Some((
RowLayout { height: cut, items: a, bottom: row.bottom, edge: 0.0, cells: Vec::new() },
RowLayout { height, items: b, bottom: row.bottom, edge: row.edge, cells: parts },
))
}
/// A cell paragraph's placed lines (not turned text): the paragraph, its lines `l0..l1` and the
/// top of line `l0`.
fn text_lines(it: &Placed) -> Option<(&ParaLayout, usize, usize, f32)> {
match it {
Placed::Lines { para, l0, l1, y, .. } if it.turned_bounds().is_none() => Some((para, *l0, *l1, *y)),
_ => None,
}
}
/// The first of lines `l0..l1` of `para` (placed with line `l0`'s top at `y`) that does not fit
/// above `limit`: the last line of a cell's last paragraph needs room down to the cell's content
/// `end`, its space after. Widow control keeps two lines on either side of the break, or moves
/// the paragraph on whole.
fn line_split(para: &ParaLayout, (l0, l1): (usize, usize), y: f32, limit: f32, end: Option<f32>) -> usize {
let n = para.lines.len();
let mut k = l0;
while k < l1 {
let Some(bottom) = crate::item_bottom(y, para, l0, k + 1) else { break };
let bottom = if k + 1 == n { end.map_or(bottom, |e| e.max(bottom)) } else { bottom };
if bottom > limit + 0.01 {
break;
}
k += 1;
}
// The body's rule differs: there a widow takes a line along only when three or more lines are
// on the page (`place_para_in_flow`).
if para.rp.widow_control && k > l0 && k < l1 {
// A widow: take one more line along.
if n - k == 1 {
k -= 1;
}
// An orphan (the first line alone): the paragraph moves on whole.
if l0 == 0 && k == 1 {
k = 0;
}
}
k
}
/// The parts of `r` above `cut` and below it, the latter moved up by `shift` (onto the
/// continuation of a row split at `cut`).
fn split_rect(r: Rect, cut: f32, shift: f32) -> (Option<Rect>, Option<Rect>) {
let above = (r.y < cut).then(|| Rect::new(r.x, r.y, r.w, (r.bottom().min(cut) - r.y).max(0.0)));
let top = (r.y - shift).max(0.0);
let below = (r.bottom() > cut).then(|| Rect::new(r.x, top, r.w, (r.bottom() - shift - top).max(0.0)));
(above, below)
}
/// A cell's table style parts.
type Region = PartSet;
/// The parts of the cell at grid column `g` spanning `span` of `ncols` columns in row `ri`.
fn region_of(t: &Table, style: Option<&TableStyleProps>, ri: usize, g: usize, span: usize, ncols: usize) -> Region {
let rows = style.and_then(|s| s.parts.band_size).unwrap_or(0);
let cols = style.and_then(|s| s.parts.col_band_size).unwrap_or(0);
PartSet::of_cell(&t.props.look, (rows, cols), (ri, t.rows.len()), (g, span, ncols))
}
/// The table style's text formatting in `region`: each part's character formatting and paragraph
/// alignment over the last, lowest precedence first (ECMA-376 §17.7.6).
fn region_text(st: &TableStyleProps, region: Region) -> CellText {
let mut chr = wordcraft_doc::CharProps::default();
let mut para = st.para.clone();
for part in region.iter() {
let f = st.part(part);
chr.overlay(&f.chr);
if f.align.is_some() {
para.align = f.align;
}
}
CellText { para, chr }
}
/// Each column's narrowest and widest content, cell margins included: its longest word and its
/// longest line. A cell spanning columns shares what they lack between them; nested tables count
/// with their own. `None` when every cell is empty.
fn column_content(ctx: &mut Ctx, t: &Table, style: Option<&TableStyleProps>, depth: usize) -> Option<Vec<(f32, f32)>> {
let margins_def = default_margins(t, style);
let mut cols = vec![(0.0f32, 0.0f32); t.cols().max(1)];
let mut spanning = Vec::new();
let mut any = false;
for (ri, row) in t.rows.iter().enumerate() {
let mut g = row.first_grid();
for cell in &row.cells {
let span = cell.span();
if cell.props.vmerge != VMerge::Continue {
let text = style.map(|st| region_text(st, region_of(t, style, ri, g, cell.span(), cols.len())));
let (lo, hi) = blocks_content(ctx, &cell.blocks, text.as_ref(), depth);
any |= hi > 0.0;
let m = cell.props.resolved_margins(margins_def);
let need = (lo + m[1] + m[3], hi + m[1] + m[3]);
match cols.get_mut(g) {
Some(col) if span == 1 => *col = (col.0.max(need.0), col.1.max(need.1)),
_ => spanning.push((g, span, need)),
}
}
g += span;
}
spanning.extend(row_gaps(row, g).map(|(g, span, w)| (g, span, (w, w))));
}
for (g, span, need) in spanning {
let Some(cs) = cols.get_mut(g..(g + span).min(t.cols())) else { continue };
let n = cs.len().max(1) as f32;
let (have_lo, have_hi) = cs.iter().fold((0.0, 0.0), |(l, h), c| (l + c.0, h + c.1));
for c in cs {
c.0 += (need.0 - have_lo).max(0.0) / n;
c.1 += (need.1 - have_hi).max(0.0) / n;
}
}
any.then_some(cols)
}
/// The grid columns row `row` leaves empty before its first cell and after its last one (which
/// ends at grid column `end`) when it gives their width (`w:wBefore`, `w:wAfter`): (first column,
/// columns, width in points). Column widths make room for them like for an empty cell.
fn row_gaps(row: &Row, end: usize) -> impl Iterator<Item = (usize, usize, f32)> {
let p = &row.props;
[(0, row.first_grid(), p.w_before), (end, p.grid_after as usize, p.w_after)]
.into_iter()
.filter(|&(_, n, _)| n > 0)
.filter_map(|(g, n, w)| w.filter(|w| w.is_finite() && *w > 0.0).map(|w| (g, n, w.min(1584.0))))
}
/// The narrowest and widest a block list can be laid out.
fn blocks_content(ctx: &mut Ctx, blocks: &Blocks, text: Option<&CellText>, depth: usize) -> (f32, f32) {
let (mut lo, mut hi) = (0.0f32, 0.0f32);
for b in blocks.iter() {
let (l, h) = match &**b {
Block::Para(p) => {
let label = list_level(ctx, p).map(|l| (String::new(), l));
ctx.para_measured(p, UNBOUNDED, text, label, para_content)
}
// A nested table: its fixed width, else its columns' but at least its preferred width
// (its own, else its grid's when it has preferred widths at all).
Block::Table(nt) if depth < 8 => {
let grid = if is_automatic(nt) { 0.0 } else { nt.grid.iter().filter(|w| w.is_finite()).map(|w| w.max(0.0)).sum() };
let width = nt.props.width.filter(|w| w.is_finite()).unwrap_or(grid).clamp(0.0, UNBOUNDED);
if nt.props.fixed && width > 0.0 {
(width, width)
} else {
let style = table_style(ctx, nt);
let cols = column_content(ctx, nt, style.as_ref(), depth + 1).unwrap_or_default();
let (l, h) = cols.iter().fold((0.0, 0.0), |(l, h), c| (l + c.0, h + c.1));
(width.max(l), width.max(h))
}
}
Block::Table(_) => (0.0, 0.0),
};
lo = lo.max(l);
hi = hi.max(h);
}
(lo, hi)
}
/// The list level `p` is in, for its indents (without counting it: measuring isn't laying out).
fn list_level(ctx: &Ctx, p: &Paragraph) -> Option<wordcraft_doc::numbering::Level> {
let n = p.props.numbering.or_else(|| ctx.doc.styles.resolve_para(&p.props).numbering).filter(|n| n.num != 0)?;
ctx.doc.numbering.level(n.num, n.level).cloned()
}
/// A width to lay a paragraph out in so that only its own line breaks end lines.
const UNBOUNDED: f32 = 100_000.0;
/// A paragraph's longest word and longest line (between line breaks), indents included.
fn para_content(pl: &ParaLayout) -> (f32, f32) {
let indents = pl.rp.indent_left.max(0.0) + pl.rp.indent_right.max(0.0) + pl.rp.indent_first.max(0.0);
let (mut word, mut line, mut spaces) = (0.0f32, 0.0f32, 0.0f32);
let (mut lo, mut hi) = (0.0f32, 0.0f32);
for c in &pl.clusters {
let adv = if c.adv.is_finite() { c.adv.clamp(0.0, UNBOUNDED) } else { 0.0 };
match c.kind {
ClKind::Space => {
lo = lo.max(word);
word = 0.0;
spaces += adv;
}
ClKind::LineBreak | ClKind::PageBreak | ClKind::ColumnBreak => {
lo = lo.max(word);
hi = hi.max(line);
(word, line, spaces) = (0.0, 0.0, 0.0);
}
_ => {
word += adv;
line += spaces + adv;
spaces = 0.0;
if c.break_after {
lo = lo.max(word);
word = 0.0;
}
}
}
}
(lo.max(word) + indents, hi.max(line) + indents)
}
/// Whether nothing in `t` asks for a width (its own or any cell's): its columns then follow
/// their content alone.
fn is_automatic(t: &Table) -> bool {
t.props.width.is_none()
&& t.props.width_pct.is_none()
&& t.rows.iter().all(|r| r.cells.iter().all(|c| c.props.width.is_none() && c.props.width_pct.is_none()))
}
/// Consistent cell widths replace a stale grid without changing the table's total width.
/// Conflicting widths, spans and percentages need the full shared-grid algorithm.
fn apply_cell_widths(t: &Table, grid: &mut [f32]) {
let Some(first) = t.rows.first() else { return };
if first.cells.len() != grid.len() {
return;
}
let widths: Option<Vec<f32>> = first.cells.iter().map(|c| c.props.width.filter(|w| w.is_finite() && *w > 0.0)).collect();
let Some(widths) = widths else { return };
let total: f32 = widths.iter().sum();
if !total.is_finite() || (total - grid.iter().sum::<f32>()).abs() > 0.01 {
return;
}
if t.rows.iter().any(|r| {
r.cells.len() != grid.len()
|| r.cells.iter().zip(&widths).any(|(c, w)| c.span() != 1 || c.props.width_pct.is_some() || c.props.width != Some(*w))
}) {
return;
}
grid.copy_from_slice(&widths);
}
/// Word's autofit for a table whose widths may follow its content (`tblLayout` not fixed).
/// - With no preferred width anywhere (the table's and every cell's automatic) the columns come
/// from their content alone: each its widest line if they all fit in `avail`, else shared out
/// between narrowest and widest in proportion to how much each can give.
/// - Otherwise the grid stands, but no column is narrower than its longest word: the columns with
/// room to spare give up the difference.
fn autofit(grid: &mut [f32], content: &[(f32, f32)], t: &Table, avail: f32) {
if grid.len() != content.len() || !avail.is_finite() {
return;
}
if is_automatic(t) {
let (lo, hi) = content.iter().fold((0.0f32, 0.0f32), |(l, h), c| (l + c.0, h + c.1.max(c.0)));
let k = if hi <= avail {
1.0
} else if lo >= avail || hi - lo < 0.01 {
0.0
} else {
(avail - lo) / (hi - lo)
};
for (w, c) in grid.iter_mut().zip(content) {
*w = (c.0 + (c.1.max(c.0) - c.0) * k).max(4.0);
}
return;
}
let short: f32 = grid.iter().zip(content).map(|(w, c)| (c.0 - *w).max(0.0)).sum();
if short < 0.01 {
return;
}
// Donors give equal widths until they reach their own content minimum.
let mut rooms: Vec<_> = grid.iter().zip(content).map(|(w, c)| *w - c.0).filter(|room| *room > 0.0).collect();
rooms.sort_by(f32::total_cmp);
let mut remaining = short;
let mut share = 0.0;
for (i, room) in rooms.iter().enumerate() {
share = remaining / (rooms.len() - i) as f32;
if share <= *room {
break;
}
remaining -= room;
}
for (w, c) in grid.iter_mut().zip(content) {
*w = if *w < c.0 { c.0 } else { *w - (*w - c.0).min(share) };
}
}
#[cfg(test)]
mod width_tests {
use super::*;
#[test]
fn autofit_shares_donor_loss() {
let table = Table::new(1, 4, 260.0);
let mut grid = [30.0, 120.0, 20.0, 90.0];
let content = [(28.0, 28.0), (40.0, 120.0), (40.0, 40.0), (30.0, 90.0)];
autofit(&mut grid, &content, &table, 260.0);
assert_eq!(grid, [28.0, 111.0, 40.0, 81.0]);
}
#[test]
fn autofit_exhausts_donors() {
let table = Table::new(1, 2, 50.0);
let mut grid = [30.0, 20.0];
autofit(&mut grid, &[(25.0, 25.0), (40.0, 40.0)], &table, 50.0);
assert_eq!(grid, [25.0, 40.0]);
}
#[test]
fn autofit_keeps_fitting_widths() {
let table = Table::new(1, 3, 120.0);
let mut grid = [30.0, 40.0, 50.0];
autofit(&mut grid, &[(10.0, 10.0), (20.0, 20.0), (30.0, 30.0)], &table, 120.0);
assert_eq!(grid, [30.0, 40.0, 50.0]);
}
}