* doc: reuse shared legacy codepage tables with parity tests
* Regenerate codepage dependency ordering in Cargo lockfile
* #778: Cargo.lock: wordcraft-doc depends on craft-codepage after merging main
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* #778: zip_guard delegates to craft-zip's entry-count guard
wordcraft_doc::zip_guard::check and zip64_entry_count_sample now call craft-zip, which also
treats a central-directory size of u32::MAX as a ZIP64 marker.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #778: Cargo.lock: wordcraft-doc depends on craft-zip
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #778: zip_guard: craft-zip test-support as a dev-dependency
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
SonarCloud's coverage shards run tests built with -Cinstrument-coverage
on shared runners. Tests that assert a wall-clock limit or a time
ratio swing there: the megabyte-word test took 5.7-23 s against a
15 s limit, and three linear-time tests read 24-31x against 24x. Each
one that flakes turns main red after merging.
Ignore under cfg(coverage) every test whose purpose is performance
(linear time, bounded, capped, fast). The gate CI still runs them all
uninstrumented. Behaviour tests that only carry a generous time guard
(fuzz, never-panics, clipboard, regex, render, recent fonts) and the
SSIP speech timeout tests keep running, so their checks stay covered.
Claude-Session: https://claude.ai/code/session_01FypgCXtSjB6e5jmy9Uq7u8
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Dependencies: take craft-libs main at 3af2280ee3
Every craft-libs git dependency (the workspace's 25, ui-egui's 3, the app's
craft-print, the web build's 2 and the krilla patch) moves to craft-libs main
at 3af2280ee3 from the 8 revisions they were pinned at. Cargo.lock changes
those sources, moves winprint and craft-winprint-gdi with craft-print, drops
the second craft-tiff, and adds cap-primitives and its dependencies, which
craft-font-store's Language pack recovery uses on native builds.
- craft_print::Job has two new fields. WordCraft sets grey_pages to false,
since it does not turn the spooled pages gray, and driver_settings to None.
- DOCX text boxes, rectangles, rounded rectangles and ellipses are now
written in mc:AlternateContent with a VML fallback that repeats the text
box's text. Two tests that counted w:txbxContent elements, or the story
text, now count the fallback copy as well.
* Printing: the isolated print helper keeps the job's grey pages and driver settings
The Windows print helper process rebuilt each job with grey_pages off and no
driver settings, whatever the job carried. Its request now carries both: the
driver settings travel as the printer's name and the DEVMODE in hex, and
damaged settings are an error. The request file may now be up to 2 MiB larger,
the size of the largest driver settings craft-print accepts, written as hex.
Comments in Cargo.toml and apps/wordcraft/Cargo.toml now describe the shared
craft-libs commit, and the shared-story text box test says where its two copies
are written.
* DOCX: open picture frame shapes and keep chart formatting from craft-libs
At this craft-libs commit the DOCX reader returns a picture frame shape
(`frame` preset) as its own shape kind, and charts carry series formatting
and axis settings. WordCraft's model had neither:
- A document with a picture frame shape failed to open with "unknown
variant `frame`". The model bridge now reads it as a rectangle, as the
earlier reader did.
- The chart model drops nothing the shared chart model holds: series fill,
line width, line and fill off, markers, smooth lines, data labels and
number format, and the category and value axes. Before, saving a chart
with any of these wrote the chart again without them.
- A new scatter series is drawn as markers without a line, as before.
Change Chart Type sets this when a chart becomes a scatter chart or stops
being one, Edit Data keeps each replaced series' formatting, and Data
Labels replaces each series' own label setting.
* Charts: draw turned axis titles and keep stored scatter charts without lines
At this craft-libs commit a chart's vertical axis title is a text item
turned -90°. WordCraft's graphic text item now carries that angle, and the
text is laid out along its box and drawn turned about the box's center.
Scatter series saved in a .json document or a recovery snapshot before
series formatting existed have no `noLine` value. They now read as markers
without a line, as they were drawn when saved.
* Charts: draw opened ODT scatter charts without lines
The ODF reader at this craft-libs commit does not read a scatter chart's
line setting, so a scatter chart saved as ODT and opened again was drawn
with lines between its points. The ODF writer writes every scatter series
as markers without a line, and the opened chart is now drawn that way.
* Charts: set scatter series to markers only for opened ODF files alone
Opened ODT, OTT and FODT files set their scatter series to markers
without a line, since the ODF reader does not read a series' line
setting. Other formats that go through the same import path keep their
series as read. A series that arrives with its markers turned off gets
them back, as Change Chart Type does, so it is still drawn.
The test for an opened scatter chart now also covers flat ODT, checks
the drawn items against the markers-only drawing, and checks that the
chart's paragraph is still written as read.
* Dependencies: take craft-libs main at 453b0da782
Every craft-libs dependency, and the craft-proof note in Cargo.toml,
moves from 3af2280ee3 to 453b0da782. Cargo.lock adds craft-dds, which
craft-image-codecs now uses; no crates.io version changes.
- craft-xls builds its test fixtures only for tests and its
test-support feature, and they return a Result. wordcraft-formats
gets a test-support feature that turns it on; sheet::xls_fixture is
behind it and returns Result, and wordcraft-engine's tests enable it.
- craft_metafile::Picture has new report fields; the render tests'
literals fill them with Default.
- The docx tests follow the writer: the font table is kept as read,
the flat OPC comparison allows a kept part without its XML
declaration, and the bibliography store is related from the main
part, once, after a second save too.
* DOCX: gutter position Top is kept when a document is opened and saved
craft-libs now reads w:gutterAtTop into its settings model instead of
keeping it as unread XML, and WordCraft's settings had no field for it,
so opening and saving a document dropped the setting. Settings gets
gutter_at_top (off by default), which the document bridge carries both
ways. Layout still puts the gutter on the left.
* Charts: Change Chart Type gives a pie chart's value axis the new type's gridlines
A pie or doughnut chart keeps a plain value axis (no gridlines, axis
line on), and a new chart of any other type has gridlines and no axis
line. Change Chart Type kept the old axis, so a pie chart changed to a
column chart was drawn and saved without gridlines and with an axis
line. Change Chart Type between a round chart and another type now sets
the value axis to the new type's when it is still the old type's; an
axis that was changed stays as it is.
The progress log gets a line for the chart formatting, turned axis
titles, picture frame shapes and gutter position this branch brings.
The gutter_at_top doc comment states only what the field holds.
* #3518: Review › Language can set the whole document's proofing language
review.language takes scope "selection" (default, unchanged) or "document":
every run and paragraph mark of every story (body, headers, footers, notes,
comments, text boxes, table cells, glossary) takes lang/noProof in one undo
step, selection unchanged, pending typing format follows. The Language dialog
gets an "Apply to the whole document" check box.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FypgCVCCEo7u5wetHaJRU4
* #3518: whole-document OK sends only what was changed in the Language dialog
With "Apply to the whole document" ticked, OK sends noProof only if the user
changed it from the dialog's opening value, and lang if changed or if neither
changed (never the engine's en-US default). Heading reads "Mark the whole
document as:" while ticked. Glossary walker uses doc's PARA_WALK_DEPTH (now pub).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FypgCVCCEo7u5wetHaJRU4
* #3518: whole-document OK with Do not check ticked marks the whole document
CodeRabbit: with the caret's text already not checked, the box was unchanged, so a
whole-document OK never sent noProof and the rest of the document stayed checked.
A ticked box is now always sent; an unticked, unchanged one is still omitted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #3518: refuse whole-document language changes under protection
Protection leaves only its exceptions editable, so document scope is refused before anything changes. The post-check now compares the glossary, which no story-level check reached.
* #3518: whole-document language keeps per-run proofing, returns selection, skips glossary while tracking
- Language dialog: whole-document OK sends noProof only when the box was changed, so a language-only change no longer marks everything no-proof.
- review.language scope "document" returns the caret/selection result (plus scope).
- Whole-document change skips glossary blocks while Track Changes and Track Formatting are on (tracked formatting does not cover them).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #3518: whole-document Do not check is sent when the box is clicked; report skipped glossary
- Language dialog: a whole-document OK sends noProof iff the Do-not-check box was clicked (proof_touched), so ticking it with the caret in a no-proof run still marks the whole document, while a language-only change leaves other runs alone.
- review.language document scope reports "glossarySkipped": true when the glossary was skipped under Track Changes + Track Formatting; doc comment reworded to state only the limitation.
- Docs status line and revision row bumped to 2026-11-16.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Martin Ball <martin.ball@ucsmedical.co.uk>
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
- Recognize a marked Fit One Page custom XML store across package and document
relationships, skipping passthrough only after loading successfully.
- Reserve the Fit One Page part path in EmbedWriter taken paths to prevent collisions.
- Track typography provenance per formatting location rather than sorted multisets.
- Validate stepPt parameter and reject non-positive or non-finite values.
- Report "already fits on one page" before an exhausted cumulative budget.
- Record original and reduced font sizes, line spacings and paragraph
spacings as multisets, so restoring neighbouring sizes clears the
exhausted budget, and paragraph spacing keeps the budget when line
spacing rounds back to its original.
- Resolve font and spacing budgets independently, letting the remaining
component be used instead of refusing the whole command.
- Persist FitState in a namespaced DOCX custom XML part and restore it on
read; add a DOCX round-trip regression.
- Anchor cumulative budget provenance to the original sizes and spacings
recorded before the first fit, instead of matching every current value
against every recorded value. Restoring body and heading sizes now resets
the budget.
- Build the initial and trial layouts from the session's hide_marked,
deletions_in_balloons and resolved numeral settings, so hidden or
ballooned deletions and numeral shaping paginate as in Session::layout.
- Add regressions for mixed body/heading sizes, mixed line spacings, and
Session::layout parity under non-default visibility and numeral settings.
- Round paragraph spacing up to hundredths, so a very small value is never
reduced by more than the requested factor.
- Ignore page-break characters in hidden or tracked-deleted runs, matching
Print layout visibility, so a hidden break cannot force a refusal.
- Scale paragraph spacing without rounding a small value to zero or
compressing past the requested factor.
- Detect resolved page_break_before and nested table paragraphs as explicit
breaks, and refuse before evaluating candidates.
- Count non-text body items in the final-page volume limit.
- Give the ribbon button an accessible label and test it through the
rendered Layout ribbon in both themes.
Add layout.fitOnePage, a speculative typography search that removes one
page from the document while preserving relative font hierarchy.
Every size receives the same half-point-rounded reduction; line spacing
never falls below 0.94 of its original and body text never below 8.5 pt.
The command refuses on explicit page breaks, excessive final-page content,
formatting protection, tracked formatting and the cumulative 1 pt budget,
and leaves the document unchanged when it refuses.
Add the Layout ribbon button, KeyTip and a code-drawn icon, cover the
command with 41 tests, and regenerate the parity checklist.
* ODT gap 19: frames keep the page they are anchored to
A frame or shape anchored to a page was written without its
text:anchor-page-number, so one on page 3 came back on page 1 in LibreOffice
after a save. Keep the number and write it back. WordCraft's own layout still
places such a frame on its anchor paragraph's page.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT: a page number only counts on a page anchor, and groups keep theirs (gap 19)
A paragraph anchor with a text:anchor-page-number was saved as a page anchor,
which LibreOffice ignores the number on and so moved the drawing. A group
anchored to a page lost its page. Found by CodeRabbit on this PR.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT: a frame moved in WordCraft loses its page number (gap 19)
Nothing cleared it, so the save put the frame back on its old page. One placed by
its paragraph is no page anchor any more, and one dropped on a page takes that page.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT tests: put the moved frame's test with the other moves (gap 19)
At the end of the file it collided with other ODT PRs' tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT: a frame moved onto its paragraph loses its page number (gap 19)
The number was only hidden while the frame was placed by its paragraph, so placing it by the page again (arrange.position) brought back the old page. The merge of main also lost the test attribute on a test from main, put back here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Gap 19: docs status lines are main's, with this change in front
Merging main had kept this branch's older status lines in ROADMAP.md and
docs/file-format-parity.md, which set main's 2026-11-08 date back and dropped
main's newer entries and the parity doc's CLI path note.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Gap 19: parity row says when a moved frame keeps its page number
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
Word for Mac sets raised and lowered text at 0.65 of the run's size
rounded to the half point (measured from digit advances in its PDF
export at base sizes 8 to 36 pt: 11 pt -> 7, 12 -> 8, 24 -> 15.5).
WordCraft used 0.65 exactly, so lines with superscripts were up to
0.5 pt wider and some broke a word early.
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>
* #5843: resolve LibreOffice semicolon-separated font fallback lists to first installed family
* #5843: address review comments on font fallback list resolution
* #5843: verify fallback list text on cache hit and accept Times New Roman in tests
* #5843: bound the fallback lists read from a DOCX and reuse their classification
A document could name thousands of fallback families in `w:ascii`/`w:hAnsi`.
Nothing bounded that value, and the name-based rules (roman classification,
Thai family, named substitutes) then scanned every candidate for each shaped
piece, so opening the document took seconds.
`w:ascii` and `w:hAnsi` now keep the leading families within MAX_FONT_NAME
bytes, cut at a `;` so no name is split, as `w:cs`/`w:eastAsia` already bound
their names. The family a real list would pick is kept: a LibreOffice list
names a few families, well under the bound.
The Thai-family and named-substitute answers are worked out once per list and
kept in the fallback-list cache, keyed by the list text, so shaping no longer
rescans candidates for every piece. Resolution still reads every candidate.
A 500-paragraph document whose font names 20,000 candidates laid out in 10.8 s;
bounded on read it takes 17 ms.
Tests:
- a hostile fallback list keeps its leading families, cut on a `;`
- a real LibreOffice list is read whole
- a later Thai or named fallback still classifies, and lists are cached
A file that turns an inline picture but writes no wp:effectExtent (or a zero
one), as other tools do, makes Word keep the picture's line to its frame and
let the turn overhang it, cut at the line's top and bottom: the frame as it
is in Word 2007 layout (compatibilityMode < 15), with width and height
exchanged from 45 to 135 and 225 to 315 degrees in Word 2013+ layout.
WordCraft always reserved the exact rotated bounds (Float::spin_pad), which
is what Word's own files encode through their effect extents, so the text
after such a picture sat up to 34 pt too low (docxide-pdf
standalone-picture-compat12-* / compat15-angle45/135).
The docx reader now marks such objects (Float::bare_turn, cleared when the
object is turned again); the inline room comes from Float::line_pad, which
display uses too to place the frame in it, and the line clips the picture's
overhang. The writer keeps writing no turn overhang for them, so a save
round-trips. Word-saved pictures are unchanged.
Co-authored-by: Naxvog <4963183+naxvog@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* WC6-07: preserve paragraph text direction
* Resolve gaps summary merge
* WC6-07: retain paragraph direction variants
* WC6-07: preserve strict paragraph direction
* Update lockfile for DOCX test dependency
* Close paragraph direction implementation
* Format paragraph direction imports
* #4985: a numFmt none list level shows its literal text
Word labels a paragraph whose level has numFmt none and lvlText RESPONSE:
with RESPONSE:. Counters::next_label returned no label for any none level;
now only a none level of placeholders shows nothing. Its own %n stays
empty, and it still doesn't count. Numbering::shows_label agrees.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: rustfmt
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: a "none" list level counts, and only its own number is empty
Taken from MeHail-02's numbering-literal-none branch (46405e66), checked in Word
for Windows: a none level is an ordinary level whose own number reads as
nothing. It counts and restarts the levels below it, stays empty in legal
numbering, and an empty label still tabs the text to the indent unless it
already ends there (the Chapter 1 list's Heading 2 stays at the margin).
Co-Authored-By: MeHail <mihailsmetanin15@gmail.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: STYLEREF in full context keeps a none-level heading's parent
Review: a none level now has a label, so its headings reach the style index,
which leaves the none level out of the levels it keeps; \w and \r took the last
kept level for the heading's own and dropped the parent. Levels now carry their
index, and the levels above are picked by it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: STYLEREF number switches on a none-level heading
Review: in full context an empty label of its own added a trailing period (3.),
and the suppress switch gave a none level's label (Note), which holds no number,
back whole. An empty own part is left out, and a none level's label with no
number is suppressed to nothing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: STYLEREF \s and \t on none-level labels
Review: \s took a none level's label (Note) for a chapter number; the heading's
count stands in, as for an unnumbered heading. \t matched any ancestor's number
inside a word (the a of Chapter); only the numbers the label's template shows
count now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: STYLEREF takes a label's numbers by their spans
Review: \s took the last word of a label (Response of "3 Response") and fell
back on the heading count for a none level showing its parent's number (%1);
\t kept text attached to a number (Section3). A label's numbers are now the
spans of the values its template shows, with their delimiters (a letter number
only standing apart from other letters); \s takes the last of them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* #4985: tests edit the numbering through Arc::make_mut after merging main
Main shares the numbering between undo steps (#2694), so `Document::numbering` is an
`Arc<Numbering>`; the tests this PR adds wrote to it directly and no longer built.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: MeHail <mihailsmetanin15@gmail.com>
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>
* fix(fields): quoted-argument backslash escapes and body PAGE in a cardinal-text section
A backslash inside a quoted field argument makes the next character literal, so TITLE "x \* Upper" shows x * Upper as Word does. A body PAGE field now uses the section's page number format (cardinal text shows words), as headers and footers already did.
* Fields: PAGE in a body table follows the section number format; parity docs
* fmt
* docs: clarify quoted field escape rule
* Fix page fields in fitted text boxes
* Move fitted cache test after layout helpers
---------
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>
A list level, page number or note mark in Word's thaiNumbers format was
read and saved as such but drawn as decimal (1, 2, 3). It now shows the
decimal number in Thai digits (U+0E50..U+0E59): ๑, ๒ … ๑๐, with the
level's text around it (๑., (๒)). Thai letters and Thai counting still
show decimal until they are drawn too.
Part of #5757 (Thai numerals in automatic numbering).
* ODT gap 19: a watermark change to a file is saved
A file's own page definitions were written back whatever happened to its
watermark, so one added, changed or removed in WordCraft was gone from the
saved file. The watermark now counts as part of the page setup, so a change
writes the page definitions from the model. A watermark with no text isn't
written, so it isn't a change.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT gap 19: Different Odd & Even Pages changed in a file is saved
A file's own page definitions were written back whatever happened to this
setting, so the left pages' header stayed after it was turned off, and turning
it on did nothing. The setting now counts as part of the page setup, as the
watermark does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ODT tests: say what the watermark and odd/even tests check (gap 19)
Doc comments for CodeRabbit's docstring coverage check on this PR.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>
The #3642 guard refused a file whose DIFAT lists more FAT sectors than
its sectors need: one FAT sector per sector_len / 4 sectors plus one
spare. Real writers reserve FAT for growth, so a real encrypted
package — a compound file of 78 sectors — lists 4 FAT sectors and was
refused: encrypted .docx files stopped opening with "the DIFAT lists
4 FAT sectors in a file of 78 sectors" (#5109, broken since the guard
landed).
Every FAT sector listed is a sector of the file, and none of them
twice (checked next), so a real file lists no more FAT sectors than
the file has. The refusal bound is now that: more than num_sectors is
either duplicates or names cfb refuses anyway, and the FAT cfb builds
(one sector of entries per listed name) stays within about the file's
own size — the #3642 blowup listed 2048 × 127 names in a ~1 MB file
and is still refused (260,096 > 2,048).
Co-authored-by: Bra1nZ <bra1nz@astraoffice.com>
* Table formulas: read the Panamanian balboa B/. in numbers
Word's table formula and sort number reader ignores a currency sign or short currency word around a number; the balboa's B/. was not in the list. Add it as a whitespace-delimited sign, as Word reads it: B/. 1,250.50 and 1,250.50 B/. are numbers, while B/.1,250.50 and B/ 1,250.50 are not. Tested in the locale reader and through =SUM(ABOVE); the localization parity doc records the behaviour.
* Table formulas: read the balboa cell in the document's language
The formula test summed a B/. cell through ev, which supplies no document
or language, so a regression in locale selection could pass. Evaluate
=SUM(ABOVE) with a document whose default text is es-PA and a pl-PL
editing language: 1,250.50 is a number under es-PA and not under pl-PL,
so the assertion fails if the formula stops consulting the cell's
language. The converse case reads B/. 1 250,50 from a pl-PL document with
an en-US fallback, which English alone cannot parse.
* docs: close the parenthesis in the localization-parity change note (#3546)
---------
Co-authored-by: Francisco Herrera <14956584+alosslessdev@users.noreply.github.com>
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
* WC6-06: preserve adjustRightInd in DOCX
* WC6-06: honor grid indent defaults and XML order
* Update lockfile for DOCX test dependency
* Format DOCX ordering assertion
* Keep main lockfile for paragraph property PR
* PR 5384: tighten grid indent docs and regression
* PR 5384: match paragraph check formatting
`comment_export_discovery_scales_with_unrelated_markers` timed `notes` on 2,000 and 16,000
markers through /proc/thread-self/schedstat. That count moves only at a scheduler tick or switch,
so a 17 µs baseline read as nothing or a whole tick, and gate failed on linear code (58.68x on
PR #3228's run 41932894607). As #5608 did for the balloon anchors, the test now counts the
characters the walk visits (`wordcraft_doc::work`): 4,010 -> 32,010 steps, 8.0x, every run. A
prefix scan per marker injected into `notes` counts 63.9x and fails it.
`work::steps` becomes public so a crate that walks a paragraph's text itself can count it; outside
tests it is still nothing.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* M6: preserve numbered REF results during field updates
Adapt and validate the patch published in WordCraft #4533, retaining main's bounded numbering renderer and adding relative-context/cache guards.
Co-authored-by: Alyosha-dot-shop <218732667+Alyosha-dot-shop@users.noreply.github.com>
* M6: reuse bookmark lookup for numbered REF updates
* REF: use explicit imports for the workspace clippy gate
* #4609: adapt synthetic fixtures to shared document data
* #4609: checkpoint review fixes and isolated validation notes
* #4609: retain rendered REF labels and isolate relative context by list
---------
Co-authored-by: Alyosha-dot-shop <218732667+Alyosha-dot-shop@users.noreply.github.com>
Co-authored-by: Martin Ball <martin.ball@ucscaregroup.co.uk>