* Fix Links panel stalls on network paths
* Fix Links review findings
* Fix Links worker gating and wasm build
A link check blocked on another document's unreachable share no longer
holds back the document now shown: after 200 ms a fresh worker takes
over (at most four alive at once) and the stale result is dropped.
relinked_asset used the native-only fingerprint_from, breaking the
wasm build; it now uses fingerprint_of. The EPS relink and IDML link
reach tests build Windows paths correctly.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Fix Links worker repaint and graphic read retry
Switching documents before the 200 ms takeover now schedules a repaint
for it: a blocked worker never asks for one, so without other input the
panel kept waiting. A failed graphic read is no longer cached as
missing, so the next check reads again. The legacy EPS fixture clears
its source fingerprint, as documents from before fingerprints have
none, so the test no longer depends on running within two seconds.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keep touched links OK and poll Links within seconds
Only a new size now marks a graphic Modified from its fingerprint. A
file with the same size but a new time (saved again, copied, restored
or checked out) is compared byte for byte, once per new time, and stays
OK when unchanged.
External changes are polled every two seconds instead of every five
minutes. The next check waits at least as long as the last one took
(up to five minutes), so a slow share stays idle half the time.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Wake Links when a worker slot frees; translate check strings
A worker thread asks for a repaint when it ends, so a document waiting
at the four-worker limit gets a worker as soon as a superseded one
returns from its blocked call, without other input.
Rows from a check keep a weak reference to their document. The key
names it by address, which a later document can't reuse while the
allocation is kept.
The Links panel's new strings ("Checking", "Checking links…", "Link
status is unavailable.") use the UI's ellipsis and are translated in
every interface language. The comments on Session::text_link_checks and
on the byte comparison in links::status describe what they do now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Links: rows from the document, statuses from a check that can stop
The panel lists links from the document on the UI thread (no file is
looked at) and fills in the statuses a worker check last found, kept
per document. An edit shows at once, and a check of an earlier edit
still supplies statuses instead of being discarded.
- A check of another document is told to stop and ends before its next
file; only one blocked in a file system call is replaced, so blocked
workers no longer hold their slots for a whole document.
- Checks are timed on the worker, and what they remember passes to the
worker that replaces them.
- A fingerprint is kept only for a file whose time was settled when it
was read, as for linked text. A file found unchanged by its bytes
gets its current size and time stored with the document
(remember_link_files), without marking a saved document as changed.
- Asset::set_link keeps link and fingerprint together; copies made by
Package and Copy Links To carry none.
- The Session holds one LinkListChecks. Update Link lists files only
when a linked story was edited in the layout.
- The unused "Checking links..." caption is removed.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* pt-BR: translate the 634 interface strings added since the catalog was completed
Covers the strings the catalog lacked after #495: tool names, menu
paths (Edit > Menus), Japanese typesetting dialogs (ruby, kinsoku,
mojikumi, warichu, kenten), special characters, Load Styles, Text
Frame auto-size, Data Merge, preflight, font replacement and About.
Tools are "Ferramenta X", parent pages "página principal", stories
"artigo"; menu paths reuse the existing menu names. IT, VA and Y:
stay as they are, as in every other language.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* pt-BR: one term per feature, the unit names, a catalog test and the docs
Review fixes for the new rows: Gradient Feather, stroke styles and type
on a path use the catalog's existing terms; warichu's Line Spacing no
longer reads as Leading; em/en dash, hair space, nonbreaking space and
hyphen and paragraph return take their Brazilian typographic names; the
file and word counts read correctly for one. File > New's unit names,
which reach tr at run time, are translated too.
The new test checks every pt-BR row's placeholders, padding and
ellipsis, and that each menu path is made of the menu names.
Docs: pt-BR re-measured at 2,416 of 2,419 strings (61% before) in
localization-parity.md, gaps.md (gap 21), target-app-parity.md and the
ROADMAP languages table and progress log. Readiness is unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Docs: trace each localization count to its commit; list Russian
The 1,724 baseline was counted at a127c6d5 (#355); recounting that
commit by the same rule gives 1,723, with pt-BR at the same 1,057. The
2,419 strings are main at 438d7b6d. The 2,021 figure is gone: it was
the same rule at the last commit of 2026-10-10, not another scale.
Russian (ru, 1,705 of 2,419) joins the other shipped languages, which
makes six beyond the twelve; ROADMAP's two "Also shipped" lines become
one that includes it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* pt-BR: the 12 strings main added since; docs re-measured at 2e36dcf9
Merging main brought window and guide commands, the InDesign shortcut-set
import and the credits' PRs column, which pt-BR didn't translate yet. Its
button label and result message reach tr as literals no other catalog
has, so they join the shared test's list of UI strings.
Re-measured at main 2e36dcf9: pt-BR translates 2,431 of 2,434 strings
(the Control panel's IT, VA and Y: stay English) and the unit names;
Russian 1,709 of 2,434. The PR adds 645 strings and 7 unit names.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Docs: pt-BR at 2,574 of 2,941 strings by i18n-report; Russian listed
Re-measured on main f3bcc886 with `cargo xtask i18n-report`: pt-BR
translates 2,574 of 2,941 interface strings (1,933 before); the 367 it
lacks are mostly strings main added since 2026-10-23. pt-BR to full is
4-6 h. localization-parity.md, gaps.md (gap 21), target-app-parity.md
and the ROADMAP languages table and progress log say so, each with its
revision row; the ROADMAP entry is dated today, as the review asked.
Russian (1,855 of 2,941) joins the other shipped languages, eight beyond
the twelve, and ROADMAP's two "Also shipped" lines become one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* pt-BR: drop two rows main removed; the test lists eight live labels
`cargo xtask i18n-report` flagged ten pt-BR rows whose keys its
interface list lacks. "/1000 em" and "Import InDesign Set…" are gone
from the code, so their rows go. The other eight still reach tr
(Auto-Sizing, Color Theme, Minimum Height/Width, No Line Breaks,
Reference Point, the shortcut-import result and the unavailable-fonts
message) and join the test's list of UI strings.
The report now counts 2,939 strings; pt-BR translates 2,572 of them
(88%), Russian 1,853. The docs say so; the PR adds 632 strings and the
7 unit names.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* pt-BR: the 364 strings main added since 2026-10-23
Keyboard Shortcuts sets (craft-keymap's dialog, import messages and
counts), panel docking, OpenType features, page presets, the Package
summary and the command dialog's parameter names. "Atalho", "conjunto"
and "escopo" for shortcut, set and scope; tab groups stay "grupo de
guias" and inks "de escala"/"especial", as in the catalog. Key names
stay as printed on keyboards.
`cargo xtask i18n-report` now lists only the Control panel's IT, VA and
Y: for pt-BR: 2,936 of 2,939 strings. The docs say so; pt-BR to full is
1-2 h (an in-app check and the older parent-page and story terms).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* pt-BR review fixes: Affix in, Dir, End, Pathfinder and three clashes; doc scales
- "Affix in" (where footnote prefix and suffix appear) is "Afixos em".
- "Dir" is a folder in file.save, recovery save and links.copyTo and a
direction in text.move: "Pasta/direção".
- The End key is "Fim", as Home is "Início".
- Pathfinder keeps its name ("Objeto/Pathfinder"), as "Pathfinder:"
already did, instead of "Localizador de caminhos" beside "demarcador".
- Pattern (stroke dashes) is "Padrão de traço", apart from Default
("Padrão"); Skipped is "Pulados", apart from "ignoradas"; Heading is
"Intertítulo", apart from Title and Section headings.
ROADMAP's "Also shipped" line now uses one scale (today's i18n-report of
2,939 strings); the pt-BR note names main ef6c3686, and Russian reads
1,855 of 2,939.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Place: a graphic dropped on a line gets its own frame and leaves the line alone (#2290)
drop_frame took any unlocked item under the pointer whose content takes a
graphic, so a file dropped on a line's stroke turned the line into a
zero-height image frame. It now also requires the item's path to enclose
an area, as the place cursor's click already does.
* Place: the next file of a drop doesn't replace a graphic that got its own frame (#2290)
A graphic dropped on a line, or on the page, gets a frame of its own at
the pointer but left frame_free set, so the next file of the same drop
found that new frame under the pointer and replaced the first graphic.
Any graphic placed by the drop now clears the flag.
* Command API: refuse a page index past the end, and say how many pages went
layout.pages.delete answered ok for an index past the last page and deleted nothing for it, so a caller off by one removed a different range and read a success; on a 336-page book that is page 337 instead of 336. Both commands now check every index against the document and refuse it with the sentence the MCP export commands have used since #70, return how many pages they touched so a no-op and a success are told apart, and state in their params that the indices count from 0.
* Pages: count the pages a delete will remove, not the indices it was given
delete_pages refused [0, 0, 1, 2] on a four-page document as if it would leave none, because the guard measured the length of the list instead of the pages it names. Same family as the index check next to it: the count was of the wrong thing. A test in the engine pins it.
* Pages: refuse a malformed page list instead of dropping what it cannot read
Review finding: page_indices used filter_map, so a list holding a string, a fraction or nothing at all was silently shortened — the caller was told a page was handled when it never was, which is the same mistake the command now refuses one level down. Every entry must be a whole number, and pages must be a list. Tests cover all three shapes, and the new test functions carry docstrings so the coverage check has nothing to report.
Found while testing crates/textimport/src/rtf.rs:43-63 (SKIP list) and :201 (`SKIP.contains(&word.as_str())`)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
craft-update moves from c5a1cbfb to c7c84230 on dotCooCoo/craft-libs, a
branch that merges storytold/craft-libs#687 and #696 into craft-libs main.
The check result window shows for every status but idle and stays closed
while the startup prompt is open. At the first start after an update it
reports that the update installed or why it failed. Its Restart to Update
calls Updater::restart_to_update. The install helper now starts when the
app exits and drops the updater.
Found while testing crates/color/src/gradient.rs:~101-110 (expanded_stops); the unclamped value comes from crates/idml/src/import/swatches.rs:128-129
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/textimport/src/docx.rs:132-133 (parse, Event::GeneralRef branch)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/doc/src/xref.rs:~100 (expand, early `return out` in the `tag_end` None branch)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Text: adjust hyperlink and placeholder source ranges on story edits (#1219)
Text-anchored hyperlinks and data-merge placeholders store byte ranges
(`start..end`) pointing into story text. When text earlier in the story is
inserted or deleted, these offsets remained unchanged, causing hyperlink
targets and data-merge placeholders to drift across unrelated text or
point to invalid UTF-8 boundaries.
This change:
- Implements `Document::adjust_text_ranges` in `designcraft-doc` to shift
or resize text hyperlinks and placeholders when story text is edited.
Pure insertions at or before the link shift it forward; insertions
inside expand the link's span; deletions before shift it back; deletions
consuming or overlapping the link clamp to surviving text or prune it if
entirely consumed.
- Cleans up obsolete hyperlinks and placeholders in `set_story_text`.
- Wires range adjustment into engine text commands: typing, deleting,
selection replacement, autocorrect, `story.replaceRange`, find/change,
and spelling change.
- Adds comprehensive unit tests in `designcraft-doc` and 7 integration
regression tests in `crates/engine/tests/hyperlink_offsets.rs`.
* fix(engine): adjust hyperlink and placeholder ranges for physically removed text under track changes
* fix(engine): adjust text ranges for formatted paste, case change, and drag move
* fix(engine): record change case adjustments per character to preserve enclosed hyperlinks
* fix(engine): adjust text ranges for track changes, inline markers, and preserve zero-length ranges
- Route Track Changes accept/reject deletions through adjust_text_ranges, and transfer replaced text hyperlinks to inserted text so links survive accept.
- Hook inline marker operations to adjust_text_ranges (notes, endnotes, xref, anchored objects, index markers, tables, editorial notes, unvar, and place text).
- Replace datamerge::shift_after with adjust_text_ranges.
- Only prune ranges that the edit emptied (s < e && s_new >= e_new), preserving zero-length ranges.
- Drop inverted ranges first and use saturating arithmetic in remap_range.
- Preserve doc.hyperlinks and placeholders ordering in move_text_ranges by mutating in place.
- Remove dead wrapper mark_deleted and call mark_deleted_reporting.
- Fix changes.track invocation parameter to 'on'.
- Add comprehensive regression tests in doc and engine covering all review items.
* fix(text): guard cell find/change, pre-record tracked replaces, shift para xref anchors
* fix(text): record stable IDs for tracked replacement redirects
* fix(text): reconcile range adjustments for review feedback
* fix(text): adjust ranges on delete surrounding and clean up dead auto edit
* Prune text ranges on TOC index and live caption regeneration
* test(engine): format regenerating_index_prunes_story_hyperlinks line wrap
About > Contributors still showed an old snapshot, because nothing
regenerated contributors/contributors.json before a release. Opt in to
the shared refresh from craft-common-actions: a read-only credits job
regenerates the snapshot for the release commit, and the shared release
workflow installs it before every source build. The snapshot is not
committed back.
Found while testing crates/color/src/cms/lab.rs:25-27 (Lab::hue)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/geom/src/hit.rs:28-35 (intersects_rect: `path.anchors().any(|(_, _, a)| r.contains(a.p))` and the early-out using `b.intersect(r)`)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Flex Layout's convert turns a multi-state group into a frame that keeps
its states. PDF export draws every object pasted into a frame, but the
white overprint check walked only the active state, so a white
overprint in another state printed unreported. Which children export
draws is now one function, drawn_children, used by both.
83d11697 reported white ruby and kenten on the frame that shows the run's
first character. The review commit before this merge does the same through
white_overprint::frames_showing, and also reports a run on every frame it
spans, leaves overset text out, and reads the composition cache only when a
story has white ruby or kenten. preflight.rs keeps that version; the tree
is unchanged by this merge.
Ruby and kenten in a threaded story were reported against the story's
first frame, on its page, wherever the text was; they are now reported
against the frame that shows them, and text no frame has room for isn't.
Parent pages' items and anchored objects print too and weren't looked
at; items on a hidden or nonprintable layer, or in a hidden or
nonprinting group, don't print and were reported. The check now walks
the items as PDF export draws them, with its rules (the parent page
items a page shows now come from one function both use), and a frame
PDF export leaves out reports none of its ruby and kenten.
A fill, stroke or stroke gap set to overprint in [Paper] or in a process
colour that lays down no ink (C=0 M=0 Y=0 K=0, gray 0, RGB white, a 0%
tint) disappears on press. Preflight now reports it as whiteOverprint, and
ruby and kenten set to overprint in such a colour too. A spot ink (a white
ink or a varnish) prints whatever colour stands in for it, so it isn't
reported. Text has no overprint of its own, so white text knocks out and
prints, and isn't reported. The old Preflight note that PDF export drops
ruby and kenten overprint is gone: export writes it.
* IDML import: a page item's or a text run's hyperlink survives (#1446)
IDML carries a hyperlink as three elements: a source, a destination and a
`Hyperlink` joining them by `Self`. Only `CrossReferenceSource` was mapped, so
an image frame linked to a photo and a run of text linked to a page were
dropped on import, although both are native here and both export to PDF.
The names are collected as the package is read and the two ends put together
in `finish`, once the pages, page items and anchors they name exist. A link
whose source or destination is not in the file is dropped with a warning
saying which, rather than silently or as a link that goes nowhere.
Carries all three review fixes:
- a source recorded in a table cell's or a footnote's story is dropped. Those
stories are built under the StoryId(0) placeholder, so the link named a
story nothing can look up, with offsets into text that is not there. The
warning says where the text was instead of claiming it was not text.
- the same for a HyperlinkTextDestination in such a story: its anchor id is
(0 << 24) | k, so it collides with other cells' anchors and belongs to no
story the cross-reference index holds; the link resolved to Anchor(1) and
the PDF left the annotation out without a word.
- the index's topic cross-references go in at the start of the first story
with text after the ranges are recorded, so the ranges for that story are
shifted with the markers; a range left alone covered the wrong nine bytes.
Rebased onto the post-split main, hunk by hunk: the Importer struct, its
fields, the element scan, hyperlink_list and finish are in import.rs, while
StoryBuilder, walk_story and the anchor list are in import/story.rs. finish
answers (Document, Vec<String>) so the warnings reach the caller.
* An anchor a hyperlink cannot mean is refused, not ignored
This branch carries its own copy of `dest_param`'s anchor arm -- the IDML import
needed anchors as link destinations and added it here independently of #1555 --
and it has the same hole, so review is right that the fix has to be here as well
rather than only on that PR.
`dest_param` returns `Result<Option<_>>`. `None` still means no destination was
given, which `hyperlink.edit` needs for an edit that only renames; an `anchor`
that is given and names none is an error.
An unknown *id* was accepted outright, since `as_u64` answers for any number, so
a link was stored to an anchor nobody has and the export quietly left it out. On
`hyperlink.edit` the `None` read as "no destination given", so the command
reported success and kept the old destination, which is the case review
reported. An unknown *name* did already fail on `hyperlink.create`, but with
"give `url`, `email`, `page` or `anchor`" -- the wrong problem named.
The same change is on #1555, deliberately: each PR introduces this code on its
own line and each has to be right on its own. Whichever lands first, the other's
conflict is the same text on both sides.
* IDML export: a document's hyperlinks are written out (#1446)
Import reads hyperlinks now, but export dropped them, so a round trip lost
every link the document had. The three elements IDML wants are written: a
destination, a source, and a `Hyperlink` joining them.
A link on a run of text is a `HyperlinkTextSource` around the ranges it
covers, and the element nests inside a `ParagraphStyleRange`, so a link whose
text crosses a paragraph is written as one source per paragraph and each piece
stays clickable. Consecutive ranges of one link share a single source element.
Carries both review fixes:
- a link to an anchor whose text has been deleted is not written. The anchor
is resolved first, as the page destination's page already was, so the file
cannot hold a `Hyperlink` whose `Destination` names no element.
- every `Self` names one element. Three links on one frame wrote three
`HyperlinkPageItemSource` elements all called i12, and each link to a page
pushed another `HyperlinkPageDestination`; the item source carries the
link's index now, and the page destination is written once per page as
url_dest_el already did for URLs. The round-trip test collects every `Self`
in the designmap and asserts none repeats, so an element added later is
covered without anyone remembering to extend it.
Rebased onto the post-split main, on top of the rebased #1652: the `Ex` state
is in export.rs, the hyperlink elements are written in export/designmap.rs
with url_dest_el, fnv and has_self beside their only caller, and the run
cutting and wrapping is in export/story.rs.
* Split the hyperlink and index-marker steps out of Importer::run, under clippy's complexity limit
* Split the hyperlink and index-marker steps out of Importer::run, under clippy's complexity limit
* Split the hyperlink steps out of story_para, under clippy's complexity limit
* IDML import: say when a link's frame is anchored in text, which PDF export cannot place yet
---------
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
* IDML import: a page item's or a text run's hyperlink survives (#1446)
IDML carries a hyperlink as three elements: a source, a destination and a
`Hyperlink` joining them by `Self`. Only `CrossReferenceSource` was mapped, so
an image frame linked to a photo and a run of text linked to a page were
dropped on import, although both are native here and both export to PDF.
The names are collected as the package is read and the two ends put together
in `finish`, once the pages, page items and anchors they name exist. A link
whose source or destination is not in the file is dropped with a warning
saying which, rather than silently or as a link that goes nowhere.
Carries all three review fixes:
- a source recorded in a table cell's or a footnote's story is dropped. Those
stories are built under the StoryId(0) placeholder, so the link named a
story nothing can look up, with offsets into text that is not there. The
warning says where the text was instead of claiming it was not text.
- the same for a HyperlinkTextDestination in such a story: its anchor id is
(0 << 24) | k, so it collides with other cells' anchors and belongs to no
story the cross-reference index holds; the link resolved to Anchor(1) and
the PDF left the annotation out without a word.
- the index's topic cross-references go in at the start of the first story
with text after the ranges are recorded, so the ranges for that story are
shifted with the markers; a range left alone covered the wrong nine bytes.
Rebased onto the post-split main, hunk by hunk: the Importer struct, its
fields, the element scan, hyperlink_list and finish are in import.rs, while
StoryBuilder, walk_story and the anchor list are in import/story.rs. finish
answers (Document, Vec<String>) so the warnings reach the caller.
* An anchor a hyperlink cannot mean is refused, not ignored
This branch carries its own copy of `dest_param`'s anchor arm -- the IDML import
needed anchors as link destinations and added it here independently of #1555 --
and it has the same hole, so review is right that the fix has to be here as well
rather than only on that PR.
`dest_param` returns `Result<Option<_>>`. `None` still means no destination was
given, which `hyperlink.edit` needs for an edit that only renames; an `anchor`
that is given and names none is an error.
An unknown *id* was accepted outright, since `as_u64` answers for any number, so
a link was stored to an anchor nobody has and the export quietly left it out. On
`hyperlink.edit` the `None` read as "no destination given", so the command
reported success and kept the old destination, which is the case review
reported. An unknown *name* did already fail on `hyperlink.create`, but with
"give `url`, `email`, `page` or `anchor`" -- the wrong problem named.
The same change is on #1555, deliberately: each PR introduces this code on its
own line and each has to be right on its own. Whichever lands first, the other's
conflict is the same text on both sides.
* Split the hyperlink and index-marker steps out of Importer::run, under clippy's complexity limit
* IDML import: say when a link's frame is anchored in text, which PDF export cannot place yet
---------
Co-authored-by: skylar taylor-barrick <hi@skylar.cc>
Dylan M. Taylor
skylar taylor-barrick
aiongg
ljkx
autocraft-bot[bot]
kuseidos
dotCooCoo
NetterSilverandClaude Haiku 5.5
Youssef Azriouil
Kushagra Kumar
Luis
Zinadin Zidan
Hamza Naveedandskylar taylor-barrick