cap-std's Windows `rename`, `remove_file` and `create_dir` are path-based:
they append the request to the handle's own path, read back without its
`\\?\` prefix, and call `std::fs` (sunfishcode/cap-std#348). Below a drive
root that still resolves; below a UNC root (`\\wsl.localhost\...`, any
network share) it becomes the relative `UNC\wsl.localhost\...` and fails
with ERROR_PATH_NOT_FOUND. `AuthorizedWorkspace::write` then left only its
`.pcsave.tmp` behind, `export_sequence` could not create a new directory,
and `ExportSequence::cleanup` could not remove its own frames.
Opening files and directories and `stat` are handle-relative and work
below any root, so everything is built on them. Files are removed through
the held directory handle (`remove_file_in`: `Dir::remove_file`, then on
the UNC `NotFound` an open relative to the same handle with
FILE_FLAG_DELETE_ON_CLOSE | FILE_FLAG_OPEN_REPARSE_POINT). The two steps
that have no handle-relative form, creating a directory and renaming the
finished temporary file into place, run on the canonical root path
(trusted launch configuration, never the request) below directories that
were each verified a plain directory through the handle of their parent
and are held open for the duration (`held_directory_chain`,
`pinned_directory_chain`). cap-std opens directories without
FILE_SHARE_DELETE, so a held ancestor cannot be renamed away or replaced
by a link while the ambient step runs. The rename is therefore the same
temp-file + rename replace as on every other root: a failed save keeps the
original, and no temporary file is stranded. Drive roots and Unix are
unchanged.
The regression tests use the temp drive's administrative share
(`\\localhost\C$\...`) and skip on a runner that does not expose it; the
pinned-ancestor test shows a held chain refusing the swap with
ERROR_SHARING_VIOLATION. The security model notes document the two
ambient operations.
Fixes#10317
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Remove the Azerbaijani menu coverage test (red on main), the crop Fill
label coverage test and the has() loop in the pen path Esc test. A
missing string falls back to English; cargo xtask i18n-coverage lists
the gaps. docs/contributing.md now says tests never require a string to
be translated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Window menu item from #11669 and the Azerbaijani catalog landed
together, so azerbaijani_covers_every_menu_label_and_path fails on main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#11005's new head adds path.toShape to the vector-path list next to
path.stroke; the batch's 95c7a5d17e had added it next to path.set. Both
merged cleanly into a duplicate or-pattern; keep the contributor's.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The bar draws inside the canvas pass, whose write-back of its own camera
dropped the fit that commit_crop asks for, so Done kept the old zoom. The
stored view now wins when the bar changed it, as a view menu row's does.
done_fits_the_cropped_document starts at 50 % and fails without the fix.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(export): preserve full JPEG chroma in Export As
* style(ui): simplify blend preview condition for Clippy
* test(export): verify full JPEG chroma at every quality
Exercise real JPEG export and decode at five qualities across U8, U16 and F32. Read all sampling factors from the main frame while skipping metadata segments and embedded thumbnails. Keep the colour-error limit at quality 100 and verify that the layered source is unchanged.
* Keep JPEG regression compatible with current Clippy
Restore the upstream blend-preview and IME predicates, removing unrelated lint-only changes. Use array chunks in the JPEG colour-error assertion for current stable Clippy.
* Use quality-dependent JPEG sampling in Export As
Keep 4:2:0 through quality 50 and full 4:4:4 above it, using the existing percentage-scale cutoff from Save for Web. Photoshop 27.11 modern Export As measurements show 4:2:0 at levels 1-4 and 4:4:4 at 5-7; no exact numerical mapping of the two scales is claimed.
Exercise both sampling regimes through 15 real JPEG exports across U8/U16/F32, check all percentage qualities in settings, and retain the Save As default. Boundary mutations fail at 50 and 51 as expected.
The automation registry check takes a `path` parameter for a host file
unless the command is listed as taking a document vector path, as the
other path commands are. path.toShape's `path` is the inline path being
drawn, so it belongs in that list; integration batch #11643 failed on it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EZbgszHyuEuWAzxaamteuU
Unit tests no longer read PHOTOCRAFT_UI_DIRECTION (cached per process, so a
developer's shell could flip the Hebrew/Urdu and shell assertions). The
Hebrew/Urdu test checks glyphs only when craft-fonts carries a face for the
script (the eager provider draws it in the first frame, as in CI with
CRAFT_FONTS_REQUIRED); without it the harness allows tofu and only the
mirroring is checked, instead of panicking on a host without the fonts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Draw the Crop task bar only on the active document's tile: a click into
another tile activates its document mid-pass, and the tile drawn after it
put the new frame's bar on itself for that frame.
- ui.inspect reports the bar hidden (and no rect) as soon as it can't show,
not only after the next draw.
- ui.set merges panels.task_bar field by field: {"at": [x, y]} keeps a pin,
{"pinned": true} keeps a dragged place; unknown nested fields are refused.
- Tests for each (each fails with its fix reverted), plus the task bar's
preset dropdown picking 1 : 1 in sync with the options bar.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The automation registry check reads a `path` param as a host filesystem
path unless the command is listed as taking a document vector path, as
path.set is. #11005's new path.toShape takes an inline vector path, so it
joins that list; CI failed registry_filesystem_path_params_are_classified.
Co-Authored-By: DingusMcRinkyDinky <DingusMcRinkyDinky@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- zoom field: a menu's I key drops İ/ı text whatever the UI language (the keyboard layout is
not the interface language).
- dropdown type-ahead: one entry per list tagged with its language, cleared on any frame in
another language, so a prefix never carries over a quick switch and back.
- Hebrew and Urdu: live shell test selects each catalog, checks mirroring and glyphs.
- Azerbaijani restart test resets the drawing language before the restarted shell loads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The pinned central font manifest has no Tamil face (storytold/craft-fonts#63 adds one) and the
Linux runner has none installed, so the strict missing-glyph policy panicked on the Tamil
labels. Tamil keeps every label and state check; a guard fails once the manifest ships Taml.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The reference application's New Guide dialog has Orientation, Position
(a text field that takes typed units such as "2 cm") and Color, but no
unit dropdown; measured on 2026-11-06. Typed units already work on main,
so the PR only adds something the reference lacks.
This reverts merge 22e03d0bcb.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Linux runner has no GPU, so the GPU tests ran on Mesa llvmpipe over
GL, which can't render Rgba32Float: 57 GPU checks (50 GPU parity, 3
canvas, 2 adjustment preview parity, 2 effect-map recovery) returned
early and passed without comparing anything. A new step after Test
installs mesa-vulkan-drivers and runs the nextest `gpu` group on
lavapipe (Vulkan), reusing Test's build, and fails unless the parity
tests report the Vulkan backend and no test printed a skip.
No code change: photocraft's tests take whatever adapter wgpu returns,
software adapters included. Follows the CI half of lightcraft#1148.
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Add an opt-in WinTab backend for pen pressure on Windows
Windows Ink stays the default. Preferences > Tools > Pen Tablet has a Tablet
input dropdown (Windows Ink (default) / WinTab). WinTab opens a WinTab 1.4
system context on PhotoCraft's window and polls pressure and the eraser end
(tilt on devices that report orientation axes; rotation only with a pen and
driver that report twist) into the stylus feed. The row says to turn Windows
Ink off in the tablet's software utility; the choice applies at next launch.
- crates/tablet: Windows-only wintab/wintab-sys backend. Devices with
orientation axes use the orientation packet; pressure-only devices open a
packet without it and report neutral tilt. Leaving is the driver's hardware
proximity message (WT_PROXIMITY) and the packet's out-of-context status, with
a long fallback for drivers that never report either. Overlap priority
follows window focus, and the message hook requests a frame for each WinTab
message so pressure-only strokes update without cursor movement.
- apps/photocraft: install the backend when the preference is on; poll it at
the start of each frame; the installed wake repaints the UI.
- ui-egui: TabletPollFn service; the Windows-only Pen Tablet row with the
Tablet input dropdown and the note.
- engine: tools.tabletInput (windowsInk/winTab), default windowsInk.
- docs: hardware-parity WinTab gap entry, qualifying tilt to orientation axes
and rotation to twist-reporting pens.
Tested locally on a Wacom tablet with Windows Ink off: pressure, tilt and the
eraser end reach the brush (rotation where the pen reports twist); the default
path is unchanged.
* WinTab: a message that finds the state borrowed is skipped, not a panic in the hook
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* WinTab: rustfmt
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Scorecard: count the Tablet input setting
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
- #11249: a recorded Blend If Set that also carries real layer effects (a
Lefx with more than the scale) stays the recorded step instead of playing
and saving back without the effects.
- #11541: the toolbar "..." flyout owns Escape, so Escape with it open no
longer deselects or leaves full screen; a failed Edit Toolbar from a left
click on "..." reports its error in the status bar like the flyout row.
- #9862: the click-on-another-layer check uses the commit's targets (the
link group as it is now), so a layer linked after the box opened is not
"another layer".
Each has a test that fails without the fix.
Co-Authored-By: NaufalDzz <NaufalDzz@users.noreply.github.com>
Co-Authored-By: TrevorBrown-dev <TrevorBrown-dev@users.noreply.github.com>
Co-Authored-By: Higgs32584 <Higgs32584@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main gained the Crop tool's "Background (default)" / "Transparent (default)"
Fill labels and the Pen tool's refused-draft messages after these catalogs were
written, and two tests require them in every catalog:
the_fill_labels_are_translated_where_the_old_label_was (az, th, ta) and
esc_ends_a_refused_draft_and_return_keeps_it (th, ta). Built from each
catalog's own terms for Background, Transparent, Default and Work Path.
Co-authored-by: Thanawatttt <89511939+Thanawatttt@users.noreply.github.com>
Co-authored-by: NewsBaba <346041388+newsbaba24x7@users.noreply.github.com>
Co-authored-by: Tamer Albayrak <56173603+Taxperia@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Present the live scratch state as current configuration rather than the cause of a file's refusal. Explain that configured scratch capacity cannot override caller/build limits or replace non-spillable memory. Keep the typed error and admission policy unchanged, and add wording regressions.
Formatting and whitespace checks passed. No local build or test execution; validation awaits hosted CI.
The Crop tool gets the reference application's contextual task bar: a small floating
bar centred 11 pt under the crop frame with a grip, the preset dropdown (the options
bar's own, so both stay in sync), Straighten, a ⋯ menu (Hide bar, Reset bar position,
Pin bar position) and, while a crop is pending, Cancel and Done. It follows the frame,
stays inside the canvas, and stays put once dragged or pinned. Window › Contextual Task
Bar (after Tools, as in the reference) shows and hides it.
The bar itself is a reusable widget (widgets/task_bar.rs): position rule, grip, ⋯ menu
and the hide/pin/place state in Panels (kept with the panel layout across restarts,
read by ui.inspect › taskBar, set by ui.set {panels}). A context supplies only its
controls and where ⋯ goes; the Crop contents are tools/crop_task_bar.rs. Generative
Expand stays out until #41.
Also: Lucide pin and rotate-ccw icons, six new strings in all 27 catalogues, parity
661/661 (FLOOR raised), control-protocol docs. Three tests about the options bar's
Ratio dropdown now hide the task bar, which shows the same dropdown.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
docs/korean-translation-validation.md records one agent session's checks
against a base from October (a "root task", a correction ZIP, other apps'
string tables); it says nothing a reader of docs/ can use. The 344 rows #8185
adds on top of main's ko.tsv stay (the merge kept main's wording for the 14
keys where both differ and dropped the 3368 rows main already has).
Co-authored-by: AKY <85772856+arksky1@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>