* fix(ui): Overscroll off keeps an image smaller than the window centred (#10488)
With Tools › Overscroll off, an image smaller than the window stayed
wherever zooming around the pointer left it, inside the window, so
zooming out drifted it off centre. #1718 measured that on Photoshop 25.4,
but two reports of Photoshop itself (#10488, #4799) say it stays centred
with Overscroll off. The owner chose the centred behaviour:
scrollbars::clamp_axis now centres an axis on which the image is no
larger than the view; a larger image still can't leave its edges, and
Overscroll on keeps the free view.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(ui): Overscroll off centres a fitting rotated image (#10488)
CodeRabbit on #11085: the rotated clamp path was only tested from an
already-centred view. A 600 px square at 45 degrees, started off centre
in a 900-point view, now ends with its screen bounds centred.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* style(ui): keep the rotated centring test within rustfmt's width
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Type: add OpenType Numerals submenu (figure styles)
Type › OpenType › Numerals sets lnum/onum/pnum/tnum as a radio style
(Default through Tabular Oldstyle). Checkmarks follow the selection
start; TYPE-218-5 scorecard/gaps updated; parity floor 640.
* refactor(engine): one style-at-character lookup for the OpenType menus
feature_on_at and numerals_style_at each walked the runs the same way;
they now share style_at. NumeralsStyle::from_id and numerals_style_of
read the existing tables instead of repeating them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(text): persist OpenType numeral tags through PSD EngineData
Map lnum/pnum/tnum as EngineData bools and the width×height combos as
FigureStyle (Krita/Illustrator), so Numerals choices survive PSD round-trip.
* docs(gaps): note Numerals PSD figure-tag round-trip
* revert(text): keep PSD EngineData to the keys Photoshop writes
Backs out the PSD half of 18610324 and its gaps note (58813de3); the
engine reformat in that commit stays.
- `LiningFigures`, `ProportionalFigures` and `TabularFigures` are not
Photoshop EngineData keys (the commit says Photoshop only has
`OldStyle`), so writing them puts keys of our own into users' PSDs.
- The `/FigureStyle` code table (1-4) has no oracle: no Photoshop-authored
file in photocraft-corpus, and neither ag-psd nor psd-tools reads it.
AGENTS.md rule 3 asks for public specs or observation. PSD persistence
of the numeral styles can come back with a Photoshop-authored oracle.
The Numerals menu still round-trips through .pcraft.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(text): lock PSD EngineData to Photoshop OldStyle for numerals
Document AGENTS.md clean-room: only onum survives TySh round-trip; lnum/pnum/tnum stay .pcraft-only until a Photoshop-authored corpus oracle.
---------
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
jobs.cancel, cancel_all_jobs and closing a document (cancel_jobs_on) cancelled
any running job, so they could report a save as cancelled while its worker
went on to replace the file. The point of no return was tracked only by the
desktop shell's own Cancel/Esc path.
The commit point now lives in the engine's JobCtx: commit() and a cancel race
on one compare-and-swap, so exactly one wins.
- Cancelled before commit: commit() fails, the save writes nothing and the
job ends Cancelled.
- Committed: every cancel path leaves the job running and says so
(Session::try_cancel_job -> CancelRefused::Committing; cancel_all_jobs and
cancel_jobs_on return a CancelReport). The job then reports its real
outcome, applied even if its document was closed meanwhile.
- jobs.cancel (engine and desktop control) replies {cancelled: n} as before,
plus notCancelled: [{job, reason: "committing"}] for a save left running.
This is not an error: the job is fine and will report how it ended.
Single-job print refusals are unchanged.
- The shell's SaveCtl::commit and Cancel/Esc use the engine gate, so the
status bar, control and scripting agree.
Tests: cancel before commit (target untouched, Cancelled), cancel after
commit (refused, Done, file written), cancel_all and document close with
mixed committed/open jobs, a commit/cancel race, and desktop jobs.cancel
before and during the write with real files.
Offer TIFF compression, image pyramid and transparency options on Save As and Save a Copy, with matching CLI and automation parameters.
Follow-up: preserve the unsaved state when a copy discards layers over the document's own file, and check the other export formats for the same issue.
Closes#8827.
* fix(ui): Ctrl+Alt + wheel zooms the canvas like Alt + wheel (#10488)
Ctrl+Alt+wheel over the canvas did nothing useful because two conditions in
wheel_nav.rs excluded it from the Alt gesture:
let latest_alt = wheel.last().map(|(_, _, m, _)| m.alt && !m.command && !m.ctrl);
and, in the raw-event accumulator, every branch was filtered with
`!m.command && !m.ctrl`. Ctrl+Alt+wheel therefore never set the tracked `alt`
and never contributed a notch to `raw`: with Preferences > General > Zoom with
Scroll Wheel ON it produced nothing at all (raw stayed zero, so the zoom read an
eased ~0), and with it OFF it fell through to classify's `_` arm and merely
scrolled vertically, because egui keeps vertical_scroll_modifier = ALT and does
not fold the delta onto x when both scroll patterns match. The combination is a
dead/incorrect gesture today, so no working behaviour is being replaced.
The new `alt_gesture` helper forgives `command` only when `ctrl` is set as well.
winit mirrors a held Ctrl on Windows (and Linux) into BOTH ctrl and command, so
a bare `!command` test can never match what a Windows reporter typed; macOS
Cmd+Alt (command set, ctrl not) is left out of the zoom gesture (egui 0.37 does not fold it sideways either: with Alt held both scroll patterns match, so its delta stays vertical). `latest_shift` stays strict, so
Ctrl+Alt+Shift+wheel zooms continuously instead of snapping to the ladder, and
the Shift-only and plain branches are untouched, so plain Ctrl+wheel still
scrolls sideways (#635). `legacy_pinch_to_zoom` is untouched and cannot
double-count a fractional Ctrl+Alt pinch step: it matches `modifiers.ctrl`
whatever else is held, but it *replaces* the wheel event with `Event::Zoom`, so
`wheel_events` never sees that step a second time.
Deliberately out of scope: the other half of the report, the centred-zoom /
overscroll complaint. The per-frame view clamp at canvas.rs:1737-1741 runs only
when Preferences > Tools > Overscroll is off (it is on by default), so with the
defaults nothing pulls the view centre back, and what the reporter is seeing is
more likely the deliberate wheel_anchor nearest-image-point clamp
(view_xform.rs:239, #3720) or Zoom Clicked Point to Center. That needs its own
reproduction and issue, and is noted as still open in the dev log.
Tests, fail-first: each_modifier_maps_like_photoshop gained Windows Ctrl+Alt
rows (preference off and on, instant and eased, a fractional pinch step counted
once, and a macOS Cmd+Alt guard row) while keeping its Ctrl-only sideways-scroll
rows; new full-app kittest test
input_tests::ctrl_alt_wheel_zooms_the_canvas_like_alt_wheel zooms five notches
around the pointer and holds the document point under it. Both failed before the
production change and pass after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(ui): Ctrl + Alt + Shift + wheel is dead today, not a zoom (#10488)
CodeRabbit 4432102938 asked for an assertion on the Ctrl + Alt + Shift combination instead of
only a comment, and for the reason: the filter rejects the wheel event. Both are right. With ctrl
held, latest_shift (wheel_nav.rs:381) is false because it excludes ctrl and command, so the alt
branch of the filter (wheel_nav.rs:413-422) runs and its "!m.shift" drops the notch out of raw;
shift stays false, so the snap branch never takes it either.
The new row in each_modifier_maps_like_photoshop first asserted what my own comment claimed - a
continuous zoom, with Zoom with Scroll Wheel off and Animated Zoom off - and failed exactly that
way: "Ctrl + Alt + Shift + wheel zooms 10%, at once: [0.0 0.0] 1 snaps 0". Pan zero, zoom 1.0,
snaps 0, and the same numbers with Animated Zoom on, while plain Alt + Shift still snaps. The
gesture is dead, and only the ctrl half of it kills it.
So the three places that asserted a continuous zoom are corrected here instead of left to rot: the
module doc (wheel_nav.rs:8-11), the comment above the filter (wheel_nav.rs:407-413) and the parity
row (docs/ui-parity.md:193). They now state that the gesture is dead, point at the test that pins
it, and name the open question. No behaviour change: the repo has no measured Photoshop evidence
for this combination (the issue names only Ctrl + Alt), so the choice - (a) accept Shift in the alt
branch when Ctrl is held, making it zoom like Alt + wheel, or (b) treat it as exactly Alt + Shift
and snap - goes to a maintainer rather than being invented. Option (b) has a measured cost worth
knowing before anyone picks it: plain Ctrl + Shift + wheel pans sideways today ([40,0], zoom 1.0,
snaps 0, and no test covers it), and under (b) that becomes a no-op with Zoom with Scroll Wheel on.
Gates: cargo test -p photocraft-ui-egui --lib -- wheel = 102 passed, 0 failed (with the promised
assertion first: 101 passed, 1 failed); cargo clippy -p photocraft-ui-egui --all-targets
-- -D warnings clean; cargo fmt -p photocraft-ui-egui -- --check clean.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(ui): pin Ctrl + Alt + Shift + wheel with Zoom with Scroll Wheel on (#10488)
Review 4433950859: both Ctrl + Alt + Shift + wheel assertions ran with Zoom with
Scroll Wheel off, while the test comment, the wheel_nav.rs filter comment and the
ui-parity.md Wheel row stated the no-action result unqualified. Measured, not
copied: a temporary probe test ran one Ctrl + Alt + Shift notch (Line unit,
gesture_with, 60 frames summed) in all four preference combinations and printed
(pan, zoom, snaps) through an assert that was made to fail:
zww=off anim=off ([0.0 0.0], 1.0, 0) | zww=off anim=on ([0.0 0.0], 1.0, 0)
zww=on anim=off ([0.0 40.0], 1.0, 0) | zww=on anim=on ([0.0 40.0], 1.0, 0)
test result: FAILED. 102 passed; 1 failed; 0 ignored; 0 measured; 4526 filtered out; finished in 2.63s
The suggestion's premise holds: the dead-gesture claim is true only with Zoom
with Scroll Wheel off. With the preference on, alt_gesture keeps the notch out of
the branch's raw, but `alt != zoom_with_wheel` is false so the zoom arms are
skipped, and egui's own smoothed scroll delta (which the raw filter never gates)
pans the notch through the ordinary scroll path: one whole notch, 40 points
vertical, no zoom, no snap, with Animated Zoom off AND on (it only spreads the
same 40 points over later frames).
Probe deleted; the qualified version landed in all three artefacts:
- wheel_nav/tests.rs: the two dead assertions now say "with Zoom with Scroll
Wheel off", two new assertions pin the preference-on pan (near [0.0 40.0],
zoom 1.0, snaps 0, both Animated Zoom modes), and the comment and the
"not even a scroll" sentence carry the preference condition and the numbers.
- wheel_nav.rs classification comment: the dead-gesture sentence is qualified
the same way.
- docs/ui-parity.md Wheel row: same qualification plus the measured numbers.
Final gates: cargo test -p photocraft-ui-egui --lib -- wheel ->
test result: ok. 102 passed; 0 failed (also re-run after cargo fmt re-wrapped
one over-long assert, same result); clippy -D warnings clean; fmt --check exit
0; cargo xtask layers: OK, 31 crates, no layering violations.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(ui): keep Cmd+Ctrl+Alt + wheel off the Alt zoom path, pin AltGr (#10488)
alt_gesture now also requires !mac_cmd when forgiving command, so macOS
Cmd+Ctrl+Alt (command, ctrl, mac_cmd, alt) is not the zoom gesture. Tests pin
the macOS and Windows/Linux modifier shapes and that AltGr reaches egui as
neither ctrl nor alt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(ui): keep Ctrl+Alt+Shift + wheel as before, correct the AltGr note, pin the Alt key fallback (#10488)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Resolve conflicts with #5331 (EXR streaming): the block-by-block EXR import now
reads from the same identity-verified handle as the rest of the open path, and
docs/scratch-disks.md keeps both the EXR streaming pointer and the RAM/scratch
admission notes.
Main moved the GPU canvas into canvas/gpu_canvas/* and gained its own
budget, LayerFit and background-redraw machinery. Conflicted files take
main's version; the viewport streaming is re-integrated on top of main's
budget in the following commit.
One RAM detector: the cgroup v1/v2, macOS reclaimable and Windows commit providers move from
photocraft-io into photocraft_raster::memory::system (L0), which the residency budget (#5330)
and photocraft_io::memory both read. Unreadable container limits still fail closed, swap is
never counted as RAM. sysinfo is unified on =0.40.1 (raster only).
While the scratch disk is spilling (on, not paused), import admission credits the tile pixels
of an estimate against spill::scratch_capacity() (free space above the reserve, within the
Scratch Size Limit); decode buffers, conversion copies and merged planes stay RAM-only. Spilling
off/paused and the web keep the RAM-only policy unchanged. PsdMemoryEstimate gains tile_bytes;
the flat TIFF streaming check credits its tile surface the same way; engine approveMemory's
fast pre-check uses the credited hard limit and the import re-checks exactly.
- import_file_for_open_with decodes single-part flat scanline EXR block by block into
tiles; headers the band decoder refuses fall back to the whole-file decoder, so files
main opened still open. Band-time failures and cancellation publish no document.
- The RAM tile budget and scratch-disk checks live in exr_native (the PSD stream module
they came from is dropped); over-budget opens fail with IoError::OutOfMemory.
- The band decoder picks colour channels with exr_color_channels, like the byte decoder.
- Tests move into the per-crate it binaries and use import_file; an engine open-job test
covers the OpenSource::Path route. The ui-egui sync open test is dropped: main's
synchronous open reads bytes, and the io test covers the oversized header.
- docs/streaming-io.md, roadmap, scratch-disks and REL-212-1 describe EXR only.
This reverts commit 89ed4cb284 from the
#5332 branch so the GPU viewport streaming lands standalone. The import
work is rescoped separately in #5331 (EXR streaming only).
Keep the PSD memory admission work and upstream's scratch storage
together. Conflicts resolved:
- Cargo.lock: both sysinfo versions are required. 0.40.1 serves
photocraft-io's adaptive import budget; 0.37.2 serves
photocraft-raster's scratch volume and memory supervision. Cargo
metadata resolved the lock with no other changes.
- crates/engine/src/lib.rs: Session keeps pending_preflights and
next_pending_preflight beside scratch_runtime and scratch_epoch.
Session derives Default, so neither set needs a manual initializer.
Upstream: 655295b2a2
Main now takes the manifest envelope from craft-pcraft-format, which has
no full_composite member. The save writes the member after the envelope
when it writes composite/full.png, and read_full_composite reads it back.
A scratch write or create failure, or a volume about to drop below max(1 GiB, 5%)
free space, now pauses spilling with exponential backoff (1 s doubling to 60 s)
instead of disabling it for the session. Tiles stay resident, set_budget keeps
the latest budget for the resume, preference changes retry at once, and
spill::epoch lets the engine re-apply History's byte limit while paused.
spill::set_size_cap adds an explicit size limit; free space comes from sysinfo's
disk feature.
One unreadable tile used to block editing, saving and exporting every open document
(spill::check_integrity was process-global). Add Tile/Surface/Document::scratch_health
(free while no tile has failed) and Document::for_each_surface (ops' history walk now
delegates to it). edit_with checks the document before and after, jobs the document they
captured, import the result, save/export the document written; check_since is now
generation-only so a failure or recovery during an operation refuses it once and a retry
works. Caches check their own sources and purge once per new failure instead of every
frame. Fault-injection tests use a duplicated handle (the Unix file has no name).
The spill manager is process-global and never dropped, so the Unix scratch file was only
removed by the next launch's reap. Unlink it immediately after opening and read/write
through the one handle with pread/pwrite (readers can no longer reopen it by path).
Windows keeps delete-on-close and per-thread readers. The abrupt-exit test now asserts
that no file remains on every platform; docs/scratch-disks.md updated.
Resolve conflicts with main's prefs UI split, session/execute.rs move, GPU memory
budget (gpuMemoryMb, 0 = auto, kept from main) and effective_memory_usage_mb history
budget. Port the Memory Usage % and Scratch Disks preferences into prefs_ui, the edit
integrity guard into session/execute.rs, and adapt main's new pixel readers to the
guarded TileBytes API. Cargo.lock regenerated; scorecard regenerated.
* TIFF: open YCbCr images as RGB (#10316)
Scanners often write JPEG-compressed YCbCr TIFFs with 2x2 chroma
subsampling. The tiff crate decodes and upsamples them but returns raw
YCbCr samples, which the decoder rejected with "colour type YCbCr(8)".
The new tiff/ycbcr.rs converts them to RGB after decoding: JPEG data
uses the JFIF full-range BT.601 conversion (as libtiff does), other
compressions honour ReferenceBlackWhite and YCbCrCoefficients with the
TIFF defaults as fallback. 8 and 16-bit. YCbCr with extra samples, or
in other sample formats, fails with a clear unsupported error.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Contributors: add Yuki-init's display name
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* TIFF YCbCr: test that YCbCrCoefficients set the luma weights
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* TIFF: YCbCr tags with the wrong value count use the defaults
tag_f64s now returns None unless the tag holds exactly n values, so a
YCbCrCoefficients or ReferenceBlackWhite tag with extra values falls
back to the TIFF defaults instead of being truncated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lrgj5pnShYtsSnrsZq5iRn
* TIFF YCbCr: tests for tiled JPEG with ICC, big-endian 16-bit and planar
Covers the layouts the review asked for: a tiled 4:2:0 JPEG file with an
embedded profile (checked against an independent JPEG decoder), a
big-endian 16-bit file, and planar data in both byte orders.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lrgj5pnShYtsSnrsZq5iRn
* TIFF YCbCr: cite the formula sources and pin the green-channel choice
Documents where the conversion comes from: TIFF 6.0 Section 21
(YCbCrCoefficients, pp. 90-91) and libtiff's tif_color.c
(TIFFYCbCrToRGBInit, TIFFYCbCrtoRGB). Red and blue follow the spec;
green follows libtiff and adds Y unscaled, which matches the spec only
when the luma weights sum to 1. The module doc records this as a
deliberate compatibility choice, and a new test pins it: with weights
0.25/0.25/0.25, mid gray stays (128, 128, 128).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lrgj5pnShYtsSnrsZq5iRn
* Codecs README: YCbCr TIFFs with three 8- or 16-bit samples decode to RGB
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lrgj5pnShYtsSnrsZq5iRn
* Format tiff_ycbcr tests with rustfmt
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(tiff): compare grouped YCbCr decoding and update support notes
* style(tiff): align grouped decode call with test formatting
* style(tiff): apply repository rustfmt to grouped YCbCr regression
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Co-authored-by: FredApps <FredApps@users.noreply.github.com>
The doc merge left five Actions rows that disagreed (version 12 unsupported vs read; 120 vs
127 events). Found by CodeRabbit on #10776.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Supported table read from the batch's events.rs. The PRs' own sections (settings and command
options retained, Render defaults, steps kept disabled) are kept; the known-gaps tables that
#10702 and #10704 close are gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Distinguish available physical RAM from swap and commit headroom
- Adaptive soft and hard import thresholds with native fallback
- Interactive preflight confirmation modal before pixel decoding
- Headless and command parameters for explicit high-memory approval
- One-shot source-bound validation rejecting altered sources or degraded RAM
- Process-wide serialized decode gate with reentrancy support
- Documentation and synthetic regression test coverage
The Group F loader only mapped descriptors with an empty name and class
`null`, and nested shapes with an empty name. The reference application
names every step descriptor after the localized step (plus NUL), uses
the event code as its class and names nested shapes too, so every real
recorded Add To / Fade / Merge Layers ... step stayed a kept step and
never played (e.g. `addTo` with name "Add To\0", class `AddT` and a
nested `Rctn` named "\0"). The header is no longer checked; the shape
is matched by class only; key, order and enum checks stay strict.
IntW, Dstr and MrgL were registered with the string IDs
`interfaceWhite`, `distort` and none; the real ones are
`intersectWith`, `distribute` and `mergeLayers`, so steps written with
them stayed kept.
Co-authored-by: yoriswind <164435053+yoriswind@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With our fix-up ea4112a4ce on top of the author's head 686f08c884. docs/atn.md kept from the batch (regenerated at the end); development.md and parity doc rows as unions.
With our fix-up 6ab24508d8 on top of the author's head 8988b6bfd0. Conflicts resolved: import lists as unions (bare_from/bare_to taken from steps_other, #10631), event table, tests and parity doc rows as unions.
With our fix-up 1f524c378b on top of the author's head 713aa92ba5. Resolved with main: the version check sits in main's parse_layout (it merged cleanly); doc comments combined; atn_contract no longer expects version 12 to be refused and pins that an empty version 12 set parses; one ATTRIBUTION row for all ATN fixtures; the fixture README states which files were encoded how; development.md no longer counts the fixtures.
Out-of-range Oil Paint and Smart Sharpen settings kept the step as a
Photoshop step; the reference application clamps them when the step
plays, so the import now clamps (stylization 10.5 -> 10, Amnt 500.5 % ->
500, Rds 0.05 px -> 0.1). Export still refuses out-of-range params.
Oil Paint's light direction (LghD) had no range: LghD 500 mapped and
saved back as 500. It now clamps to the command's -180..180 on import
(500 -> 180) and export refuses {"angle": 181}.
Smart Sharpen mapped the legacy Lens Blur form, which the command's
Gaussian sharpen does not play, and wrote every step as legacy Lens
Blur. It now follows the PSD smart-filter mapping (photocraft-io
smart_map): only useLegacy false + blur GsnB maps (noiseReduction ->
reduceNoise); the reference default (useLegacy true, lensBlur), Motion
Blur and legacy steps stay Photoshop steps; export writes useLegacy
false, Amnt, Rds, noiseReduction, blur GsnB.
Co-authored-by: hectorviov <1501713+hectorviov@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With our fix-up 363e0f5e89 on top of the author's head 244082fb2c. Conflicts with main resolved as unions (new tests next to main's; parity doc revision rows).