`blocks` and `fnv` exist only for `harris_and_describe_hashes_recorded_from_the
_serial_code`, which is `#[cfg(all(target_os = "linux", target_arch =
"x86_64"))]` because the hashes it compares against go through the platform's
hypot / ln / cos / atan2 / sin_cos. Everywhere else the test is compiled out and
the two helpers are left unused, so `-D warnings` fails on `dead_code`:
error: function `blocks` is never used crates/algo/src/features.rs:704
error: function `fnv` is never used crates/algo/src/features.rs:708
They now carry the same cfg as the only thing that calls them.
CI's matrix is ubuntu-latest and windows-latest and runs clippy on both, so this
fails the Windows leg as much as a local run on macOS: `target_os = "linux"` is
false for both.
#5344 added the export_presets field to Persisted; #5616 rewrote load_presets_json to read each set with take() and its Persisted initializer lacked the new field, so main stopped compiling (E0063 at presets/mod.rs:180). Read export_presets the same way as the other sets; the existing load step and round-trip test cover it.
Co-authored-by: 00xJ0 <[email protected]>
N is the whole algo-side Photomerge pipeline on 3 × 8 MP photos (register at the engine's
scale with the Auto layout, gain + vignette fit, seams, multi-band blend), 16 threads:
1042 ms before, 264 ms after, 3.9×, rounded to 4×. The engine's own glue (reading regions,
building weights, writing layers) is not in this number.
Registration and multi-band blending ran almost entirely on one thread per image (four
images on 16 threads), so most of the machine sat idle. Every output is bit for bit what
the serial code produced: a probe that hashes the public Photomerge entry points
(detect_many, match_pairs, align for every layout, photometric, seam_labels, multiband at
1 to 7 levels with 1 and 3 channels, with holes in the alpha, with images sticking out of
the canvas, plus direct harris / describe calls) matches on all 349 outputs before and
after, on 4 × 2 MP, 3 × 8 MP, 97×61, 211×157, 333×251, 320×240 two-image and 40×35 sets,
at 16, 3 and 1 threads.
- features `blur` (Harris window and BRIEF smoothing): the row pass runs rows in
parallel; the column pass cuts the image into bands of 64 columns, each holding the
`&mut` row segments of its own columns (no unsafe), run in parallel. Every column keeps
its own f32 running sum, started from its first window summed in order and updated in
the same order, so only the loop nest changes.
- `harris`: Sobel, the corner response and the non-maximum scan are parallel by rows (the
candidates are fully ordered by score, y, x afterwards, so arrival order is irrelevant).
`describe`: one descriptor per corner in parallel, kept in corner order.
- `reduce`, `expand`, `push_pull`: per-row maps in parallel. `reduce`'s column pass adds
five whole rows in tap order (same per-element sum, now vectorised); `expand` works out
its source columns once and truncates instead of calling libm `floor` (the clamped
position is never negative); `push_pull` builds each coarse row from its two fine rows in
raster order, which is the order the serial scatter-add used.
- `multiband`: grid fill per image, Laplacian difference, band accumulation (every row
walks the bands in the original order), normalisation, collapse and the coverage union
are parallel. Float additions are never reordered.
Left serial on purpose: the peak fold in `harris` (about 1 ms, an order-free max but not
worth the proof), the corner spacing grid, `seam_labels` (a dynamic-programming path whose
rows depend on the previous one), `photometric` sample gathering, and the pyramid levels
themselves (each level needs the one above).
Median of 5 (8 MP rows: 7), release, ms, interleaved before/after runs. The laptop was
busy with unrelated builds the whole time (load average 11 to 18 on 16 threads), so the
1-thread rows are noisy (a single thread is time-sliced against them) and rows can move
by 10 to 20 percent.
| Case (probe) | before (16 thr) | after (16 thr) | speedup | before (1 thr) | after (1 thr) |
|--------------------------------|----------------:|---------------:|--------:|---------------:|--------------:|
| register, 4 × 2 MP | 154 | 60 | 2.6× | 336 | 167 |
| multiband, 4 × 2 MP, 3 ch | 407 | 108 | 3.8× | 591 | 407 |
| multiband, 4 × 2 MP, 1 ch | 307 | 73 | 4.2× | 813 | 566 |
| whole pipeline, 4 × 2 MP | 437 | 160 | 2.7× | 716 | 504 |
| register, 3 × 8 MP (full size) | 327 | 124 | 2.6× | 978 | 391 |
| multiband, 3 × 8 MP, 3 ch | 2348 | 916 | 2.6× | 1742 | 1150 |
| whole pipeline, 3 × 8 MP | 1042 | 264 | 3.9× | 2264 | 1280 |
The 8 MP multiband row at 16 threads is the noisiest (minimum of 7: 959 to 224 ms, 4.3×);
the whole-pipeline row is steady. Max RSS for the 3 × 8 MP whole pipeline (the probe's
inputs are about 1.05 GB of it): 2.79 GB before, 2.79 GB after (three runs each).
Tests: the serial `blur`, `reduce`, `expand`, `push_pull` and `multiband` are kept in the
test modules as references and compared bit for bit over sizes that cross every band and
row boundary (1×1, 1-wide, 1-high, odd sizes, 63 / 64 / 65 / 129 columns, radius 0 to
wider than the image, holes, partial and zero weights, ROIs outside the canvas, 1 to 7
levels). Hard-coded hashes recorded from the old code cover multiband, push_pull,
reduce + expand everywhere (plain f32 arithmetic), and harris + describe on Linux x86-64
(they go through the platform's hypot / ln / cos / atan2 / sin_cos).
Claude-Session: https://claude.ai/code/session_01AfRHB1wHUjrJx6agBYGnVu
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Save: ask before overwriting a file another app changed (#6596) [WIP]
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Save: ask before overwriting a file another app changed (#6596)
File › Save now notes the size and modification time of a document's
own file when it is opened or saved, and checks them again before
writing back. If another app changed the file in between, a prompt asks
whether to replace it (Cancel keeps both; Save As saves elsewhere)
instead of silently overwriting the other app's changes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
* fix(windows): keep crash-recovery autosaves out of the roaming profile (#5648)
* test(windows): cover the local recovery folder and moving old autosaves (#5648)
A TySh warpValue, warpPerspective or warpPerspectiveOther beyond f32 became inf and the warp was silently dropped; they are now clamped to -100..100 like type.warpText.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Layer › Smart Objects › Stack Mode (`algo::stack::combine`) was one serial
per-pixel loop (1.0x parallelism, identical at 1 and 16 threads), and Entropy
built a HashMap per channel per pixel.
- Output pixels are computed in parallel in 4096-pixel chunks (`par_rows`, so
wasm keeps its serial path). Each chunk owns its scratch buffers.
- Each pixel uses exactly the same arithmetic in the same order as before, so
every mode except Entropy is bit-identical: 110 output hashes (10 modes x 11
cases: 3 x 24 MP, 7 layers of odd/mismatched sizes, transparency, HDR values,
1 layer, no layers) match the old code.
- Entropy: a `[u32; 256]` histogram plus a short list of the bins touched
replaces the HashMap; touched bins are summed in ascending order, so the result
is now deterministic. The old HashMap summed in random iteration order and
differed run to run (up to 1.2e-7 between two runs of the old code). New vs old
output: max abs difference 1.19e-7 (one f32 ulp near 1.0), on up to ~10% of
values in the 7-layer cases; the 3-layer 24 MP case is identical.
- Modes that never use the mean (Median, Min, Max, Range, Summation, Entropy) no
longer compute it; Median uses an unstable sort (same result under total_cmp).
3 x 24 MP, release, median of 5 interleaved before/after runs (load average
~3.5-4.5 from an unrelated build running alongside), ms:
| Mode | before | after (16 thr) | speedup | after (1 thr) |
|------------|-------:|---------------:|--------:|--------------:|
| Median | 1100 | 99 | 11.1x | 1064 |
| Mean | 308 | 36 | 8.6x | 338 |
| StdDev | 342 | 41 | 8.3x | 380 |
| Entropy | 4727 | 269 | 17.5x | 2730 |
| Kurtosis | 458 | 55 | 8.3x | 488 |
| Maximum | 368 | 36 | 10.1x | 346 |
The other modes land between 8x and 10x. Max RSS for Median is unchanged
(1.50 GB, of which 1.13 GB is the input layers).
Tests: hard-coded bit patterns recorded from the old code for every mode but
Entropy; Entropy against hand-computed values and run-to-run determinism; and a
chunk-boundary test comparing a multi-chunk image to per-pixel results.
Claude-Session: https://claude.ai/code/session_01AfRHB1wHUjrJx6agBYGnVu
Co-authored-by: Claude Opus 5.5 <[email protected]>
The image crate's Targa encoder let run-length packets run on from the end of
one row into the next. TGA 2.0 says a packet must not cross a scanline, and
readers that decode a row at a time misread such files: Pillow fails on them
with "buffer overrun", and texture tools and game engines reject them.
Targa is now written by our own encoder (codecs::tga), as Photoshop writes
it: RLE true colour or grayscale, every packet within its row, rows stored
bottom-up (the origin readers that ignore the descriptor assume), the alpha
bits in the descriptor and the TGA 2.0 footer. Reading stays with the image
crate.
One malformed set in the preferences file (gradients, swatches, styles...)
made load_presets_json ignore the whole presets blob, so saved New Document
presets vanished at every launch: the shell merges back only changed keys,
leaving the damaged set on disk. Each set now loads on its own; null and
malformed sets keep their defaults.
Refs storytold/photocraft#5602
Co-authored-by: Cursor <[email protected]>
Horizontal scale was applied after line breaking, so paragraph text below
100% wrapped early and text above 100% overflowed the box. Break box text
at the available width divided by the paragraph's horizontal scale (the
largest one when runs mix), scale the first-line indent and alignment
offset back to the box, and keep the scaled trailing space out of the
line extent.
Fixesstorytold/photocraft#5565
Co-authored-by: Cursor <[email protected]>
The prefix cache's chunk textures are only copied to and from
(COPY_SRC | COPY_DST), but went through Tex::new, which always creates
a default view. Vulkan rejects a view on an image with no sampled,
storage or attachment usage (VUID-VkImageViewCreateInfo-image-04441);
wgpu doesn't check it, so it only shows with the validation layers on,
once per new chunk while painting. The view was never used. Chunks now
hold a bare wgpu::Texture. No rendering change.
Co-authored-by: Claude Opus 5.5 <[email protected]>
usvg builds the glyph outlines of a text element (Text::flattened) with
identity absolute transforms, and the SVG importer read each path's
abs_transform(). Text positioned by its own or an ancestor's transform,
as Illustrator writes every text element (matrix(1 0 0 1 x y)), lost it:
it landed off the canvas, so nothing showed, or at its untransformed
place and size.
The converter now composes each node's transform from its parent chain,
which equals abs_transform() outside text, and sizes raster fallbacks
from the same transform. A new test checks translate, matrix, scale and
transformed groups against resvg's own render; it fails on main.
Co-authored-by: Claude Opus 5.5 <[email protected]>
On a document over about 5 MP with Low Resolution Previews on, a filter
dialog's live preview ran on a reduced proxy at every zoom, so at 100 % the
canvas showed the image in k-pixel blocks and hid exactly what Unsharp Mask,
High Pass and Smart Blur do.
ensure_filter_preview now caps its factor with proxy::view_factor(k, zoom),
the same rule adjustment previews got in #5764, so from 100 % the preview is
full resolution and stays reduced when zoomed out. The preview already runs on
a cancellable background worker, so the full-resolution path does not block the
UI thread.
Also in this change:
- canvas.rs and filter_preview_ui.rs share one canvas::filter_preview_factor,
so the in-dialog preview window cannot size textures from an uncapped factor
and rebuild them when the first result arrives.
- The cap is gated on cfg!(target_arch = "wasm32"), so the web build keeps its
proxy rather than running a full-resolution filter inline on the UI thread.
- The end-to-end test moved into mod tests and covers Low Resolution Previews
off; a new factor test covers 25/50/75/100/400 %, NaN and zero zoom.
The Community packages heading from #1149 landed before the maintainers' releasing note and the IMPORTANT status admonition, so those two paragraphs rendered under Community packages instead of Get started. Move the section to just above Documentation; the content is unchanged.
Strict workspace Clippy stopped on collapsible_match in affinity, raw, plugins, engine, ui-egui and xtask. Pure conditions become guarded match arms (each falling through to an existing empty arm), and the ruler drag state and the measurement command result are bound to locals, preserving evaluation order, error messages and fallbacks.
The same lint fixes also appear as a side commit in #5157 (David2766), which can drop it on rebase.
With a document open in the web app, Place Embedded was disabled because its menu enablement inherited the engine's native-filesystem check, although the menu path reads the picker's bytes and never touches a path. Place Embedded is now enabled whenever a document and a file-dialog service exist; Place Linked keeps the engine's native guard, since the web could only link to a bare file name (caught in #5817 by Higgs32584). An enablement test runs with the native check forced off, and a file-dialog test covers the browser answer shape, cancel, invalid bytes and the Free Transform handoff.
Fixes#1209. Fixes#5749.
egui places font runs left to right with no bidi reordering, so a right-to-left layer name showed every word correctly but in reversed order. bidi_text::visual puts a label into display order with the Unicode Bidirectional Algorithm (unicode-bidi, already in the tree via usvg): right-to-left words stay in logical order for the shaper, everything else in a right-to-left run is reversed around them and brackets are mirrored. The Layers panel's rows and the app-drawn title bar use it; the title keeps a left-to-right paragraph so '— PhotoCraft' stays at the right, and a long right-to-left name is cut at its logical end so its start survives. Display only: stored names and the OS title keep the logical text.
Fixes#5159.
Co-authored-by: ahmedhosnypro <[email protected]>
A command that fails on a layer id or name the document doesn't have now says which layers there are, top-down as id and name (at most 20, names cut at 64 characters), so an agent can recover from the error alone: ids are not positions, and a freshly opened PNG's Background is layer 2. NoLayer becomes UnknownLayer { id, known } at both command entry points, keeping the 'no such layer' prefix. A layer name that matches no layer exactly matches regardless of case; an exact match always wins.
Fixes#5058.
photocraft_algo::faces::detect finds faces in an RGBA image with a score, a box and five landmarks, through a pure-Rust port of libfacedetection's YuNet network (Shiqi Yu, BSD-3-Clause; licence, NOTICE and ATTRIBUTION rows included). The weights are the library's, converted to half precision (107 KB), parsed defensively. image.detectFaces reports the document's faces to scripts and agents; it is read-only with no history step and no menu entry. The network is compiled out of the browser build, where the command reports itself unavailable. Head outputs are pinned against the C++ reference.
Part of #3603.
Adds a Community packages heading to the README, noting these installers are maintained by community members, with the Gentoo recipe for media-gfx/photocraft-bin from the ::snakebyte overlay. The overlay's ebuild fetches the official Linux x86_64 release tarball with Manifest-pinned hashes that match the release's SHA256SUMS.
A single oversized input event reached egui text layout in a numeric field even though it could never become a valid value. number_edit now drops Text, Paste and IME events longer than 128 scalar values while the field has focus, and the custom parser rejects accumulated text over the cap before normalizing. DragValue keeps focus, accessibility, drag and arithmetic behaviour. Six harness tests cover the event kinds, the Unicode boundary, accumulation, undo, and document text fields staying unrestricted.
Part of #2150.
brush.defineFromSelection allocated the sampled tip and the selection mask infallibly over the whole selection, so a large selection could abort the process. Both now go through the engine's fallible allocator and return a normal error that leaves the brush, presets, history and document untouched. A failpoint test asserts all four.
Part of #212.
The caret's blink phase was tied to the wall clock, so an arrow key or keystroke that landed in the off half-cycle moved an invisible caret. A transient caret_blink_time on UiState is reset on every handled key event, on a click that places the caret, and while dragging; draw_overlay computes the phase from it so the caret is solid for 530 ms after each interaction, and the repaint is scheduled to the next phase boundary.
Fixes#5511.
With Only Web Colors on, every picker edit (field, sliders, hex, Lab/CMYK, screen sampler, Current) snaps through the clamped 216-colour ramp. The checkbox is remembered on the Color panel state so the foreground, field, command, Layer Style and type-style pickers open with the same setting. Opening with it on snaps the colour and its original together, so OK without an edit leaves a fill layer unchanged. The hex field follows the colour when another path changes it while focused.
Fixes#4995.
Creating a Solid Color fill from the Layer menu or the Layers panel footer now opens the shared Color Picker seeded with the foreground colour. The layer is created on OK in one history step; Cancel leaves the document and history untouched. Parameterised calls (CLI, MCP, actions) still create the layer directly.
Fixes#5144.
memory-stats reads /proc/self/smaps by default on Linux: every sample formats and parses one
block per mapping while holding the mmap lock. The benches' RssSampler polls it every 2 ms, so
on a profile of bevel_bench --size 4000x3000 the rss-sampler thread was ~22% of all samples
(smaps_account 14.7%, smaps_pte_entry 6.3%), contending with the measured threads.
Enable the crate's always_use_statm feature: one short read of the kernel's RSS counters, the
same number VmRSS reports. The measured code is unchanged; only the harness overhead goes.
parley's default overflow wrap only breaks at UAX #14 opportunities, so
a word longer than the paragraph box ran out past its edge and resizing
the box changed nothing. Paragraph text now sets OverflowWrap::BreakWord,
so such a word wraps at the box edge as Photoshop does; a space or hyphen
still wins over a mid-word break, a box that fits the word lays out as
before, and point text is untouched.
* fix(linux): avoid duplicate AppImage desktop launcher when managed by AM / AppMan (#6392)
* test(linux): cover AM / AppMan launcher detection in AppRun smoke test (#6392)
The local Laplacian filter behind Shadows and Highlights builds a whole
pyramid per reference value (up to 96). Three costs dominated it:
- blur_step called mirror() (an integer remainder) for every tap of every
pixel; the source rows are now found once per row and columns mirror only
within two pixels of an edge.
- up() blurred a zero-stuffed plane, three samples in four zero; it now adds
only the nonzero taps, in the same order.
- the small pyramid levels were split into one-row rayon tasks, costing more
in scheduling than the rows; planes under 32K samples run on one thread.
The output is bit for bit the same: a new test checks blur_step and up
against the plain versions on odd, even, tiny and parallel-sized planes.
Fixes#3708; part of #1540 (Flatpak needs its own follow-up: each sandbox has its own loopback).
A second launch hands its files to the running instance over an authenticated loopback channel (random port, 32-byte token in a private instance.json, constant-time check, bounded clients, requests and paths). Files are acknowledged only once the UI has taken them; the instance shuts the hand-off down before a CPU-compatibility relaunch; it runs only in the per-user settings mode. --new-instance and PHOTOCRAFT_NEW_INSTANCE=1 opt out.
Thanks to @MrCloudy2 (#2358), whose Linux version first handled the relaunch and per-user endpoints, and @Desayuno64 (#5273), whose shutdown flag and read deadline informed the final design.
Both harnesses built their device with egui_kittest's .wgpu() (egui's default limits) instead of
gpu_canvas::wgpu_setup() (the adapter's own limits, what the app requests), as #5954 fixed for the
background_redraw and navigator tests. On that device:
- canvas_16f's sixteen_bit_gradient_on_screen and thirty_two_bit_preview_exposes_values_above_one
panicked in egui-wgpu ("Failed to create staging buffer for index data", renderer.rs:984), every
run, single-threaded too;
- arbitrary_rotation_preview's two tests failed "the fixture must use the GPU canvas".
Both binaries pass on the app's device.
* Save As on Linux: the file type picked sets the extension (#6121, #4328, #4394)
The GTK file chooser behind the desktop portal keeps the typed name when
another file type is picked, and rfd 0.17.2 reads only the file name from
the portal's answer, so Untitled-1.psd with PNG picked was written as a
PSD. On Linux, Save As now asks the portal's FileChooser.SaveFile itself
(over the zbus already used for the portal), reads current_filter, and
gives the name that type's extension: an extension PhotoCraft saves is
replaced, none or another is appended. If that names an existing file, the
dialog opens again under the new name so the portal asks before replacing
it. rfd remains the fallback without a portal, and this can go once an rfd
release reads current_filter (PolyMeilex/rfd#341).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Save As on Linux: keep typed extensions, open on the stem, follow the request handle
Review follow-ups, checked on real GNOME 46 and KDE Plasma 6.7.5 sessions
(Wayland and X11):
- An extension the user typed is kept, as in PolyMeilex/rfd#341: only a
name with none, or still with the extension the dialog opened with, takes
the picked type's. GTK reports the type even when it is untouched, so
art.png with the PSD type left alone was saved as art.psd.
- The dialog opens on the stem (test), as on Windows (#4407): with
test.psd in the name and PNG picked, GTK asked to replace test.psd,
which the save leaves alone. A name that then exists is asked about
again under that name; the rule compares against the suggested name, so
a test.psd GTK completes back still becomes test.png.
- Response 2 (GNOME's dialog closed with Esc) is a cancel: it fell back to
rfd, which opened a second Save As.
- The answer is awaited on the handle SaveFile returns, which a portal
older than 0.9 picks itself (the Request documentation).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* style(automation): format smart filter refresh closure
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: skylar taylor-barrick <[email protected]>
* GPU memory: a graphics card's own memory isn't charged for the document; GPU Memory preference; GPU Memory Full in the status bar
Many layers hit a cliff: past what the GPU memory budget leaves for layer pages, every edit
that redraws the whole document (opacity, visibility, a nudge) goes to the CPU compositor.
Measured on 3000x2000 full-canvas layers (RTX 5090, 31 GB RAM, Memory Usage 8 GB): 150
layers took 2.6 s per such edit, 300 layers 6 s. The budget was Memory Usage minus the
document's pixels and History, which is right where the GPU shares system memory, but a
discrete card's memory doesn't hold the document, so 150 layers of 6 MP (3.6 GB) squeezed
the budget below what they needed.
- Automatic budget: discrete adapters are no longer charged for the document's pixels
(integrated and software adapters still are). 150 layers: 2.6 s -> 25 ms.
- Preferences › Performance › GPU Memory (MB, 0 = Automatic; performance.gpuMemoryMb) sets
the budget outright. 300 layers with 12 GB: 6 s -> 45-72 ms.
- The status bar shows "GPU Memory Full" while the active document's layers don't fit
(both status bar styles); hovering gives the numbers, clicking opens Preferences ›
Performance. Its Graphics rows show the budget and what the layers need. The slow CPU
redraw notice gives the numbers and points at GPU Memory.
- Headless: the preference is a normal pref (prefs.set), and ui.state's perf.timings
reports gpuLayerFit {doc, needed, budget} next to gpuBudgetMb.
Next (owner-approved plan): after #5370, read the card's real limit from the driver
(wgpu memory budget thresholds) and shrink on out-of-memory; then redraw the visible
area on the GPU first when layers still don't fit.
* GPU memory: typed discrete check; the status and budget follow a GPU Memory change at once (measure_fit); perf rows P60/P61 for whole-document edits among 150 full-canvas layers under the app's budget
* GPU memory on top of #5234: memory_budget takes the card's memory (vram) and the GPU Memory value; the document's pixels are charged unless the card's own memory is known; a set GPU Memory is capped at the card's share; benches use video_memory
* perf(gpu): cap the GPU memory budgets by a discrete GPU's own memory
The compositor and canvas budgets came from system RAM only, so a 16 GB card in a 96 GB machine planned for up to 8 GB of compositor pages plus 24 GB of display textures. Detect a discrete adapter's dedicated memory (DXGI on Windows through winsafe, amdgpu sysfs on Linux) and cap each budget at three eighths of it. PHOTOCRAFT_VRAM_MB simulates a smaller card. Fixes#4536.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* perf(gpu): keep a 16-bit canvas on the GPU in 8-bit when floats don't fit
With the canvas budget capped by a small card's memory, a 16-bit document under the float-canvas pixel limit could need more than the budget as Rgba16Float and fall back to the CPU canvas although its Rgba8Unorm set fits. Choose the float canvas only when it fits the budget, which is now shared atomically so format_for can read it while the renderer is locked.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* perf(gpu): detect a card's memory once per adapter, and test the Linux reader
video_memory built a DXGI factory each time it was asked, three times at
startup; it now remembers the answer per (vendor, device), so a GPU retry
that picks another adapter still gets its own. The amdgpu sysfs reader
takes the drm folder as a parameter and is tested against a temporary
folder: the matching card's VRAM, a card without the file, garbage ids, a
malformed size, no match and no folder.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
- The CPU canvas texture and the crop's beyond-canvas bands were reduced to fit MAX_TEXTURE (4096)
whatever the device allows. On a device with a smaller limit (2048 on some older GPUs and WebGL
contexts) a 2600 px document was uploaded at full size: egui asserts and the backend can't
create it. They now fit the smaller of the two. Test: cpu_canvas_texture_fits_the_device_limit.
- background_redraw and navigator GPU tests built their harness with egui's default wgpu limits
instead of the app's wgpu_setup (the adapter's own limits), and the GPU canvas lost the device
on it ("Device is lost", every run on an RTX 5090). They now use wgpu_setup.
- Eleven tests built a wgpu device without photocraft_testkit::gpu_lock, which testkit requires so
devices aren't created concurrently; they now hold it.
* Security: enforce the automation read boundary on stored smart-filter params
Smart filters store a command id and params inside the document, so a
crafted .pcraft could resolve stored ambient paths (displace mapPath,
color-lookup LUT files) at render time without the authorize gate:
synchronous smart-object rendering has no Session argument, spawned job
threads drop the requesting session's authority, and a trusted UI
render could warm the process-wide stack cache with a stage the gate
never saw.
Thread the requesting session's gate through a thread-local in jobs.rs
(entered around command execution, reinstalled on spawned workers and on
job apply, restored on drop), re-check stored filter commands in
render_in before source resolution and any cache lookup and in
apply_filter_to_surface, and re-authorize layer.smartFilter.setParam
with the complete merged params after merging stored document params —
including edits whose request contains no path and actions.play steps.
Denial is conservative: the document is unchanged and no worker starts.
Regression tests in the automation crate cover rerender, cached-stage
reuse, merged-param edits, action playback, background jobs, and filter
transfer, for both the displace map and color-lookup LUT vectors.
* Review: carry render authority through Session::edit and surface denials
Session::edit is the engine's universal mutation entry (linked-file
refresh, transforms, place); renders reached through it ran with no
authorization scope installed, so stored smart-filter params read
ambiently there even after the execute-path fix. Install the scope in
Session::edit as well, keyed on the session's authorize gate.
apply_filter_to_surface also conflated a denied filter with an unknown
one: authorize_render_command's Result was discarded with .ok()?, so a
denial silently rendered unfiltered pixels and the stage entered the
process-wide stack cache. The signature becomes
Result<Option<Surface>>: Err for denials (callers propagate), None for
ids that aren't pixel filters (pass-through semantics unchanged), and an
unreadable map keeps its deliberate pass-through. apply_smart_filter and
apply_smart_filters propagate instead of cur.clone().
Two tests: a refresh driven outside command execution (Session::edit)
is denied and leaves the document unchanged; a denied render caches no
stage. Nine authorization tests pass, engine suite and strict Clippy
clean, panic hunt green.
* ui_set description: name the layerFilter field main added
The Layers-panel filter dropdown (#5048) added layerFilter to the
control method's UI_SET_FIELDS without extending the ui_set tool's
description, so the drift guard (mcp.rs, ui_set_description_names_every_
accepted_field) fails on Windows CI for any branch built on main. Name
the field with its patch shape: by (kind, name or mode), kinds (a list
of layer kinds), name (matches part of the layer name), mode (a blend
mode name). The 40 MCP tests pass; strict Clippy clean.
* wasm size: compose and algo join cms at opt-level "z" to clear the 24 MiB gate
Main's browser build is 171 KB over the 24 MiB wasm gate in CI
(25,337,343 bytes), so every branch built on current main fails the
web size check. Demote three pixel crates in the wasm-release profile:
- cms: transforms are built once and cached (transform::cached),
so "z" costs little runtime (measured -42 KB).
- compose: the browser app composites per frame on the GPU (crates/gpu,
WebGPU/WebGL2), so the CPU compositor is not the browser's hot path
there (measured -122 KB).
- algo: filters run per pixel on every interactive filter application.
"z" there is the lever of last resort (measured -331 KB), taken
because the gate is enforced in CI; if browser filters feel too slow,
restore 3 here and reclaim the bytes from a feature-gated decoder.
Final size 24,811,949 bytes locally, 354 KB under the gate; the native
profiles are untouched. If a maintainer prefers keeping algo at 3, the
alternative is raising the self-imposed 24 MiB early-warning limit
(Cloudflare's hard cap is 25 MiB) - a policy call in packaging/web/
package.sh, not a code fix.
* Tests: Free Transform applies when another tool, layer or New Layer is picked (#5665, #5667)
Whole-app regression tests for the paths #5665 and #5667 report (a moved box, and a placed image moved or untouched; toolbar tool click, Layers panel row click, New Layer). They pass on main, so these issues don't reproduce here; the tests pin Photoshop's behaviour (apply, never silently cancel) while the reporters narrow down their environment.
* New Layer while Free Transform is open applies the transform first (History: the transform, then New Layer; a placed layer's step no longer swallows New Layer); tests: a real placed fixture, the New Layer button, History order
* Free Transform + New Layer: check the command can run before applying the transform; a refused transform keeps the box and the placed layer and returns the refusal
* Free Transform + New Layer: Session::precheck validates the call (params shape, New Layer's values, precondition) before the box is applied
* Free Transform + New Layer: commit-first only for PRECHECKED commands; precheck shares dispatch's admission checks (named-layer precondition, jobs, hidden target) and checks a named layer exists; Group from Layers documents its layer param; test: precheck agrees with execute
* Free Transform + New Layer: precheck resolves a layer given by name as the command does; Layer via Copy/Cut leave PRECHECKED (their failures show only when they run); rebased (lib.rs mod lines)
* Shortcuts without the UI: shell commands and tools in the engine, reserved keys, set files
A retro against "everything is drivable headless; the UI is a layer" found the shortcut
work leaning on the UI. Fixed:
- The shell's command table (UI_COMMANDS: View › Zoom In, the panels…) moves into the
engine as data (shell_commands::SHELL_COMMANDS; menus::UI_COMMANDS is now an alias), with
the tools (shell_commands::TOOLS: id, Photoshop name, default key; a test keeps it equal
to the toolbar). edit.keyboardShortcuts binds, lists, imports, exports and checks them for
conflicts with no UI: a headless .kys import takes tool keys, and an export is the whole set.
Tools may share a key (Shift cycling); a command given a tool's key takes it from them.
- Reserved system keys are refused by the engine (the dialog uses the same check) and left
out of imported sets.
- importKys / exportKys / summarize take {"path": …} to read or write files, like
actions.save; the automation gate refuses those paths for MCP.
* Keyboard shortcuts: bindings commit only after their exports are written (a bad target changes nothing); docs: the command knows the tools
* Keyboard shortcuts: both export targets are resolved and checked before either file is written; a failed later write removes the earlier one
* docs: reserved keys per platform
* Keyboard shortcuts export: files are staged next to their targets and moved into place only when all are written; a failure leaves the user's files untouched
* Keyboard shortcut exports all or nothing: unique create_new staging and backups beside each target, restored on any failure; the same target twice refused; docs: Ctrl+Shift+Esc
* Shortcut exports: a failed backup move removes its placeholder; undo restores backups without deleting first and names any it couldn't; read-only targets refused up front
* PHOTOCRAFT_PRESENT_MODE / PHOTOCRAFT_FRAME_LATENCY launch overrides for latency A/B tests (#1343)
Linux traces put the remaining stroke latency in presentation queueing (the app's frame fits
in one refresh; the rest waits on the swap chain), and an NVIDIA user cured it by forcing
vsync off. Testers can now compare without rebuilding: PHOTOCRAFT_PRESENT_MODE (fifo,
mailbox, immediate, autovsync, autonovsync) and PHOTOCRAFT_FRAME_LATENCY (1-3), applied over
the platform default. Bad values are logged and ignored. The active mode is logged at start,
kept in GpuInfo.present and shown in every stroke trace line.
RTX 5090, Vulkan, 120 Hz: default autovsync/latency 1 waits 6 ms per 8.3 ms frame; mailbox
renders every 2.3 ms with 0.2 ms wait, so each vsync shows the newest frame.
* docs: present mode and frame latency overrides
* Present mode overrides never request a mode the driver may lack (Mailbox crashed on X11/NVIDIA)
* Preferences › Performance › Low-Latency Drawing, on by default: no vsync wait while drawing (#1343)
Testers' traces put the ~50 ms stroke lag in the swap chain's vsync wait. With vsync off it
disappears (X11/NVIDIA; Windows/NVIDIA via the driver), and a shorter queue doesn't help. The
surface now presents with AutoNoVsync by default: wgpu picks Immediate, else Mailbox, else
Fifo by what the driver offers, so no unsupported mode is ever requested. Turning the
preference off brings vsync back (smooth, but late). It applies at the next launch, like the
other graphics settings. Where only Immediate exists the canvas may tear while it changes;
frames render as fast as input arrives during a stroke and stop when idle.
* Runtime and launch controls: ui.set strokeTrace, --stroke-trace, --present-mode, --frame-latency
The retro's remaining gap: the stroke trace and the presentation overrides were env-only. The trace now switches at runtime over the control channel (ui.set {"strokeTrace": true|false|"stderr"|path}) and with --stroke-trace; --present-mode / --frame-latency sit beside their env vars and win over them. Bad values are reported and ignored; none fails a launch.
* docs: strokeTrace beside the other ui.set fields
* docs: regenerate the scorecard after rebasing onto main
* Low-Latency Drawing: frames capped at 240 fps without vsync (after #4472); ui.set strokeTrace takes on/off only (a CSV path is launch-only); tearing documented as usual (Immediate); pref hidden on web; doc comment order
* stroke_trace: from_value doc says ui.set passes only on or off
* Low-Latency Drawing ships off (opt-in), per the owner's decision
* docs: present mode default is vsync (Low-Latency Drawing ships off)
Keep the full workspace and feature build selection while running the craft-fonts pass only for text, io and ui-egui. Ordinary workspace test coverage remains in the first test pass.
Run workspace tests with pinned cargo-nextest, a CI fail-fast profile and a ten-minute per-test hang limit. Keep cheap checks first and doctests in a separate step so failures propagate on Windows.
* Layers panel: blend mode, Opacity, Fill and locks change every selected layer; Photoshop's History names
With several layers selected, the Layers panel's blend mode, Opacity, Fill and lock controls
changed only the active layer, silently (the layers audit's top panel gap). They now change
every selected layer, as in Photoshop, in one undo step. layer.setProps takes
"layers": [id, …] for that (headless too); a list with an unknown id, a bad shape, or a
name for several layers is an error that changes nothing.
History names follow Photoshop for a single kind of change: Opacity Change, Fill Opacity
Change, Blending Change, Rename Layer (a mix stays Layer Properties).
* Layers panel with several selected: Pass Through only for groups (refused for other layers), the Basic lock toggle and the blend-mode hover preview take every selected layer, the controls grey with the Background selected; History icon for opacity/fill/blend changes
* Layers panel: leave the Background-in-a-selection behaviour as it was until Photoshop's is confirmed (#5856)
* Layers panel: a selected Background is left out of several-layer Opacity, Fill, blend and lock changes until Photoshop's behaviour is confirmed (#5856)
* Keyboard Shortcuts: Accept and Go to Conflict / Undo Changes, and reserved keys (K4a)
A key that another command holds now offers Photoshop's two answers under the warning.
Accept and Go to Conflict takes the key from the other command at once (its main key or
an alternate) and selects it for a new one. Undo Changes puts the edited command's main and
further keys back as they were. OK without either still accepts, as in Photoshop.
Keys the system takes before the app sees them are refused with a message: Alt+Tab,
Alt+F4, Ctrl+Alt+Delete, Ctrl+Esc and Ctrl+Shift+Esc off the Mac; Cmd+Tab, Cmd+` and
Cmd+Opt+Esc on it.
* Keyboard Shortcuts: a conflict goes stale when the bindings change after it (Reset All, Use Default, Delete, ×, Escape, a set switch); Accept takes the key from every holder; Linux reserves its own keys; doc comment back on shortcut_text
* Paste Layer Style carries Blend If and channels; align to an active selection with several layers
From the layers audit (L2):
- Copy / Paste Layer Style copied the effects, blend mode, Fill and advanced blending but
dropped Blend If and the channel restrictions; Photoshop's layer style carries all of its
Blending Options. Across colour modes only Blend If's Gray range carries over and the
channels reset (RGB channel ranges mean nothing in CMYK). The copied style is now a named
struct (CopiedStyle) instead of a tuple.
- With an active selection, aligning (Layer › Align, "auto") aligned two or more layers to
their own bounds; Photoshop aligns them to the selection (Layer › Align Layers to
Selection), as it already did for one layer.
* Paste Layer Style keeps Blend If normalised across modes; Align docs say it follows an active selection for any count; undo, group and 16-bit tests
* Clip groups: Release Clipping Mask takes the layers above; merges keep clipping; Delete Hidden settles
From the layers audit (L2, silently wrong results):
- Release Clipping Mask on a clipped layer also releases the clipped layers above it, as
Adobe documents ("removes the selected layer and any layers above it from the clipping
mask"). Before, they stayed clipped to the released layer.
- Merge Down of two clipped layers, and Merge Layers of clipped layers of one chain, give a
layer still clipped to their base (it was always unclipped). Merge Down's result also keeps
the lower layer's colour label and link; Merge Layers' keeps the top layer's label.
- Delete Hidden Layers releases the layers clipped to a deleted hidden base (clip_group rule
3, as single Delete already does), instead of leaving them clipped to the layer below.
* Merge Layers keeps the result clipped only when the clipped layers share one clip chain
* fmt
* Check deep EXR tile count tables against max_alloc
A deep tile's sample-count table was decompressed at the header's full tile size with no
budget check, so a 20 KB file could build a 4 MiB table and a scratch copy under a 1 MiB
max_alloc. Try that size only when max_alloc has room for both; otherwise read the clipped size.
Fixes#4958.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Keep deep EXR decompression within the size it was admitted at
The count-table check admits the table and its scratch copy, 2x the table, but Vec doubling
gave a 4 MiB ZIPS table 8 MiB of capacity (a 3 MiB RLE table 4 MiB), so an admitted table
and its copy could take 3x. Grow the decompressed buffer by doubling capped at the expected
size, inflating straight into it, and allocate it and the scratch copy fallibly.
Tests: decompressed buffers never reserve past the expected size; a near-maximal tile with a
clipped table decodes under every budget; per-tile admission; half and uint channels after
full-size tables; the 8 MiB boundary for uncompressed tables too.
Refs #4958.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Tests: show deep tile tables are admitted one tile at a time
The 17x17 image in 16x16 tiles had an interior tile, which never reaches the max_alloc check,
so only three edge tables (6 KiB) were checked against the 8000-byte budget and a check that
summed admissions across tiles still passed. Use a 48x1 image instead: its three tiles are all
edge tiles, and under 4096 bytes a summed check fails at the third tile.
Refs #4958.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Complete deep EXR count-table allocation admission
Account for scratch space on clipped and interior compressed tables while preserving raw-table budgets. Borrow packed input before admission and make initial RLE reservations and raw copies fallible.
Add single-channel admission boundaries, raw fallback controls, and hostile-size borrowed-read regressions.
* ci: run deep EXR regressions independently
* Drop the separate deep EXR CI job
It was added so these tests could run while main's workspace suite was
failing. Main is green again, and the regular test jobs run them.
This reverts commit 5aa4869b45.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: jonathnwxh <[email protected]>
The only limit on an unauthenticated control connection was the 30 s
idle timeout, which starts again with every read. A local client that
kept sending bytes or blank lines without ever completing `auth` kept
its connection, and one of the 16 connection permits, for as long as it
liked; sixteen of them locked every other client out.
AuthDeadlineStream wraps the read half of a connection: until the
connection authenticates, reads fail once AUTH_DEADLINE (10 s) has
passed since it connected, and a blocking read waits no longer than
what is left. Both transports use it (the desktop app's control server
and the headless `photocraft-cli serve`). After `auth` succeeds only
the idle timeout applies, as before.
Co-authored-by: Claude Opus 5.5 <[email protected]>
ruzstd before 0.9.1 lets a crafted zstd input under 128 KB allocate
several gigabytes at once. PhotoCraft decodes zstd when it opens .pcraft
and Affinity files, so such a file could end the process with an
out-of-memory abort.
Raise the minimum in both crates that use it and update the lockfile
from 0.9.0 to 0.9.1.
Co-authored-by: Claude Opus 5.5 <[email protected]>
While a path was being drawn, the Pen's Auto Add/Delete was off, so adding
a point meant finishing the path and switching to Add Anchor Point. Now a
click on a segment of the path in progress adds an anchor, a click on a
middle anchor deletes it, and drawing carries on from the last anchor. The
first anchor still closes the path and the last still breaks its handle.
The hover highlight draws over the path in progress, and the rubber band
hides where a click adds or deletes instead of placing an anchor.
Claude-Session: https://claude.ai/code/session_01JFTxZ57wCmWEDXxCn1Zuni
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Quick Mask: ⌘I inverts the whole mask, unstored black included (#6095)
Image › Adjustments on an alpha channel or the Quick Mask ran on the surface's
stored content only. The black outside a selection-made Quick Mask (or a new
black alpha channel) is the surface default, never stored, so Invert left it
black: the whole mask turned red and leaving Quick Mask selected nothing.
Channel targets now go through `pixels::adjust_mask`, as layer masks do, which
adjusts the default pixel too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Quick Mask: test ⌘I with Color Indicates Selected Areas too (#6095)
A rectangle comes back as everything but the rectangle, and with nothing
selected ⌘I selects everything. Both failed on main.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Run the existing Clippy, dependency layering and scorecard checks before the test passes so failures are reported earlier.
Preserve the existing commands, test coverage and runtime limits. Linux, Windows, corpus, WASM, web and packaging checks passed.
Choose PSB block length width from the canonical big-endian key before converting the key for little-endian TIFF output. Preserve foreign payloads, padding and conversion of following blocks.
Addresses #5599, finding 2 only.
Fixes#6104.
Keep effect-identity memo lookup, pure hashing and insertion atomic across concurrent render callers. Preserve existing identities and add a contention regression with eight synchronized callers.
The pass reran every test in the workspace, but CRAFT_FONTS is only read in photocraft-text,
photocraft-io (svg) and photocraft-ui-egui. The build selection stays the same (the disk warning
above still applies); a nextest filter runs those three packages' tests: ~3,800 of ~8,900.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
cargo test runs the ~258 test binaries one at a time; nextest runs them in parallel. Its new `ci`
profile stops at the first failure (a red PR frees its runner early), fails a test that hangs for
10 minutes instead of letting it hold the runner, and repeats every failure at the end of the log;
the GPU test group still applies. Same --workspace --locked --features heif selection, so the build
is unchanged. Doctests stay on cargo test --doc.
postscript_only_style_shows_actual_subfamily relied on another test in the same process having
started the font scan; under nextest it runs alone, so it starts the scan itself (as #2933 does).
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
They take about 1.5 min together but ran after all the test passes, so a clippy failure (15 s)
showed red only 41-48 min into the job. Same steps, reordered: no cost on a green run.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Fix two ui-egui tests broken on main by concurrent landings
- profile_open_tests: #5614 renamed the built-in profile to "(PhotoCraft)" after #5208
hard-coded "(Photocraft)" in the Use-embedded button label.
- smart_object_curves_preview_keeps_its_size_on_a_reduced_proxy (#5779): the reduced
square is in the right place, but its corner pixels are 220/255 after the smart
object's resampling; assert full coverage inside and over half at the corners.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Web: ship the wasm as compressed parts behind a boot module (#5302)
The release site held the app's wasm as one file, which sits at the 25 MiB
per-file cap of Cloudflare Pages, so the 24 MiB gate failed unrelated PRs.
packaging/web/package.sh now gzips the wasm and cuts it into parts of at
most 20 MiB under app/, and ships the page of a new crate,
apps/photocraft-web-boot. That module fetches the parts in order, has the
browser inflate them (DecompressionStream) and passes the stream to the
app's own wasm-bindgen loader, so the browser still compiles while it
downloads. A failed download is retried twice, then the page says so.
The check is now that no file in the site is over 24 MiB. trunk build in
apps/photocraft-web and the Docker image still give one .wasm.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Web boot: inflate the parts in the module where the browser can't (#5302)
* Tests: the adjustment-layer drag preview test follows #5778 (#5926)
#5778 caps a live adjustment-layer preview by the view's zoom, so at 100 %
the drag shows full resolution even with Low Resolution Previews on. The
test still expected the reduced copy there. It now checks that at 100 % every
column shows by default, and moves the low-res on/off comparison to 50 %,
where the reduced copy is still used: on, the drag shows only the (inverted)
black columns; off, it matches the committed view.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* TIFF: read and write CIE L*a*b* (CIELab, ICCLab); Lab documents stay Lab
Photoshop saves Lab documents as CIELab TIFFs (PhotometricInterpretation
8, a* and b* signed), which PhotoCraft could not open, and PhotoCraft
saved Lab documents as sRGB TIFFs.
- codecs: a Lab channel layout (Lab, LabA) holding a* and b* as
PhotoCraft's Lab documents do ((a + 128) / 255; 16 bits like the PSD
importer), with its own Lab <-> sRGB conversion (same formulas as
photocraft-color; codecs stays standalone) for formats without Lab,
which now warn LabConverted.
- TIFF read: CIELab and ICCLab (8 and 16 bits, alpha) are handed to the
tiff crate as RGB and labelled Lab; a* and b* are converted from the
file's encoding (signed a / 256a, or offset).
- TIFF write: Lab and LabA as CIELab, as Photoshop writes it; float Lab
is written with 16 bits (DepthReduced).
- io: a Lab TIFF opens as a Lab document; a Lab document with one plain
layer is written to TIFF in Lab (other formats still use sRGB).
Photoshop 2025's Lab TIFFs open with the colours Photoshop shows (its
own RGB conversion matches within one level at 8 bits and exactly at 16
bits), and Photoshop reads PhotoCraft's Lab TIFFs back with the same
L*a*b* values.
Fixes#4936
* PSD: a smart object's filter mask shows when the file opens
Photoshop stores a smart object's layer pixels with its filters applied
but not its filter mask. When it composites, it mixes the unfiltered
pixels (cached in FEid) back in through the mask. We showed the stored
pixels as they were, so a PSD whose smart filter has a mask opened with
the filter everywhere. Only after something re-rendered the smart object
did the mask appear, because our re-render mixes it.
- compose::filter_mask holds the mix, moved from the engine, and its
inverse.
- Import decodes the unfiltered pixels from FEid
(smart_map::unfiltered_from_item) and keeps the mixed result as the
smart object's pixels, as the engine's re-render does.
- Export writes the layer record the way Photoshop does: unmixed with
the unfiltered pixels it writes (or keeps) in FEid, so Photoshop mixes
back to what the document showed. Transparent pixels stay empty.
Photoshop oracles: sf-filter-mask-hard, sf-filter-mask-gradient and
sf-stack-blend-and-mask pass (145 -> 148 of 258); floor 145 -> 148.
psd-tools and io corpora unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* DNG: decode lossy (JPEG-compressed) DNG tiles (FILE-215-11)
DNG 1.4 lossy JPEG (compression 34892) stores each tile or strip as a
baseline-DCT JPEG of 8-bit samples: a gray stream for CFA data, a YCbCr
stream for three-sample LinearRaw. The decoder arm from #2021 read the
streams but widened every sample by 257, so the values no longer matched
the file's levels: an 8-bit WhiteLevel (or its default, 255) clipped
everything to white and a 256-entry LinearizationTable was indexed past
its end. It also required the JPEG frame to match the clipped row count,
which rejected the padded edge tiles real tiled files carry.
The arm is now `lossy_jpeg()` in sensor.rs. The decoded 8-bit values are
the stored samples, and BlackLevel, WhiteLevel (default 255 for 8-bit
data) and the LinearizationTable apply to them exactly as for every other
compression. A tile may be coded at the full tile size (edge tiles padded,
as the spec says) or at the image's row count; a strip at its row count;
the width must be the segment's. BitsPerSample must be 8 and the JPEG's
component count must equal SamplesPerPixel (1 or 3); anything else is an
error, never a silent conversion. The frame header is checked before any
pixel is decoded and the JPEG decoder's own size cap is the segment the
IFD declares (already charged against the limits), so a hostile SOF cannot
size a buffer past the tile.
testgen writes true 8-bit lossy fixtures (`LossyJpegTiles`, and strips via
the new `LossyJpegStrips`), accepts pre-encoded `lossy_segments` for
hostile-input tests, exposes `baseline_jpeg()`, and leaves WhiteLevel out
when `white` is 0. The DNG tests cover CFA and LinearRaw RGB from tiles
and strips at 37x23 (partial edge tiles, a short last strip), the default
white level, black/white scaling and a LinearizationTable against the
uncompressed path, and hostile inputs: a truncated tile, frame-size and
component-count mismatches, a 65535x65535 SOF under default and tight
limits, a tile over the allocation budget, and 12/16-bit or 4-sample
files. The hostile mutation set gains a lossy CFA and a lossy LinearRaw
file. The parity doc's DNG row is updated; the scorecard row is left for
the coordinating branch.
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* File > Save As / Save a Copy / CLI can write PDF: one flattened page with the embedded profile (#3401)
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* Vector: bound flattening memory, keep over-budget dashes, cheaper edge build (#5195)
- flatten: a path's flattened points are capped at 1 << 20; past it every
curve keeps a proportional share of its pieces (at least one), so hostile
handles from files (vmsk, SVG, effect outlines) cannot exhaust memory.
fill_rasterizer flattens the whole path under one budget.
- stroke: past the dash step budget (or below coordinate precision),
dash_polyline keeps the dashes made so far and draws the rest solid
instead of dropping the stroke. Open polylines are no longer copied.
- raster: coordinates are clamped to +-1e15 and edges with a non-finite
slope (vanishingly short) are skipped, so huge finite coordinates no
longer produce NaN coverage. fill_rasterizer sorts edges once instead of
after every component (stable sort, identical order).
- Cleanups: dead miter `half`, dead loop-length check in trace_mask,
shared mask render body, map_knots helper, shared straight-segment test.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: cap pattern memory across a parse, reject mixed-depth pattern channels (#5198)
- Patterns: one PatternBudget (decoded bytes + zero padding) is shared by
every pattern of a block, a .pat file and the pattern blocks of one
document; .abr patt sections are charged to the samples' budget.
- Patterns: a channel whose depth differs from the first is rejected (the
parsed pattern could not be written back).
- Slices: descriptor numbers no longer wrap; out-of-range ids are no id,
coordinates clamp to i32, and a v6 trailing entry without a usable id
attaches its outsets to no slice.
- LayerInfo::unpadded_len is computed from field sizes instead of
serializing every record into a scratch buffer.
- abr desc presets are moved out instead of cloned; UnFl and shmd counts
use the bounded preallocation; ImageData::decode_channel (decoded every
plane to return one, unused outside its test) is removed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Vector: bound stroke outline points (#5211)
- stroke_polygons dashes all lines first, then counts round caps and
joins. When their full-resolution tessellation would exceed 1 << 20
outline points, every round cap and join gets the same smaller segment
count (down to 4 per circle, 1 per join; area-exact radius kept).
Strokes under the budget are unchanged; no piece is dropped.
- The dash step budget (100000) is now shared by all lines of a stroke
instead of per polyline, so many subpaths cannot multiply the dash
(and cap) count. Past it the rest is drawn solid, as before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: timothydaly1 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Fabio Lissi <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Finn Newick <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* TIFF: read and write CIE L*a*b* (CIELab, ICCLab); Lab documents stay Lab
Photoshop saves Lab documents as CIELab TIFFs (PhotometricInterpretation
8, a* and b* signed), which PhotoCraft could not open, and PhotoCraft
saved Lab documents as sRGB TIFFs.
- codecs: a Lab channel layout (Lab, LabA) holding a* and b* as
PhotoCraft's Lab documents do ((a + 128) / 255; 16 bits like the PSD
importer), with its own Lab <-> sRGB conversion (same formulas as
photocraft-color; codecs stays standalone) for formats without Lab,
which now warn LabConverted.
- TIFF read: CIELab and ICCLab (8 and 16 bits, alpha) are handed to the
tiff crate as RGB and labelled Lab; a* and b* are converted from the
file's encoding (signed a / 256a, or offset).
- TIFF write: Lab and LabA as CIELab, as Photoshop writes it; float Lab
is written with 16 bits (DepthReduced).
- io: a Lab TIFF opens as a Lab document; a Lab document with one plain
layer is written to TIFF in Lab (other formats still use sRGB).
Photoshop 2025's Lab TIFFs open with the colours Photoshop shows (its
own RGB conversion matches within one level at 8 bits and exactly at 16
bits), and Photoshop reads PhotoCraft's Lab TIFFs back with the same
L*a*b* values.
Fixes#4936
* PSD: a smart object's filter mask shows when the file opens
Photoshop stores a smart object's layer pixels with its filters applied
but not its filter mask. When it composites, it mixes the unfiltered
pixels (cached in FEid) back in through the mask. We showed the stored
pixels as they were, so a PSD whose smart filter has a mask opened with
the filter everywhere. Only after something re-rendered the smart object
did the mask appear, because our re-render mixes it.
- compose::filter_mask holds the mix, moved from the engine, and its
inverse.
- Import decodes the unfiltered pixels from FEid
(smart_map::unfiltered_from_item) and keeps the mixed result as the
smart object's pixels, as the engine's re-render does.
- Export writes the layer record the way Photoshop does: unmixed with
the unfiltered pixels it writes (or keeps) in FEid, so Photoshop mixes
back to what the document showed. Transparent pixels stay empty.
Photoshop oracles: sf-filter-mask-hard, sf-filter-mask-gradient and
sf-stack-blend-and-mask pass (145 -> 148 of 258); floor 145 -> 148.
psd-tools and io corpora unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* DNG: decode lossy (JPEG-compressed) DNG tiles (FILE-215-11)
DNG 1.4 lossy JPEG (compression 34892) stores each tile or strip as a
baseline-DCT JPEG of 8-bit samples: a gray stream for CFA data, a YCbCr
stream for three-sample LinearRaw. The decoder arm from #2021 read the
streams but widened every sample by 257, so the values no longer matched
the file's levels: an 8-bit WhiteLevel (or its default, 255) clipped
everything to white and a 256-entry LinearizationTable was indexed past
its end. It also required the JPEG frame to match the clipped row count,
which rejected the padded edge tiles real tiled files carry.
The arm is now `lossy_jpeg()` in sensor.rs. The decoded 8-bit values are
the stored samples, and BlackLevel, WhiteLevel (default 255 for 8-bit
data) and the LinearizationTable apply to them exactly as for every other
compression. A tile may be coded at the full tile size (edge tiles padded,
as the spec says) or at the image's row count; a strip at its row count;
the width must be the segment's. BitsPerSample must be 8 and the JPEG's
component count must equal SamplesPerPixel (1 or 3); anything else is an
error, never a silent conversion. The frame header is checked before any
pixel is decoded and the JPEG decoder's own size cap is the segment the
IFD declares (already charged against the limits), so a hostile SOF cannot
size a buffer past the tile.
testgen writes true 8-bit lossy fixtures (`LossyJpegTiles`, and strips via
the new `LossyJpegStrips`), accepts pre-encoded `lossy_segments` for
hostile-input tests, exposes `baseline_jpeg()`, and leaves WhiteLevel out
when `white` is 0. The DNG tests cover CFA and LinearRaw RGB from tiles
and strips at 37x23 (partial edge tiles, a short last strip), the default
white level, black/white scaling and a LinearizationTable against the
uncompressed path, and hostile inputs: a truncated tile, frame-size and
component-count mismatches, a 65535x65535 SOF under default and tight
limits, a tile over the allocation budget, and 12/16-bit or 4-sample
files. The hostile mutation set gains a lossy CFA and a lossy LinearRaw
file. The parity doc's DNG row is updated; the scorecard row is left for
the coordinating branch.
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* Vector: bound flattening memory, keep over-budget dashes, cheaper edge build (#5195)
- flatten: a path's flattened points are capped at 1 << 20; past it every
curve keeps a proportional share of its pieces (at least one), so hostile
handles from files (vmsk, SVG, effect outlines) cannot exhaust memory.
fill_rasterizer flattens the whole path under one budget.
- stroke: past the dash step budget (or below coordinate precision),
dash_polyline keeps the dashes made so far and draws the rest solid
instead of dropping the stroke. Open polylines are no longer copied.
- raster: coordinates are clamped to +-1e15 and edges with a non-finite
slope (vanishingly short) are skipped, so huge finite coordinates no
longer produce NaN coverage. fill_rasterizer sorts edges once instead of
after every component (stable sort, identical order).
- Cleanups: dead miter `half`, dead loop-length check in trace_mask,
shared mask render body, map_knots helper, shared straight-segment test.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: cap pattern memory across a parse, reject mixed-depth pattern channels (#5198)
- Patterns: one PatternBudget (decoded bytes + zero padding) is shared by
every pattern of a block, a .pat file and the pattern blocks of one
document; .abr patt sections are charged to the samples' budget.
- Patterns: a channel whose depth differs from the first is rejected (the
parsed pattern could not be written back).
- Slices: descriptor numbers no longer wrap; out-of-range ids are no id,
coordinates clamp to i32, and a v6 trailing entry without a usable id
attaches its outsets to no slice.
- LayerInfo::unpadded_len is computed from field sizes instead of
serializing every record into a scratch buffer.
- abr desc presets are moved out instead of cloned; UnFl and shmd counts
use the bounded preallocation; ImageData::decode_channel (decoded every
plane to return one, unused outside its test) is removed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Vector: bound stroke outline points (#5211)
- stroke_polygons dashes all lines first, then counts round caps and
joins. When their full-resolution tessellation would exceed 1 << 20
outline points, every round cap and join gets the same smaller segment
count (down to 4 per circle, 1 per join; area-exact radius kept).
Strokes under the budget are unchanged; no piece is dropped.
- The dash step budget (100000) is now shared by all lines of a stroke
instead of per polyline, so many subpaths cannot multiply the dash
(and cap) count. Past it the rest is drawn solid, as before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Fabio Lissi <[email protected]>
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Finn Newick <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* TIFF: read and write CIE L*a*b* (CIELab, ICCLab); Lab documents stay Lab
Photoshop saves Lab documents as CIELab TIFFs (PhotometricInterpretation
8, a* and b* signed), which PhotoCraft could not open, and PhotoCraft
saved Lab documents as sRGB TIFFs.
- codecs: a Lab channel layout (Lab, LabA) holding a* and b* as
PhotoCraft's Lab documents do ((a + 128) / 255; 16 bits like the PSD
importer), with its own Lab <-> sRGB conversion (same formulas as
photocraft-color; codecs stays standalone) for formats without Lab,
which now warn LabConverted.
- TIFF read: CIELab and ICCLab (8 and 16 bits, alpha) are handed to the
tiff crate as RGB and labelled Lab; a* and b* are converted from the
file's encoding (signed a / 256a, or offset).
- TIFF write: Lab and LabA as CIELab, as Photoshop writes it; float Lab
is written with 16 bits (DepthReduced).
- io: a Lab TIFF opens as a Lab document; a Lab document with one plain
layer is written to TIFF in Lab (other formats still use sRGB).
Photoshop 2025's Lab TIFFs open with the colours Photoshop shows (its
own RGB conversion matches within one level at 8 bits and exactly at 16
bits), and Photoshop reads PhotoCraft's Lab TIFFs back with the same
L*a*b* values.
Fixes#4936
* DNG: decode lossy (JPEG-compressed) DNG tiles (FILE-215-11)
DNG 1.4 lossy JPEG (compression 34892) stores each tile or strip as a
baseline-DCT JPEG of 8-bit samples: a gray stream for CFA data, a YCbCr
stream for three-sample LinearRaw. The decoder arm from #2021 read the
streams but widened every sample by 257, so the values no longer matched
the file's levels: an 8-bit WhiteLevel (or its default, 255) clipped
everything to white and a 256-entry LinearizationTable was indexed past
its end. It also required the JPEG frame to match the clipped row count,
which rejected the padded edge tiles real tiled files carry.
The arm is now `lossy_jpeg()` in sensor.rs. The decoded 8-bit values are
the stored samples, and BlackLevel, WhiteLevel (default 255 for 8-bit
data) and the LinearizationTable apply to them exactly as for every other
compression. A tile may be coded at the full tile size (edge tiles padded,
as the spec says) or at the image's row count; a strip at its row count;
the width must be the segment's. BitsPerSample must be 8 and the JPEG's
component count must equal SamplesPerPixel (1 or 3); anything else is an
error, never a silent conversion. The frame header is checked before any
pixel is decoded and the JPEG decoder's own size cap is the segment the
IFD declares (already charged against the limits), so a hostile SOF cannot
size a buffer past the tile.
testgen writes true 8-bit lossy fixtures (`LossyJpegTiles`, and strips via
the new `LossyJpegStrips`), accepts pre-encoded `lossy_segments` for
hostile-input tests, exposes `baseline_jpeg()`, and leaves WhiteLevel out
when `white` is 0. The DNG tests cover CFA and LinearRaw RGB from tiles
and strips at 37x23 (partial edge tiles, a short last strip), the default
white level, black/white scaling and a LinearizationTable against the
uncompressed path, and hostile inputs: a truncated tile, frame-size and
component-count mismatches, a 65535x65535 SOF under default and tight
limits, a tile over the allocation budget, and 12/16-bit or 4-sample
files. The hostile mutation set gains a lossy CFA and a lossy LinearRaw
file. The parity doc's DNG row is updated; the scorecard row is left for
the coordinating branch.
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* Vector: bound flattening memory, keep over-budget dashes, cheaper edge build (#5195)
- flatten: a path's flattened points are capped at 1 << 20; past it every
curve keeps a proportional share of its pieces (at least one), so hostile
handles from files (vmsk, SVG, effect outlines) cannot exhaust memory.
fill_rasterizer flattens the whole path under one budget.
- stroke: past the dash step budget (or below coordinate precision),
dash_polyline keeps the dashes made so far and draws the rest solid
instead of dropping the stroke. Open polylines are no longer copied.
- raster: coordinates are clamped to +-1e15 and edges with a non-finite
slope (vanishingly short) are skipped, so huge finite coordinates no
longer produce NaN coverage. fill_rasterizer sorts edges once instead of
after every component (stable sort, identical order).
- Cleanups: dead miter `half`, dead loop-length check in trace_mask,
shared mask render body, map_knots helper, shared straight-segment test.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: cap pattern memory across a parse, reject mixed-depth pattern channels (#5198)
- Patterns: one PatternBudget (decoded bytes + zero padding) is shared by
every pattern of a block, a .pat file and the pattern blocks of one
document; .abr patt sections are charged to the samples' budget.
- Patterns: a channel whose depth differs from the first is rejected (the
parsed pattern could not be written back).
- Slices: descriptor numbers no longer wrap; out-of-range ids are no id,
coordinates clamp to i32, and a v6 trailing entry without a usable id
attaches its outsets to no slice.
- LayerInfo::unpadded_len is computed from field sizes instead of
serializing every record into a scratch buffer.
- abr desc presets are moved out instead of cloned; UnFl and shmd counts
use the bounded preallocation; ImageData::decode_channel (decoded every
plane to return one, unused outside its test) is removed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Vector: bound stroke outline points (#5211)
- stroke_polygons dashes all lines first, then counts round caps and
joins. When their full-resolution tessellation would exceed 1 << 20
outline points, every round cap and join gets the same smaller segment
count (down to 4 per circle, 1 per join; area-exact radius kept).
Strokes under the budget are unchanged; no piece is dropped.
- The dash step budget (100000) is now shared by all lines of a stroke
instead of per polyline, so many subpaths cannot multiply the dash
(and cap) count. Past it the rest is drawn solid, as before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Finn Newick <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* TIFF: read and write CIE L*a*b* (CIELab, ICCLab); Lab documents stay Lab
Photoshop saves Lab documents as CIELab TIFFs (PhotometricInterpretation
8, a* and b* signed), which PhotoCraft could not open, and PhotoCraft
saved Lab documents as sRGB TIFFs.
- codecs: a Lab channel layout (Lab, LabA) holding a* and b* as
PhotoCraft's Lab documents do ((a + 128) / 255; 16 bits like the PSD
importer), with its own Lab <-> sRGB conversion (same formulas as
photocraft-color; codecs stays standalone) for formats without Lab,
which now warn LabConverted.
- TIFF read: CIELab and ICCLab (8 and 16 bits, alpha) are handed to the
tiff crate as RGB and labelled Lab; a* and b* are converted from the
file's encoding (signed a / 256a, or offset).
- TIFF write: Lab and LabA as CIELab, as Photoshop writes it; float Lab
is written with 16 bits (DepthReduced).
- io: a Lab TIFF opens as a Lab document; a Lab document with one plain
layer is written to TIFF in Lab (other formats still use sRGB).
Photoshop 2025's Lab TIFFs open with the colours Photoshop shows (its
own RGB conversion matches within one level at 8 bits and exactly at 16
bits), and Photoshop reads PhotoCraft's Lab TIFFs back with the same
L*a*b* values.
Fixes#4936
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: cap pattern memory across a parse, reject mixed-depth pattern channels (#5198)
- Patterns: one PatternBudget (decoded bytes + zero padding) is shared by
every pattern of a block, a .pat file and the pattern blocks of one
document; .abr patt sections are charged to the samples' budget.
- Patterns: a channel whose depth differs from the first is rejected (the
parsed pattern could not be written back).
- Slices: descriptor numbers no longer wrap; out-of-range ids are no id,
coordinates clamp to i32, and a v6 trailing entry without a usable id
attaches its outsets to no slice.
- LayerInfo::unpadded_len is computed from field sizes instead of
serializing every record into a scratch buffer.
- abr desc presets are moved out instead of cloned; UnFl and shmd counts
use the bounded preallocation; ImageData::decode_channel (decoded every
plane to return one, unused outside its test) is removed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Finn Newick <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* TIFF: read and write CIE L*a*b* (CIELab, ICCLab); Lab documents stay Lab
Photoshop saves Lab documents as CIELab TIFFs (PhotometricInterpretation
8, a* and b* signed), which PhotoCraft could not open, and PhotoCraft
saved Lab documents as sRGB TIFFs.
- codecs: a Lab channel layout (Lab, LabA) holding a* and b* as
PhotoCraft's Lab documents do ((a + 128) / 255; 16 bits like the PSD
importer), with its own Lab <-> sRGB conversion (same formulas as
photocraft-color; codecs stays standalone) for formats without Lab,
which now warn LabConverted.
- TIFF read: CIELab and ICCLab (8 and 16 bits, alpha) are handed to the
tiff crate as RGB and labelled Lab; a* and b* are converted from the
file's encoding (signed a / 256a, or offset).
- TIFF write: Lab and LabA as CIELab, as Photoshop writes it; float Lab
is written with 16 bits (DepthReduced).
- io: a Lab TIFF opens as a Lab document; a Lab document with one plain
layer is written to TIFF in Lab (other formats still use sRGB).
Photoshop 2025's Lab TIFFs open with the colours Photoshop shows (its
own RGB conversion matches within one level at 8 bits and exactly at 16
bits), and Photoshop reads PhotoCraft's Lab TIFFs back with the same
L*a*b* values.
Fixes#4936
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Finn Newick <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* perf(compose,gpu): effect caches reserve tile addresses instead of pinning layers
Effect map cache entries kept a clone of their layer (a deep clone for groups) so the tile addresses the cache key hashes couldn't be reused. That kept an edited or deleted layer's pixels alive, uncounted, until the entry was evicted. Keep a Weak to each hashed tile instead: it reserves the address without keeping the pixels. The GPU compositor's styled-group pins and shape-source tile snapshots do the same. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(history): count snapshot-only tiles against the history budget
History panel snapshots outlived every budget: a Merged Layers snapshot is a full flattened copy that neither the history budget, the GPU budget's document bytes nor Purge accounted for. Trim now counts the tiles only snapshots hold as fixed cost, the GPU budget includes them, and Purge > Histories no longer reports tiles a kept snapshot still holds as freed. Part of #4530.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Diego Chagas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Select Subject: keep subject parts coloured like the backdrop, leave enclosed gaps open
Users on #2788 reported Select Subject dropping clothes that are close in
colour to a plain wall and bridging the gap between an arm and the torso.
Both come from the same place: the saliency pass labels low-contrast parts
of the subject as background, the background colour model then learns
them and explains them away, and the final blind hole fill closes every
enclosed gap up to 5 % of the mask.
Two changes in crates/algo:
- A second GrabCut pass whose background model only knows the backdrop. A
colour model fitted to the image frame (minus a margin around the
salient blob, in case the subject runs out of the picture) marks every
pixel it explains as probable background; the blob and whatever the
backdrop cannot account for start as probable foreground. Only
components that touch the first pass's result survive.
- Hole filling that consults the colour models: specks (<= 0.2 % of the
mask) always fill, larger holes up to 5 % fill only when their pixels
fit the foreground model better than the background's.
Tests: two regressions (a cream-shouldered figure on a near-white wall
keeps its shoulders; a ring keeps its backdrop-coloured hole open), plus
ignored timing and picture-dump helpers. The 6 MP first run goes from
about 100 ms to about 200 ms on one core; the P29 repeat path is untouched.
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Gert Pellin <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Content-Aware Fill: measure a thin hole by its thickness, not its bounds (#1366)
Removing long, thin objects (coiled guitar strings, wires) left their
outline as darker and lighter blotches and copied pieces of other strings
into the fill. The pyramid shrank until the hole's bounding box was four
patches across: for a string across the image that is a level where the
hole has swallowed its surroundings, so the matches found there copy
whatever looks like the blur, and the finer levels (refine-only for wide
bounds) keep them.
A hole at most a quarter as thick as its bounds are long now counts as
thin: the pyramid stops while it is still about two patches thick, and
that level starts from a membrane across it (over a few pixels, the
surroundings themselves) instead of the onion peel, whose exhaustive
search is what made a shallow pyramid unaffordable. Compact holes are
unchanged bit for bit.
Refs #1366
* Content-Aware Fill: every part of a thin selection keeps its surroundings (#1366)
Two more causes of ghosts when removing strings and the small knobs
selected with them:
* Auto sampling summarises the selection's surroundings by a few
clusters. Around a long selection those come from the strings'
surroundings, so a small part elsewhere (a tuning knob in a differently
coloured corner) lost its own surroundings and was filled from far away
in another colour. Around a thin selection the Auto area now always
reaches as far as the selection is thick.
* The membrane that starts a thin hole now starts a patch radius out, past
the anti-aliased fringe or halo that a selection drawn close to the
object leaves outside it.
Refs #1366
* Content-Aware Fill: peel a thin hole where that stays affordable (#1366)
Starting every thin hole from a membrane smoothed Spot Healing on film
grain (spot_healing_keeps_film_grain_through_overlapping_strokes: grain
9.0 against 12.1 around). The onion peel copies the grain; it was only
given up because its exhaustive search is too slow on the big level a
long thin hole stops at. It now starts the thin hole whenever hole cells
times sources stay within 2^28 comparisons, and the membrane only beyond.
Refs #1366
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: vangel76 <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* PSD: write tiff blocks without copies, inflate merged ZIP once
- ImageSourceData::to_bytes writes the global mask and global blocks
straight from the model with 4-byte alignment instead of cloning each
block (data included) to set its padding. Same bytes.
- The tiff transcoder's nested regions and blocks write into the parent's
buffer (taken and handed back, error or not) instead of building their
own output and copying it up per nesting level. Bodies that fail or are
not transcoded are truncated before being copied as stored or dropped.
Same bytes.
- PsdFile::from_bytes_and_merged decodes the merged image with the parse
when asked: a ZIP stream is inflated once (the decode's stream check is
the parse's); the parse-only check runs only when the decode fails, to
keep from_bytes' errors. PSD import uses it with merged_needed.
- PsdBuilder::composite documents that Gray/CMYK composites stay straight.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Keyboard shortcut sets: Set menu, New Set, Save, Delete, .kys Export and Summarize (K2)
Edit › Keyboard Shortcuts gets Photoshop's set management. A set is a saved copy of the
bindings (Preferences.shortcutSets, the one in use in shortcutSet; "" is PhotoCraft Defaults),
and the Set menu shows "(modified)" when the bindings differ from it. New Set…, Save and
Delete wait for OK like the rest of the dialog. Export… writes the dialog's bindings as a .kys
file (commands then tools, Photoshop's macOS key spelling), each row carrying our id as pc-id
so PhotoCraft imports its own sets back exactly, whatever the label; Summarize… writes
Photoshop's HTML list of every bound command by menu.
edit.keyboardShortcuts takes useSet (before the call's edits), saveSet, deleteSet, sets and
activeSet (after them), exportKys and summarize; every reply reports the sets, the set in use
and whether it is modified. Bad names, keys or counts are errors that change nothing.
* .kys import: tool keys and sets saved on Windows (K3)
Photoshop's <tool> rows now import: each matches our tool by name (our tool names are
Photoshop's) or by pc-id in sets PhotoCraft exported, and a letter other than the tool's own
becomes its override, while its own letter keeps the tool cycling with the others (#5036).
A set saved on Windows spells the primary modifier Ctrl where a Mac set writes Cmd and keeps
Control for the Control key. A set with no Cmd or Opt anywhere but some Ctrl is read as a
Windows set, its Ctrl as our Cmd: Ctrl+L is the Cmd+L default, not a Control-key override that
would also mean Control on a Mac. The import reports `windows`.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Vikram Bhaskaran <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Tanish Manku <[email protected]>
* Raw: measure the Nikon D90 image area (4288x2848 at (10, 10))
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
* Codecs: open Cineon (.cin) film frames read-only
Cineon is the 10-bit logarithmic film-scan format DPX grew from. Read it from
the same Kodak header family: the generic header (big-endian, orientation at
192, channel count at 193, the channel-1 bits/dimensions), one or three
channels, three 10-bit samples packed into each 32-bit word on a word
boundary. The 10-bit code values open as U16 (round(c x 65535 / 1023)); the
file's log black/white points are a display choice and are not applied. A
depth other than 10 bits or a channel count other than 1/3 is Unsupported;
truncation errors rather than panicking. Read-only (no Cineon writer), so
.cin joins the open filters and Cineon read closes part of #1049's G10 gap.
* cargo fmt the extension lists
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Aligning the decoded D90 sensor plane with the camera's own JPEG puts the
image at (10, 10) in the 4352x2868 sensor plane, giving the camera's
4288x2848 frame (10-pixel margins; the right border reads the camera's flat
255 from column 4310). Add the measured ImageArea so a D90 NEF opens on the
camera's frame instead of the uncropped plane, and make a measured profile
leave a file that carries its own colour calibration (a DNG) untouched.
Co-authored-by: Brandon Thomas <[email protected]>
* fix(ui): brush outline matches the stroke on Wayland at fractional scales
On Hyprland at 1.25 the brush outline was 1.25x too large and sat right of
and below where the brush painted, by more the larger the brush. The engine
places dabs exactly; the OS cursor bitmap was wrong. winit 0.30 sizes custom
cursor bitmaps on Wayland with a wp_viewport destination whenever the scale
isn't 1, and Hyprland (like Mutter before GNOME 48) ignores viewports on
cursor surfaces: it shows the bitmap one pixel per logical pixel with the
hotspot winit divided by the scale.
A client can't tell which compositors do this, so on Wayland at any scale
but 1 the brush outline and crosshairs are painted in-app, as on the web.
Other platforms and Wayland at scale 1 keep the low-latency OS bitmap.
Measured on Hyprland 0.56.2 at 1.25 (100 % zoom, 100 px brush): dab centre
(288.5, 560.5), diameter 97 px; outline before (303, 574.5), ~130 px;
after (288.5, 562), 107 px including its border.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Merge main; cargo fmt
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Fix Info panel guide coordinates during ruler drags (storytold/photocraft#5728).
Show the dragged guide position on one axis while the pointer is outside the canvas, matching Photoshop.
Co-authored-by: Cursor <[email protected]>
* Merge main; fill GuideDrag's new press field in the test; cargo fmt
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Cursor <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Dialogs drawn by the shared renderer had no close box: the only ways
out were Cancel/OK or Esc. As in Photoshop on Windows, the title band now
ends in a × (the icon button the floating panels close with) on Windows,
Linux and the web, none on macOS. A click on it joins Esc in the single
cancel path of `dialogs::show` (outcome = Cancel -> `cancel()`), so the
live preview goes and nothing reaches the document or history. The × is
left out of the title band's drag area, so a drag starting on it neither
moves nor closes the dialog; the rest of the band still drags.
Check Spelling (an untitled window) gets the same close box; it is Done,
as every change there is already made.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
* Warp: a thin layer stretched and bent keeps its curve on commit (#5793)
The mesh warp sized its triangle grid by the source rectangle only (one cell per 4 source px). A 3 px column scaled into a wide band and then warped got a single cell across, so the committed result was a straight, affine band while the preview showed the curve. The grid now also follows the destination length of the mapped rows and columns.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Warp: module doc describes the grid sized on both sides of the map (#5793)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
As in Photoshop, with several layers selected (Shift-click with Auto-Select,
or the Move marquee) Show Transform Controls draws one frame around all of
them (the Background excluded) and outlines every selected shape layer.
Before, the frame showed only the active layer, so the other selected layers
had no feedback on the canvas. The frame joins the boxes Free Transform would
open on each layer (transform_frame), and it hides while a corner radius
widget is dragged.
Free Transform works on one layer, so the shared frame has no handles and a
press on it doesn't start Free Transform; a drag inside still moves every
selected layer.
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Free Transform: shapes and paths stay on whole pixels (#5751)
With Snap Vector Tools and Transforms to Pixel Grid on, only the pointer was
rounded, and only while a shape tool was active. Under the Move tool, or with
a grab offset or a proportional scale, the box landed between pixels (a
rectangle dragged to 540 px came out 539.6). An upright box on a shape layer
or path now keeps whole-pixel edges whatever the tool.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Format (cargo fmt)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
As in Photoshop, with the Layer Style dialog on a Drop Shadow or Inner Shadow
page, dragging on the image moves the shadow: its offset follows the pointer
and is stored as the effect's angle and distance (whole degrees and pixels),
moving the shared global light angle when the effect uses it. Space and the
middle button still pan; colour pickers still sample.
Co-authored-by: Claude Opus 5.5 <[email protected]>
A finished work path is drawn whenever the Pen, Freeform Pen or Path
Selection tool is active, and it is saved with the document, as in
Photoshop. But once the path was finished, Delete/Backspace fell through
to Edit > Clear (the pixels under it), so the only ways to remove it were
the right-click Delete Path and the Paths panel, and it looked impossible
to delete.
With those tools, no path being drawn and the work path the one they
edit (no shape layer or targeted vector mask in front of it), an
unmodified Delete/Backspace now runs path.delete on the work path: one
undoable Delete Path step. Direct Selection and the Curvature Pen are
left alone, since Photoshop deletes their selected anchors, not the path.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The dialogs converted points and picas at a fixed 72 pt/in, while the
rulers, Info, layer properties and the shape dialog use Preferences >
Units & Rulers > Point/Pica Size. With Traditional (72.27) a document
showed different point sizes in the rulers and in Image Size.
The dialog now stores the preference's points per inch when it opens
(__ptPerIn, 72 when missing) and uses it for points, picas (points / 12)
and typed units in the Width/Height fields.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
The toolbar's Switch Colors and Default Colors buttons ran their commands
without type_tool::foreground_changed, so characters highlighted in the
type layer being edited kept their colour. As in Photoshop, they now take
the new foreground, like picking a colour in the Color or Swatches panel.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
* File › Scripts › Browse rejects Photoshop ExtendScript.
The picker listed .js/.jsx files that parse_script cannot run. Filter to
JSON/text/droplet scripts and return a clear error for .js/.jsx paths.
* File › Scripts › Browse keeps Photoshop extensions in the picker.
Review feedback on #5435: filtering .js/.jsx out of the picker hid the
rejection error from UI users. List Photoshop's script types (as #2452
asked) alongside PhotoCraft's so selecting one shows the ExtendScript
error instead of hiding the file.
* fix(ui): restore previous tool when committing or canceling text edit initiated via double-click
* fix(ui): restore return tool only while on type and stay on type for layer thumbs
- commit_with/cancel: restore return tool only while app.ui.tool.is_type()
- edit_active: keep tool on Type when editing from layer thumbnail (matches Photoshop)
- clear text_edit_return_tool when text_edit is dropped on lost text or stale transform
- test tool switch mid-edit and edit_active remaining on Type
---------
Co-authored-by: safvanedge <[email protected]>
egui's DragValue clamps an existing value to its range and reports a change, so showing a
type layer scaled past 1296 pt (or below 0.1 pt) sent type.setStyle with the clamped size.
Value fields now clamp only edits, never a value they merely show.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Interface > Enable Animations off already made scroll areas jump straight
to their target, but the canvas still panned from egui's smoothed wheel
delta, gliding for several frames after each notch. With the preference off
a plain wheel scroll now moves the canvas by the whole notch in the frame it
arrives and stops there.
Co-authored-by: Claude Opus 5.5 <[email protected]>
An Image > Adjustments recorded on a smart object (Brightness/Contrast,
Levels, Curves...) stores image.adjustments.<kind>. The Layers panel row
called smart_ui::open_editor, which only knew filter_dialog and silently did
nothing for those commands, so the filter could never be reopened. Route
adjustments to adjust_dialog, seed it with the recorded params and let OK run
layer.smartFilter.setParams (#5689).
A preset, group or folder dragged in the Brushes panel or the Brush Preset
picker couldn't reach rows out of view: the list didn't scroll while
dragging, so the drag stopped at the last visible row. Holding the drag at
the list's top or bottom edge (or past it) now scrolls it, with the same
ramp as the Layers panel's row drags.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Clouds, Difference Clouds, Add Noise, Fibers and the other seeded filters
exposed a seed number, but a new pattern meant typing a different one. A
Random button beside every seed field now picks a fresh seed, which the live
preview renders at once. "Random" gains its row in the zh-Hans, zh-Hant, ru,
fr and id catalogs, which claim complete coverage.
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Type: keep text selection highlight visible while Command/Ctrl is held
* Type: match Photoshop 1px dark frame and handle styling for type transform
* Merge main; keep the transform box and underline colours from the theme tokens
The selection-highlight fix stands on its own; the hard-coded box and
underline colours and the removed reference point were unrelated to it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: safvanedge <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Dragging selected anchors with the Direct Selection tool ignored Shift,
so lengthening a slanted bar also nudged it vertically. As in Photoshop,
Shift now locks the anchor move to multiples of 45° from the press, in
the live preview and in the committed path.moveAnchors. It works whether
Shift is held from the press or added during the drag; Shift at the
press still toggles the anchor's selection as before.
Reuses stroke_constraint::snap45 (the Gradient tool's Shift lock).
Fixes#5475
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The overlay that keeps the title strip's free gap draggable under a dialog's
modal layer (#4913) only ran with the custom title bar. macOS keeps the OS
traffic lights over the app's own title strip (custom_titlebar is false), so
with the Embedded Profile Mismatch prompt or any dialog open the window could
not be moved. The overlay now runs whenever the strip is drawn.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The main window set a runtime PNG icon everywhere. On macOS that replaces the
app bundle's PhotoCraft.icns in the Dock, so the system couldn't render it
in its own style (Liquid Glass and its tinted and dark variants). The runtime
icon is now skipped when running from an .app bundle; unbundled runs and the
Windows and Linux builds still set it.
Co-authored-by: Claude Opus 5.5 <[email protected]>
German showed "Ja (Y)" next to "(N)ein". Keys now go inside the words only when
every translated label starts with its key; otherwise all labels append it
("Ja (Y)", "Nein (N)", "Abbrechen (C)").
Co-authored-by: Cursor <[email protected]>
Dragging a Curves point (or any control) in the Properties panel of an
adjustment layer previews through `canvas::ensure_proxy_preview`, which
picked its downscale factor from the document size alone
(`proxy::preview_factor`): any document over 5 MP was drawn from a 2-5x
reduced copy at every zoom, so at 100 % the canvas turned blocky until
the mouse was released.
Cap the factor by the view's zoom with the rule the Image > Adjustments
dialogs already use (`adjust_preview::proxy_factor`: the proxy stays at
least at screen resolution) when the wgpu compositor can draw the
full-size document. At 50 % and closer the drag now shows the full-size
composite; zoomed out, the proxy still saves the work. Documents the GPU
compositor doesn't cover keep the old proxy (the full size would be a
CPU composite per change).
Co-authored-by: Claude Opus 5.5 <[email protected]>
As in Photoshop, dragging a guide out of the top ruler with Alt/Option held
makes a vertical guide, and out of the left ruler a horizontal one; without
Alt each ruler makes its own kind, as before.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The Studio, Adwaita and Breeze themes' Move tool options bar had only Auto-Select and a hint, so Show Transform Controls could not be turned on there and a click on a layer never showed its box and handles. The toggle is now in every theme, as Auto-Select is (#1275).
Co-authored-by: Claude Opus 5.5 <[email protected]>
Files dropped from another app's window (a file manager) were imported, but
the source window stayed in front of PhotoCraft. The frame that takes a drop
now asks for focus, so PhotoCraft comes forward as Photoshop does.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The canvas cursor match mapped Tool::Move to CursorIcon::Move, the 4-way
move cursor, while Photoshop shows its standard arrow pointer and marks the
movable layer with the hover outline. Use CursorIcon::Default for the tool
(#5646).
* UI: centre text vertically in shared text fields
TextEdit draws its line at the top of its rect, and the theme's 24 pt
interact height made the edit taller than the field's inside, so text in
widgets::text_field sat a few pixels high (Preferences search, hex fields,
name fields) while dropdowns and value fields were centred. Fill the box
and centre the line in it.
* Fixed failing test
---------
Co-authored-by: Stevan Marjanović <[email protected]>
Three prompts used the formal verb form ("Adjon", "Válasszon",
"kattintson"); the catalog addresses the user informally throughout.
Two colour profile messages put an article before a placeholder
("A(z) „{name}”") and glued a suffix to one ("{mode}-profilja"). The
correct form depends on the inserted value, so the placeholders now
stand after a colon or in brackets.
The Ask On Mismatch hint kept the English option name; it now names
the Hungarian label of that option in Color Settings.
* Filter dialogs: paint the preview computed mid-drag instead of waiting for the pointer to stop
A preview that finished while the slider was still moving was thrown away before it
could reach the canvas, because the worker result had to match the request on every
field including the parameters. Every completion of a drag lands after the value has
moved on, so none of them matched, the worker was handed the newest value instead, and
that one went stale in turn. Nothing was ever painted while the pointer moved: the
effect appeared only once the slider had been still for about a frame of preview
time, which is what "I have to stop the slider and wait" looks like.
The display gate a few lines below already had the opposite rule. It compared the
document, revision, dialog, layer, channel target, command and proxy factor and
deliberately left the parameters out, so that a slightly old frame stays on screen
mid-drag. The publish gate disagreed with it and won, so the frame the display gate
was written to keep never existed.
Publish a successful result when the finished request matches the current one on
everything except the parameters, and let the existing display gate decide whether to
keep it. The two gates now share one predicate, FilterPreviewKey::same_target, so
they cannot drift apart again.
Errors keep the exact match. A value the pointer has already dragged past should not
raise a notice, and a failure is not something to show stale.
What this deliberately does not change:
- A reopened dialog has a new id, a committed document a new revision, a different
command or channel target a different key, so a result from another preview is still
never painted. Covered by a_reopened_dialog_discards_the_previous_dialogs_finished_preview.
- The freshly computed request is still what decides whether a new computation starts,
so the worker still converges on the value the dialog actually holds.
- One preview runs at a time and a stale one is still cancelled only once the values
settle, so this adds no work to the worker; it only stops discarding completed work.
- OK still recomputes at full resolution, so a frame shown mid-drag is never what gets
written to the document.
Test: a_preview_computed_mid_drag_reaches_the_canvas starts a blur at one radius, moves
the slider before that preview lands, and requires the completed frame to reach the
canvas while the slider is still on the new value. It fails on the old gate, timing out
after 10 s with no frame published.
* Filter dialogs: three guards on a mid-drag preview, after review
The first version of this let any completed frame through that matched the dialog
on everything but the parameters. Review found three ways that is wrong, and all
three are now closed and pinned by a test.
A stale result with nothing to show. Storing it replaces a good frame with an empty
one, which the display gate paints as "draw nothing": the canvas blinks back to the
unfiltered layer mid-drag. The exact request may still store an empty result, which
is how a filter that no longer applies clears the canvas, so the guard belongs on
the stale arm only. Splitting the match arm is what makes that separable; the
earlier single arm forced a false choice between the two.
A late frame replacing an exact one. Drag the slider away and straight back: the
frame on screen is already exact for the dialog, and the in-flight frame for the
value on the way out would have overwritten it, showing a blur the user has already
undone. A stale result is now dropped when the stored preview is already exact.
Errors were already scoped to the exact request and stay there: a value the pointer
has dragged past should not raise a notice, and the value it comes to rest on will
raise its own when it is recomputed.
Also pinned by test, because both are easy to undo by accident:
- The reopen test was not reaching the gate. discard_closed reclaims a closed
dialog's job a frame earlier in the real app, and the test called it too, so the
stale result never arrived and the test passed even with the dialog id removed
from the comparison. It now drives the gate directly. Verified by mutation.
- The mid-drag test asserted nothing about what the display gate does with the
published frame, so re-adding params there passed. It now asserts the canvas
actually returns something to draw, and it keeps polling until a frame for the
dialog's current value lands, which is what pins fresh to an exact match.
Tests are deterministic: the worker job is gated on a channel and signals when its
result is already in the worker channel, so nothing sleeps and nothing races the
150 ms settle timer that cancels a job whose values stopped changing.
Verified by mutating each guard and confirming a test catches it:
publish gate back to exact-match ................. killed
drop the is_some guard ........................... killed
drop the already-exact guard .................... killed
same_target ignores the dialog id ............... killed
display gate compares params again .............. killed
fresh relaxed to same_target .................... killed
same_target ignores k ........................... killed
k stays in the comparison on purpose. A 36 MP document previews at k=4, and
preview_factor lowers k as a pixel-sized parameter approaches its minimum, so k does
change under the slider. Relaxing it would paint a frame at a scale the preview
policy exists to avoid, and would loosen the display gate further than upstream
wrote it. The cost is a one-computation pause when the slider crosses a threshold
on a large document; that test keeps the choice visible.
* Filter dialogs: fix a race in the new tests that let CI pass them
The four new tests signalled each other through a channel the worker sends from inside
`compute`. But the worker only puts the result into the channel after `compute` returns:
let result = compute(&work);
let _ = tx.send(result);
So a test that woke on the signal and called `ensure_filter_preview` immediately could
reach `poll()` before that send. On a loaded CI runner it did: ubuntu failed
`a_mid_drag_empty_result_leaves_the_preview_alone` with no frame published at all. The
comment in the test claimed the result was already in the channel, which was the one
thing it was not.
The tests now drive the canvas until the state they need appears, which does not depend
on the order of that send. They also check their invariant on every pass rather than
only at the end: a wrongly published frame can be replaced by the next, correct one
before the loop stops, which had already left the reopen, late-frame and empty-result
tests unable to see their own mutation.
Unchanged: the assertions, the production code, and the four mutations that broke each
guard — re-verified, 7 of 7 still caught.
Runs: 30 back-to-back runs of the four tests on 12 cores, no failure; the full crate
twice, 2800 pass.
---------
Co-authored-by: agent-346022 <[email protected]>
The disabled Number field still drew a DragValue with range 1..=1000,
which clamps the default rows 0 to 1 and reports a change; writing that
back made the Rows checkbox's `positive` fallback turn on, so OK added
guides at the top and bottom. Disabled number fields no longer write
back. Kittest regression covers Rows and Columns.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
The ⇧ square (and a 1:1 fixed ratio) is built in continuous document
coordinates, and `marquee_px` then snapped each axis to the pixel grid on
its own: floor the near edge, ceil the far one. A press between pixels,
dragged 100.5 px, snapped x to 101 and y to 102, so the readout and the
committed selection were off by one. `marquee_preview_px` now grows the
shorter side to the longer one, away from the press point, so the square
still covers every pixel the drag touched. Ants, readout and commit share
this path.
Fixes#5472
Holding Up/Down in a value field (Character panel tracking, size, …)
stepped it a couple of times, then jumped focus to the field above.
Each arrow step re-requested focus on the field that already had it.
egui's request_focus resets the field's event filter, so on the next
frame egui no longer knew the field takes vertical arrows. A key repeat
landing on that frame was read as "move focus up". Single presses
rarely hit that frame; held keys always do.
Only request focus when the field lost it (after a typed sum gives it up
to be worked out). Regression test feeds key repeats on consecutive
frames with a field above and failed before the fix (focus moved after
two steps).
Fixes#5474
Co-authored-by: Claude Opus 5.5 <[email protected]>
Double-clicking the empty workspace shows File › Open (#5367)
Also covers the same fix as #5461; thanks to its author.
Co-authored-by: Nigel Ihza Firdausy <[email protected]>
fix(ui): expose toolbar column toggle in all themes
Also covers the same fix as #4680; thanks to its author.
Co-authored-by: Marco Arnulfo <[email protected]>
fix(heif): a grid's ICC profile is inherited from its tiles
Also covers the same fix as #5591; thanks to its author.
Co-authored-by: nrq6993-a11y <[email protected]>
Limit whole-selection transform traces to a padded one-million-cell grid
while preserving normal viewport caching and thin boundary samples.
Use double-precision intermediates for clipping long perspective lines.
Add regressions for large/disconnected selections and viewport accuracy
near the projective horizon. Preserve the full-resolution selection and
the transform behavior already landed in #5337.
Related: #5297, #3402.
Co-authored-by: LabDennis <[email protected]>
Curves: apply to masks, alpha, and Quick Mask (#5335)
Also covers the same fix as #5578; thanks to its author.
Co-authored-by: Zer0C00L01 <[email protected]>
Adjustments and filters on a smart object add a smart filter, so their live
preview re-renders the smart object from its source through its placement.
The preview runs on a proxy document scaled down by k (large documents), and
proxy_document scaled the smart object's cache but left its transform (and
perspective and filter mask) in full-size document coordinates. The re-render
then came out k times too large on the proxy canvas, so the object looked
zoomed in until OK committed the full-size result.
proxy_document now scales the placement and the filter mask with the proxy.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Options bar: Brush and Eraser size go to 5000 px in every theme (#5708)
Also covers the same fix as #5724; thanks to its author.
Co-authored-by: nrq6993-a11y <[email protected]>
A 16/32-bit document saved as PSD has its layers in a global Lr16/Lr32 block, which the writer put first among the global blocks. A Mtrn/Mt16/Mt32 marker kept from the opened file then followed it, and Photoshop refuses such a file (verified in Photoshop 2026 with psd-tools' transparentbg.psd re-saved both ways: Mt32 first opens, Lr32 first doesn't). The layer block now goes right after the last kept marker, as Photoshop orders them. Fixes#5068.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The Width/Height fields in Properties > Document and in the New Document
dialog clamped their shown value to at least 0.01 of the unit. A 1 x 1 px
document at 300 ppi is 0.0033 in, so showing it in inches clamped it to
0.01 in, the field reported a change and the document became 3 x 3 px
(two history steps, no user edit). The fields now accept down to 0; the
pixel conversion afterwards already clamps to at least 1 px.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
FITS: clamp 32-bit saves to 0..1 and warn, so they reopen as saved (#5627)
Also covers the same fix as #5642; thanks to its author.
Co-authored-by: nrq6993-a11y <[email protected]>
AppImageLauncher exports DESKTOPINTEGRATION even after
X-AppImage-Integrate=false makes it skip the AppImage. AppRun treated that
variable as "already integrated" and wrote no .desktop file, so the menu
stayed empty. Skip our entry only when another .desktop file's Exec already
points at this AppImage.
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: nrq6993-a11y <[email protected]>
warp() computed the source-centre fallback as (x0 + x1) in i32 and the
inverse-mapped tile footprint as `floor() as i32 - 3` / `ceil() as i32 + 3`.
With a strongly skewed finite affine (for example Homography
[1, 1e12, 0, 0, 1, 0, 0, 0, 1] on a 32x32 clip) the float-to-int cast
saturates at i32::MIN/MAX and the following -3/+3 overflows, panicking in
debug builds and wrapping in release. A source rect with coordinates near
1.5e9 would overflow the centre sum the same way.
Widen the centre sum to f64 and use saturating_sub/saturating_add for the
footprint, which is intersected with src_rect afterwards anyway.
Add crates/algo/tests/warp_clipped.rs covering clipped-vs-full equivalence,
huge scale allocation, the extreme-affine footprint, and the far-origin
source centre.
Refs #3446
* fix(doc): variable definitions follow layer id remapping on copy and load
A duplicated document (image.duplicate, history.newDocument) or a .pcraft
loaded with remapped ids assigned fresh layer ids but cloned VariableDef.layer
unchanged, so datasets silently drove no layer (or, after remap, a colliding
one). Remap variable targets through the same id map as layer comps and
slices; a dangling target keeps a non-colliding placeholder. A merged-only
duplicate drops the whole variable binding set, matching how it drops comps.
Closes#4969
* style: cargo fmt duplicate_variables test
---------
Co-authored-by: highperfocused <[email protected]>
fix(export): prompt for a Render Video output directory (#3365)
Also covers the same fix as #3390; thanks to its author.
Co-authored-by: buivanvu94 <[email protected]>
* Vector: bound flattening memory, keep over-budget dashes, cheaper edge build (#5195)
- flatten: a path's flattened points are capped at 1 << 20; past it every
curve keeps a proportional share of its pieces (at least one), so hostile
handles from files (vmsk, SVG, effect outlines) cannot exhaust memory.
fill_rasterizer flattens the whole path under one budget.
- stroke: past the dash step budget (or below coordinate precision),
dash_polyline keeps the dashes made so far and draws the rest solid
instead of dropping the stroke. Open polylines are no longer copied.
- raster: coordinates are clamped to +-1e15 and edges with a non-finite
slope (vanishingly short) are skipped, so huge finite coordinates no
longer produce NaN coverage. fill_rasterizer sorts edges once instead of
after every component (stable sort, identical order).
- Cleanups: dead miter `half`, dead loop-length check in trace_mask,
shared mask render body, map_knots helper, shared straight-segment test.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Vector: bound stroke outline points (#5211)
- stroke_polygons dashes all lines first, then counts round caps and
joins. When their full-resolution tessellation would exceed 1 << 20
outline points, every round cap and join gets the same smaller segment
count (down to 4 per circle, 1 per join; area-exact radius kept).
Strokes under the budget are unchanged; no piece is dropped.
- The dash step budget (100000) is now shared by all lines of a stroke
instead of per polyline, so many subpaths cannot multiply the dash
(and cap) count. Past it the rest is drawn solid, as before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Blur Gallery: Field Blur pins on the image (#3373)
Field Blur opened a parameter dialog with one centre, though the engine
already blends any number of pins. It now works as Photoshop's Blur
Gallery workspace: the image under the dialog takes the pointer. A
click adds a pin and selects it, a drag moves a pin, and a drag around
the selected pin's ring turns its blur up or down (one turn is the
0-500 px range; the arc shows the amount). Delete or Backspace removes
the selected pin. The dialog shows the selected pin's Blur and Preview,
and opens at the right edge, beside the image, so the pins stay in view.
The pins live in the dialog's fields (`pins`, `__selected`), so the
existing filter preview follows every change, OK applies
filter.blurGallery.fieldBlur with them in one step, and automation
drives it with ui.dialog.set. Pins are fractions of the selection's
bounds when there is one, else the canvas, as the engine reads them.
Proxy previews now scale Blur Gallery pins' pixel blur like other pixel
params.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Blur Gallery: Iris Blur's pins and handles on the image (#3373)
Iris Blur opened a parameter dialog with one ellipse. It now uses the
Field Blur workspace too, with Photoshop's handles on the selected pin:
the four on its outline resize that axis and turn the shape, the square
at its corner rounds it towards a rectangle, and the four dots inside
set the sharp core, where the blur starts. A click adds a pin with the
selected pin's shape and blur; pins move, their ring sets the blur,
Delete removes them, and the preview follows.
The pins keep the command's own keys (radiusX/radiusY as fractions of
the bounds' shorter side, angle, roundness, feather), and the handles
use the engine's superellipse geometry, so what the outline shows is
where the filter blurs. Values from automation are cleaned before use.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Blur Gallery: Tilt-Shift and Spin Blur on the image (#3373)
Tilt-Shift and Spin Blur opened parameter dialogs; they now use the
Blur Gallery workspace with Photoshop's handles.
Tilt-Shift: the solid lines bound the sharp band (drag to widen or
narrow it), the dashed lines mark where the blur is full (drag to set
the transition), the dots on the solid lines turn the band, the pin
moves it and its ring sets the blur. The engine takes one band, so a
click adds none and the dialog keeps the command's own centerX,
centerY, angle, focus, transition and blur.
Spin Blur: pins like Iris Blur's, with the outline handles resizing and
turning the ellipse; the ring and the dialog set the blur angle (one
turn of the ring is 360 degrees). Pins keep the command's keys
(radiusX, radiusY, angle, blurAngle).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(vi): translate the Field Blur pin strings
main added the Vietnamese catalog after this branch was cut, so
every_tl_literal_is_translated failed on the merge for "No pin
selected" and the pin hint. Terms follow the catalog's own: ghim
(pin), Bấm (click), Delete kept as the key name.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(vi): translate the Iris Blur handle hint
main added the Vietnamese catalog after this branch was cut; terms
follow the catalog's own (tay nắm for handle, Kéo for drag).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(vi): translate the Tilt-Shift and Spin Blur hints
main added the Vietnamese catalog after this branch was cut; terms
follow the catalog's own (ghim, tay nắm, Chuyển tiếp, Góc làm mờ).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* Blur Gallery: Alt+Cancel Reset starts again from one pin
main's Alt+Cancel → Reset (#3871) restores a filter dialog's schema
defaults, and the schema has none for Field Blur's pins, so Reset
left the workspace with no pins. blur_pins::reset now puts back the
opening pin (in the middle, selected), and reset_fields runs it for
the Blur Gallery. init is reset plus capturing the bounds, which
Reset keeps.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(hu, tr): translate the Field Blur pin strings
main added complete Hungarian and Turkish catalogs after this branch
was cut. Terms follow each catalog: tű / Mezőelmosás (informal
Hungarian), pim / Alan Bulanıklığı (formal Turkish).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(hu, tr): translate the Iris Blur handle hint
Terms follow each catalog: fogantyú for handle (Hungarian), tutamaç
for handle (Turkish).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(hu, tr): translate the Tilt-Shift and Spin Blur hints
Terms follow each catalog: tű, fogantyú, Átmenet (Hungarian); pim,
tutamaç, Geçiş (Turkish).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(sr): translate the Field Blur pin strings
Main gained a complete Serbian catalogue (#4436); every_tl_literal_is_translated needs these strings there too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(sr): translate the Iris Blur handle hint
Main gained a complete Serbian catalogue (#4436); every_tl_literal_is_translated needs these strings there too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(sr): translate the Tilt-Shift and Spin Blur hints
Main gained a complete Serbian catalogue (#4436); every_tl_literal_is_translated needs these strings there too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n: keep main's catalog row where both added the same label
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs/gaps.md: the Tilt-Shift and Spin paragraph replaces the Field/Iris one
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Arrow-key nudges: a run of nudges is one History state, with Tools preferences (#2259)
Tapping or holding an arrow key nudged the selected layers one pixel per
History state, so one Undo took back one pixel. Nudges of the same thing
that follow each other now share one `coalesce` key (the engine already
folds edits with the same key into one step), so a run is one state and
one Undo takes it back.
- nudge_bundle.rs decides which run a nudge belongs to: the previous nudge
was of the same thing (layers, selection outline), came within the
pause, and left the document as it is now (no other edit, Undo/Redo,
layer selection or save in between).
- move_mods.rs runs layer.translate and select.transformSelection through
it; Alt-nudges (a copy each) never join a run. Selected pixels float
until dropped, which is one state already.
- Free Transform keeps its own session history; its box nudges fold into
one session step (transform_tool.rs).
- Tools preferences: Bundle Arrow-Key Nudges (default on; off gives a
state per press as before) and Nudge Bundle Pause (ms) (default 1000,
100..10000, hidden while bundling is off). Labels translated in all
catalogues.
The two #1428 tests that asserted a state per press for the selection
outline now cover both settings.
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Brandon Thomas <[email protected]>
* Export As: exact pixels, named presets, and profile options
Users can set an exact pixel size with the aspect ratio locked or free,
save and reuse named export presets in the preferences document, and
choose whether to convert to sRGB, embed the colour profile, and keep
EXIF or all metadata. Percent scale and the last-used settings stay.
Part of #2289
Co-authored-by: Zer0C00L01 <[email protected]>
* Translate Export As's Color Profile and EXIF labels in every catalog
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* feat(engine): view.editGuide changes a guide in one step
Photoshop's Edit Selected Guides (the guide context menu) reopens New
Guide on the selected guide. view.editGuide takes the guide by orientation
and index and changes whatever is given: "newOrientation", "position"
(pixels or a length with a unit, % of the new orientation's extent) and
"color" (a name or "#rrggbb"), as one undoable "Edit Guide" step. A guide
that keeps its orientation keeps its index; one that turns goes to the
end of the other list. Without "color" the guide keeps its own colour, or
none. Bad params are refused without a history step.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* feat(ui): select guides, the guide context menu and Edit Selected Guides
As in Photoshop, a click on a guide with the Move tool (or ⌘ with a
selection tool) selects it, drawn in Photoshop's selected-guide blue
(#0000ff); a press anywhere else deselects it, and Delete or Backspace
deletes it, before Edit › Clear can reach the pixels. A guide that is
dragged stays selected, and a click no longer nudges a guide to the
pointer. One guide at a time.
Right-click on a guide selects it and opens Photoshop's guide menu: Edit
Selected Guides, Lock Guides, Clear Guides, Clear Canvas Guides, Clear
Selected Guides and Clear Selected Artboard Guides. Edit Selected Guides
opens the New Guide dialog on the guide (its orientation, its position in
the ruler unit and its colour, Cyan for a guide dragged from a ruler) and
OK runs view.editGuide, sending the colour only when it was changed.
The selection is view state (UiState.selected_guide, readable over the
control channel), keyed by document, orientation and position like guide
colours, so it lapses by itself when its guide goes. The menu's two new
labels are in every catalogue.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* feat(ui): Alt-click turns a guide; double-click opens the guide preferences
As in Photoshop, with the Move tool:
- Alt-clicking a guide turns it: a vertical guide becomes horizontal
through the point clicked, and the other way round. It keeps its colour
and stays selected; one undoable Edit Guide step (view.editGuide).
Locked or hidden guides can't be turned.
- Double-clicking a guide opens Preferences › Guides, Grid & Slices.
docs/ui-parity.md lists the guide interactions.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* feat(i18n): Hungarian and Turkish strings for the guide menu
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(ui): the spelling rows call engine_call with the app
main's spelling suggestions (choose_spelling) came in calling
engine_call(menu, command); on this branch engine_call takes the app
first (the guide menu needs it), so the merge didn't compile.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(sr): translate the guide context menu items
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* fix(ui): give main's move-hover test guide drag a press point
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Shortcuts: a command can answer a second key (F2/F3/F4, Shift+F7, Cmd++) (#3665)
Photoshop's default set gives some commands a second key besides the one
its menu shows: F2/F3/F4 Cut/Copy/Paste beside Cmd+X/C/V, Shift+F7 Inverse
beside Cmd+Shift+I, Cmd++ beside Cmd+= for Zoom In, Shift+Backspace for
Fill. A command could hold only one key here, so these couldn't exist and
a .kys import kept only a row's first key (a set listing Copy as Cmd+C and
F3 could leave Cmd+C unbound).
- prefs::ALTERNATE_SHORTCUTS holds Photoshop's second keys;
Preferences.alternate_shortcuts (alternateShortcuts) replaces a
command's defaults, [] removes them. Old settings files load unchanged.
- edit.keyboardShortcuts takes `alternates`; reset, moving a key (taken
from another command's main key or alternates) and the conflict list
count every key, and a call's changes now apply atomically.
- .kys import keeps a row's other keys as the command's alternates.
- Dispatch matches every key (a user's main key wins over a default
alternate); the UI-only SECONDARY table is folded in, so Fill's
Shift+Backspace survives an override of its main key. Menus still show
only the main key.
- Keyboard Shortcuts dialog shows each command's alternates, removable
one by one, with Add Shortcut to add one; Use Default / Reset All
restore them.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Translate Add Shortcut into Hungarian; regenerate the scorecard
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Scorecard: regenerate for the new alternate_shortcuts preference
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* docs: regenerate scorecard after merging main
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Regenerate docs/scorecard.md after merging main
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
As Photoshop's menus on Windows: a tap on Alt (or F10) puts the keyboard
on the in-window menu bar, with File lit; ← and → move along it, ↓, ↩
and Space open the lit menu with its first row lit, Esc or a click
leave. Alt+letter opens the menu with that mnemonic, open or not, and
in an open menu a letter runs the row with that mnemonic or opens its
submenu.
Mnemonics are assigned per menu level from the labels as shown (so
translated), unique within the level: Photoshop's F E I L Y S T V W H
for the top menus where the label has the letter, else the first unused
letter, word starts first; only ASCII letters and digits, since those
are the keys the keyboard reports. They are underlined while Alt is
held, the bar is focused or the menu was opened from the keyboard; a
menu opened with the mouse shows none until Alt, as on Windows. The
underline is a layout job with the placeholder colour, so disabled rows
stay dim; without underlines the text path is unchanged.
The shortcuts defer to the focused bar as they do to an open menu, so E
opens Edit instead of picking the Eraser and the arrows don't nudge the
layer. Rows carry their mnemonic in the navigation state. macOS keeps
the native menu bar, where this doesn't run.
Tests: two unit tests (assignment, key letters) and seven harness tests
(Alt focus and arrows, second Alt / Esc / click, Alt+letter and Enter,
letters in an open menu, underline conditions, the layout job, and the
whole app's shortcut deferral).
Co-authored-by: Brandon Thomas <[email protected]>
* Add Shift+scroll snap zoom to Photoshop zoom ladder levels
When zooming with the scroll wheel (either via Alt+scroll or via the
"Zoom with Scroll Wheel" preference), holding Shift snaps to the next
Photoshop zoom ladder level instead of the continuous ×1.1 step.
- zoom_with_wheel ON: Shift+scroll snaps to the ladder
- zoom_with_wheel OFF: Shift+Alt+scroll snaps to the ladder
Co-authored-by: Claude <[email protected]>
* Handle Shift+wheel arriving as horizontal delta in snap zoom
Some systems (notably egui on certain platforms) deliver Shift+scroll
as a horizontal delta (x != 0, y = 0) rather than preserving the
vertical axis. Use `raw.y + raw.x` so the snap-zoom path catches
both orientations, matching the existing Shift+Alt pattern on the
line above.
Co-authored-by: Claude <[email protected]>
* Snap zoom: per-viewport fractional accumulation, Camera Raw ladder stepping, and horizontal-delta handling
Port the improvements from TheSeanLavery/photocraft#33:
- `read()` gains `owner: Id` and `hovered: bool` so each viewport
(canvas, navigator, Camera Raw) accumulates its own snap residual
without leaking into others.
- Fractional scroll input (sub-notch point events) now accumulates
across frames with an idle timeout, so precision trackpads and
high-resolution mice reach the next ladder rung.
- Camera Raw snap zoom steps through the ladder via `step_in_range()`
instead of being a no-op.
- Shift+wheel arriving as a horizontal delta (x != 0, y == 0) is
handled by using `raw.y + raw.x` in the snap path.
- `zoom_levels::step_in_range()` extracted for reuse with custom bounds.
- 8 new tests (24 wheel_nav total), 4 new Camera Raw snap tests,
1 new zoom_levels test.
Co-authored-by: Claude <[email protected]>
* Fix test destructure after merge with main (#5080) and reformat
#5080 added a test that destructures gesture() as a 2-tuple, but this
branch makes it return 3 values. Update the destructure and run
cargo fmt --all.
---------
Co-authored-by: Claude <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Crop tool: Content-Aware fill of the area a rotated or extended crop uncovers (#3823)
image.crop gains contentAware (bool, default false). After a turned crop or one that extends
past the canvas, the area the original image doesn't cover is filled from its surroundings on
the Background (else the active pixel layer), in the crop's one undo step. Each connected piece
(corner, edge strip) gets its own Content-Aware sampling window and samples only covered pixels;
the patch search is Edit > Fill > Content-Aware's (fill_cmds::content_aware_native, shared). A
crop that uncovers nothing is unchanged; one with nothing to sample from fails and rolls back.
Crop options bar: Fill: Background (Default) / Content-Aware Fill (ToolOptions
crop_content_aware, ui.set cropFill). Generative Expand is not offered.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Crop: adapt to Document::background after main removed is_background
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Crop: translate Background (Default) for Hungarian and Turkish
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Fix the merge with main: UI_SET_FIELDS length, Serbian Background (Default)
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Merge main: UI_SET_FIELDS counts both new fields
---------
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Shapes: on-canvas corner radius widgets for live rectangles (TOOL-214-12)
As in Photoshop, with Path Selection or the Rectangle tool the active shape
layer's live rectangle shows a round widget at the centre of each corner arc
(a few screen points inside a sharp corner). Dragging one rounds every corner
alike, Alt drags that corner alone; the radius follows the move along the
corner's diagonal from the press, within half the shorter side. The drag
previews the new outline and commits one shape.edit with radii on release,
one history step; a click changes nothing.
Docs: ui-parity row done, TOOL-214-12 note updated (still partial: shape
feather), target-app-parity, regenerated scorecard.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Corner radius widgets: every selected rectangle, Move tool, cursor and readout
As in Photoshop:
- every selected live rectangle shows its widgets, also with the Move tool
when Show Transform Controls is on;
- dragging any widget gives every selected rectangle the same radius (each
capped at half its own shorter side); Alt changes that corner on each;
all the shape.edit calls share one history step (coalesce);
- the pointer over a widget is the arrow with a rounded-corner badge, and a
drag shows an "R: 25.0 px" readout;
- while dragging, the transform box and the Move rollover highlight hide, and
the Move tool's snapping started by the press is dropped.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
#5267 and #5263 gave 16-bit adjustment layers Photoshop's measured Posterize and
Threshold rules, but the destructive commands still passed no depth for 16-bit
layers and used the 8-bit rules. The live preview (an adjustment layer) then
disagreed with what OK wrote by up to 92/255, failing ui-egui's
adjust_preview::preview_composite_matches_the_destructive_command on main.
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Blur Gallery: Field Blur pins on the image (#3373)
Field Blur opened a parameter dialog with one centre, though the engine
already blends any number of pins. It now works as Photoshop's Blur
Gallery workspace: the image under the dialog takes the pointer. A
click adds a pin and selects it, a drag moves a pin, and a drag around
the selected pin's ring turns its blur up or down (one turn is the
0-500 px range; the arc shows the amount). Delete or Backspace removes
the selected pin. The dialog shows the selected pin's Blur and Preview,
and opens at the right edge, beside the image, so the pins stay in view.
The pins live in the dialog's fields (`pins`, `__selected`), so the
existing filter preview follows every change, OK applies
filter.blurGallery.fieldBlur with them in one step, and automation
drives it with ui.dialog.set. Pins are fractions of the selection's
bounds when there is one, else the canvas, as the engine reads them.
Proxy previews now scale Blur Gallery pins' pixel blur like other pixel
params.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Blur Gallery: Iris Blur's pins and handles on the image (#3373)
Iris Blur opened a parameter dialog with one ellipse. It now uses the
Field Blur workspace too, with Photoshop's handles on the selected pin:
the four on its outline resize that axis and turn the shape, the square
at its corner rounds it towards a rectangle, and the four dots inside
set the sharp core, where the blur starts. A click adds a pin with the
selected pin's shape and blur; pins move, their ring sets the blur,
Delete removes them, and the preview follows.
The pins keep the command's own keys (radiusX/radiusY as fractions of
the bounds' shorter side, angle, roundness, feather), and the handles
use the engine's superellipse geometry, so what the outline shows is
where the filter blurs. Values from automation are cleaned before use.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(vi): translate the Field Blur pin strings
main added the Vietnamese catalog after this branch was cut, so
every_tl_literal_is_translated failed on the merge for "No pin
selected" and the pin hint. Terms follow the catalog's own: ghim
(pin), Bấm (click), Delete kept as the key name.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(vi): translate the Iris Blur handle hint
main added the Vietnamese catalog after this branch was cut; terms
follow the catalog's own (tay nắm for handle, Kéo for drag).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* Blur Gallery: Alt+Cancel Reset starts again from one pin
main's Alt+Cancel → Reset (#3871) restores a filter dialog's schema
defaults, and the schema has none for Field Blur's pins, so Reset
left the workspace with no pins. blur_pins::reset now puts back the
opening pin (in the middle, selected), and reset_fields runs it for
the Blur Gallery. init is reset plus capturing the bounds, which
Reset keeps.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(hu, tr): translate the Field Blur pin strings
main added complete Hungarian and Turkish catalogs after this branch
was cut. Terms follow each catalog: tű / Mezőelmosás (informal
Hungarian), pim / Alan Bulanıklığı (formal Turkish).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(hu, tr): translate the Iris Blur handle hint
Terms follow each catalog: fogantyú for handle (Hungarian), tutamaç
for handle (Turkish).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_018CiFWEfKibLnwb6oSy6ZBE
* i18n(sr): translate the Field Blur pin strings
Main gained a complete Serbian catalogue (#4436); every_tl_literal_is_translated needs these strings there too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(sr): translate the Iris Blur handle hint
Main gained a complete Serbian catalogue (#4436); every_tl_literal_is_translated needs these strings there too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Curvature Pen: drag an anchor of the active path without a modifier
As in Photoshop, the Curvature Pen with no path being drawn shows the active
path's anchors (work path, active shape layer or targeted vector mask), and a
press-drag on one moves it with no Ctrl/Cmd. The curve reflows: every smooth
knot of that subpath gets its handles recomputed with the Curvature Pen's own
curvature_knots, so it keeps flowing through all its anchors; corners keep
their handles. The drag previews live and commits one history step.
- vector: edit::move_anchors_curvature
- engine: path.moveAnchors takes "curvature": bool=false
- ui-egui: direct_select routes the idle Curvature Pen press (Edit::Curvature)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Curvature Pen: add anchors, toggle smooth/corner and adapt the cursor on finished paths
As in Photoshop, with no path being drawn the Curvature Pen also:
- adds an anchor where a segment of the active path is clicked
(path.addAnchor), and a drag pulls it like any anchor; adding and dragging
are one history step;
- toggles an anchor between smooth and corner on a double-click
(path.convertPoint with the new "curvature" param, vector
edit::toggle_curvature_point): a Pen-drawn corner rounds off and the
segments around it curve through it; a second double-click makes it a
corner again;
- shows what a press does in its cursor: + over a segment, four arrows over
an anchor (also an anchor of the path being drawn).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* rustfmt
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* feat(psd): map Filter Gallery smart filters to filter.filterGallery (Chrome)
A PSD smart filter from Filter > Filter Gallery (`GEfc`) was kept verbatim
as psd.unsupportedFilter, so it could not be edited. Its effect stack
(`GEfs` list of `GEfc` effects with `GEfk` effect enum and `GELv`
visibility) now maps to `filter.filterGallery {effects, foreground,
background}` when every effect and setting is modelled, and is written
back as `GEfc` on export. Layout taken from a Photopea-authored PSD.
The table starts with Chrome (`Chrm`: `Dtl ` detail, `Smth` smoothness).
Stacks with other effects or unknown settings stay verbatim as before.
* feat(smart-filters): edit Filter Gallery smart filters in the gallery dialog
Double-clicking a Filter Gallery smart filter in the Layers panel did
nothing. It now opens the Filter Gallery dialog with the recorded effect
stack and colours, previewing on the pixels below that filter; OK updates
the filter in place with layer.smartFilter.setParams (one history step).
* fix(psd): read and write Photoshop's inline Filter Gallery effect
Photoshop 27 writes a Filter Gallery smart filter with a single effect
inline in `Fltr` (`GEfc { GEfk, <settings> }`) rather than as a `GEfs`
list, so such files still opened as psd.unsupportedFilter. Both layouts
are read now. A single visible effect is written inline as Photoshop does;
several or hidden effects keep the list, which is the only layout with a
visibility flag (`GELv`).
---------
Co-authored-by: Brandon Thomas <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Modern Brightness/Contrast: Photoshop's brightness curve
The modern brightness curve was a line rolled off by a v^P term. It sat
up to 4 levels off Photoshop's in the highlights, and up to 2.3 levels
off for lowered brightness.
The curve is refit to the Photoshop ground truth we have:
- the ±50 samples from 4759d12b;
- the photoshop corpus's brightness +25 / contrast +30 files,
inverted through the (exact) contrast curve.
Raised, it is the line s*x (s = 1.375^(b/50)) up to a knee at 80/255,
then a cubic Hermite to white with end slope s^-6.6. Lowered, it is the
exact inverse of the curve raised by as much (the -50 samples confirm
this to within a level). Every measured point is within about a level.
For strong values the knee moves down so the curve stays monotone
across -150..150.
Brightness then contrast are evaluated through one 4096-entry table,
built in 55 us. The GPU uses that same table through the Levels/Curves
kernel instead of rebuilding the curves in WGSL, so GPU and CPU agree
by construction.
Photoshop oracles: rgb8/16/32 and gray8/16/32 brightness-contrast now
pass (145 -> 151 of 258). The floor goes 133 -> 151.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Brightness/Contrast curves each ink, as Photoshop does
Photoshop applies Brightness/Contrast in a CMYK document to each ink's
stored value, 1 - ink, black included. Both the modern and the legacy
curve do this; the cmyk8 brightness-contrast(-legacy) oracles match it
within a level on every pixel of the files' own inks. Raising
brightness takes ink away. We applied it to the screen colour instead.
- cmyk_channel_adjustment gains Brightness/Contrast. apply_in builds
the channel curve once (brightness_contrast_lut) and runs each
separated ink through it (ink_through).
- Destructive Image > Adjustments on a CMYK layer curves the stored
inks directly, so it is exact; a selection mixes them.
adjusted_channels now takes the per-pixel operation, and works out
luma only for CMYK Threshold, which compares it.
- The GPU leaves CMYK Brightness/Contrast to the CPU, like the other
ink rules.
cmyk8/16 brightness-contrast oracles, bad pixels:
- modern: 72% -> 48-57%, max error 0.19 -> 0.09
- legacy: 71% -> 51-53%
The rest is our separation of the RGB composite, not the rule.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK Threshold compares the inks' own luma, as Photoshop does
Photoshop's CMYK Threshold splits on the Rec. 601 luma of the inks' plain
complement, (1 - c)(1 - k) and so on, whatever the document's profile.
On the files' own inks this puts every pixel of cmyk8 and cmyk16
threshold-128.psd on the right side of level 128. The profile's RGB luma
missed 100.
stored_channels works the luma out from the inks itself (ink_luma), so it
no longer takes one; the engine's stored-channel path no longer converts
each pixel to RGB at all. Through the separated RGB composite the cmyk
threshold oracles drop from 1.6% to 0.2% bad pixels.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK legacy Brightness/Contrast pivots at 128 in ink
Photoshop's legacy contrast pivots at level 128 of the ink, which is 127
of 255 of the stored value (1 - ink) that the curve runs on, not 128 as
in RGB. On Photoshop's own 16-bit inks (cmyk16
brightness-contrast-legacy.psd) ours sat a constant 0.43 of a level
above. That is exactly the one-level pivot shift times contrast +30's
slope minus one; with the pivot moved, the inks match within 0.01.
brightness_contrast_lut is now ink_brightness_contrast_lut, since its
legacy curve is the CMYK one.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: adjustment layers composite on the document's own inks
The compositor works in display RGB. In a CMYK document, an adjustment
layer that Photoshop evaluates on the inks could only act on our
separation of the RGB composite, and that separation's black differs
from the file's. So even exact ink rules missed: CMYK Invert was 10%
off, Levels and Curves 50-65%.
The ink-native compositor (compose::native, until now only for
flattening plain raster stacks) now takes adjustment layers too:
- Invert, Posterize, Threshold and Brightness/Contrast (modern and
legacy), and Levels and Curves on CMYK channels, prepared once
(InkStack, InkOp) and applied to the straight inks.
- Opacity and an unfeathered layer mask mix the inks linearly, as in
Photoshop; raster layers' masks scale their coverage the same way.
- A CMYK stack of plain raster layers and such adjustments renders this
way for display, export and reduced views (render_cmyk, tiles in
parallel); flattening and merging keep the adjusted inks exactly.
- Stacks of raster layers alone keep the RGB compositor, which the GPU
matches. The GPU already leaves these adjustments to the CPU.
Photoshop oracles: cmyk8 and cmyk16 invert, posterize-4, threshold-128,
levels, curves, brightness-contrast and brightness-contrast-legacy all
pass (14 files), 151 -> 165 of 258; floor 151 -> 165. psd-tools and io
corpora unchanged. 24 MP CMYK8 with one adjustment renders in 286 ms on
the inks, against 442 ms through the RGB compositor.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK ink compositing inlines open groups
The ink compositor gave up on any group. Now a group at full opacity
without a mask is inlined:
- a pass-through group's children composite straight onto the backdrop,
so an adjustment inside it reaches the layers below;
- an isolated (Normal) group of raster layers alone composites as its
children would, since Normal is associative.
An isolated group with an adjustment, which only sees the group's
content, keeps the RGB compositor. So do a translucent group and an
artboard. Nesting is bounded.
psd-tools' CMYK adjustment files keep their swatches in pass-through
groups. With #4948's fair comparison, curves_cmyk drops from 37% to 0.3%
bad pixels and levels_cmyk from 14% to 0.2%. The photoshop corpus is
unchanged at 165.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Hue/Saturation and Color Balance keep the black ink
In a CMYK document Photoshop applies Hue/Saturation (not Colorize) and
Color Balance to C, M and Y as the complement of an RGB colour, and
leaves the black ink alone. On the photoshop corpus's own inks,
cmyk16 hue-saturation.psd follows this on every pixel within half a
level (cmyk8 within two), and color-balance within our RGB Color
Balance's own error. We applied both to the separated screen colour,
which also rewrote K.
- adjust::cmy_adjustment and apply_cmy run the RGB adjustment once per
band on (1 - c, 1 - m, 1 - y).
- The ink compositor (InkOp::Cmy), the RGB compositor's CMYK branch and
destructive Image > Adjustments (pixels::adjusted_cmy) use them. The
GPU leaves these to the CPU in CMYK documents, as it does the other
ink rules.
- Colorize separates its colour again, black included, so it keeps the
screen-colour path.
cmyk8/16 oracles, bad pixels:
- hue-saturation: 100% -> 0.4-0.7%
- color-balance: 74% -> 25%
Neither passes yet. Half a level of cyan ink is three levels of red
through the profile at saturated cyan, and our RGB models leave that
much. The pass count stays at 165.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Modern brightness: knee at output 126, fit out to -95 and +150
psd-tools' brightnesscontrast_rgb.psd masks four Brightness/Contrast
layers to their own columns. Read through the exact contrast curve, it
gives Photoshop's brightness at -95, -31, +23 and +150, well past the
+/-50 samples the first fit had. The +/-50 curve extrapolated badly
there: 22 levels off at -95 and 25 at +150.
Every setting puts the knee at about the same output level, not input.
The roll-off's end slope falls as 1/(1 + 0.084 b). With the knee at
126/255 and that end slope, the cubic Hermite fits -95, -50, -31, +23,
+25 and +50 within 1.4 levels. Past about +110 no such cubic stays
below white, so there the roll-off is 1 - (1 - y0)(1 - u)^alpha, with
the slope continuous at the knee. That gives +150 within 6 levels,
where every curve family tried bottoms out. Lowered curves are still
the inverse of raised ones.
The photoshop corpus's six brightness-contrast oracles still pass.
psd-tools brightnesscontrast_rgb goes from 0.13 to 0.020 max error and
_grayscale from 0.098 to 0.016. The new measurements are pinned in
modern_brightness_matches_photoshop.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Lab: adjustment layers composite on the stored L, a and b
The compositor reads a Lab document's layers through display RGB, which
clips colours outside its gamut, so the adjustments Photoshop evaluates
on L, a and b (Invert, Posterize, Threshold, Brightness/Contrast, Levels
and Curves with Lab records) only saw the clipped colours: the photoshop
corpus's Lab oracles were up to 11 levels off on saturated colours.
compose::native_lab works a Lab stack of plain layers out on the stored
values instead, like the CMYK ink path: Normal mixes L, a and b
linearly, an opaque layer keeps them exactly, a layer mask scales the
mix, and adjustment layers run adjust::lab_stored_op on each pixel.
Stacks with such adjustments render for display this way; flattening a
plain Lab stack keeps its stored values (exports no longer clip).
Anything else (other blend modes, effects, other adjustments) keeps the
RGB compositor.
Photoshop oracles 145 -> 152: lab16 invert, posterize-4, levels, curves
and brightness-contrast-legacy; lab8 invert and posterize-4. lab8
levels and legacy B/C come within a level (their 8-bit rules are #5143
and #5132). psd-tools and io unchanged; every round trip unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Posterize works on the stored values (1 - ink)
Photoshop posterizes a CMYK channel's stored value, 1 - ink, not the
ink: with levels whose steps are uneven (3, 7, 13, 21) the two differ,
e.g. 3 levels take an ink of 85 to 128, where posterizing the ink drops
it to none. psd-tools posterize_cmyk.psd goes from 13.8 % of pixels off
to exact; psd-tools 241 -> 242, floor 242.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Raise corpus floors to the results of the landed adjustment fixes
Photoshop-authored 165 -> 185 of 258, io 146 -> 148 of 170, psd-tools 242 -> 249 of 309,
measured with the CMYK, Lab, 32-bit, Levels, Threshold, Posterize, feather and
pass-through PRs merged together.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Modern Brightness/Contrast: Photoshop's brightness curve
The modern brightness curve was a line rolled off by a v^P term. It sat
up to 4 levels off Photoshop's in the highlights, and up to 2.3 levels
off for lowered brightness.
The curve is refit to the Photoshop ground truth we have:
- the ±50 samples from 4759d12b;
- the photoshop corpus's brightness +25 / contrast +30 files,
inverted through the (exact) contrast curve.
Raised, it is the line s*x (s = 1.375^(b/50)) up to a knee at 80/255,
then a cubic Hermite to white with end slope s^-6.6. Lowered, it is the
exact inverse of the curve raised by as much (the -50 samples confirm
this to within a level). Every measured point is within about a level.
For strong values the knee moves down so the curve stays monotone
across -150..150.
Brightness then contrast are evaluated through one 4096-entry table,
built in 55 us. The GPU uses that same table through the Levels/Curves
kernel instead of rebuilding the curves in WGSL, so GPU and CPU agree
by construction.
Photoshop oracles: rgb8/16/32 and gray8/16/32 brightness-contrast now
pass (145 -> 151 of 258). The floor goes 133 -> 151.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Brightness/Contrast curves each ink, as Photoshop does
Photoshop applies Brightness/Contrast in a CMYK document to each ink's
stored value, 1 - ink, black included. Both the modern and the legacy
curve do this; the cmyk8 brightness-contrast(-legacy) oracles match it
within a level on every pixel of the files' own inks. Raising
brightness takes ink away. We applied it to the screen colour instead.
- cmyk_channel_adjustment gains Brightness/Contrast. apply_in builds
the channel curve once (brightness_contrast_lut) and runs each
separated ink through it (ink_through).
- Destructive Image > Adjustments on a CMYK layer curves the stored
inks directly, so it is exact; a selection mixes them.
adjusted_channels now takes the per-pixel operation, and works out
luma only for CMYK Threshold, which compares it.
- The GPU leaves CMYK Brightness/Contrast to the CPU, like the other
ink rules.
cmyk8/16 brightness-contrast oracles, bad pixels:
- modern: 72% -> 48-57%, max error 0.19 -> 0.09
- legacy: 71% -> 51-53%
The rest is our separation of the RGB composite, not the rule.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK Threshold compares the inks' own luma, as Photoshop does
Photoshop's CMYK Threshold splits on the Rec. 601 luma of the inks' plain
complement, (1 - c)(1 - k) and so on, whatever the document's profile.
On the files' own inks this puts every pixel of cmyk8 and cmyk16
threshold-128.psd on the right side of level 128. The profile's RGB luma
missed 100.
stored_channels works the luma out from the inks itself (ink_luma), so it
no longer takes one; the engine's stored-channel path no longer converts
each pixel to RGB at all. Through the separated RGB composite the cmyk
threshold oracles drop from 1.6% to 0.2% bad pixels.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK legacy Brightness/Contrast pivots at 128 in ink
Photoshop's legacy contrast pivots at level 128 of the ink, which is 127
of 255 of the stored value (1 - ink) that the curve runs on, not 128 as
in RGB. On Photoshop's own 16-bit inks (cmyk16
brightness-contrast-legacy.psd) ours sat a constant 0.43 of a level
above. That is exactly the one-level pivot shift times contrast +30's
slope minus one; with the pivot moved, the inks match within 0.01.
brightness_contrast_lut is now ink_brightness_contrast_lut, since its
legacy curve is the CMYK one.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: adjustment layers composite on the document's own inks
The compositor works in display RGB. In a CMYK document, an adjustment
layer that Photoshop evaluates on the inks could only act on our
separation of the RGB composite, and that separation's black differs
from the file's. So even exact ink rules missed: CMYK Invert was 10%
off, Levels and Curves 50-65%.
The ink-native compositor (compose::native, until now only for
flattening plain raster stacks) now takes adjustment layers too:
- Invert, Posterize, Threshold and Brightness/Contrast (modern and
legacy), and Levels and Curves on CMYK channels, prepared once
(InkStack, InkOp) and applied to the straight inks.
- Opacity and an unfeathered layer mask mix the inks linearly, as in
Photoshop; raster layers' masks scale their coverage the same way.
- A CMYK stack of plain raster layers and such adjustments renders this
way for display, export and reduced views (render_cmyk, tiles in
parallel); flattening and merging keep the adjusted inks exactly.
- Stacks of raster layers alone keep the RGB compositor, which the GPU
matches. The GPU already leaves these adjustments to the CPU.
Photoshop oracles: cmyk8 and cmyk16 invert, posterize-4, threshold-128,
levels, curves, brightness-contrast and brightness-contrast-legacy all
pass (14 files), 151 -> 165 of 258; floor 151 -> 165. psd-tools and io
corpora unchanged. 24 MP CMYK8 with one adjustment renders in 286 ms on
the inks, against 442 ms through the RGB compositor.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK ink compositing inlines open groups
The ink compositor gave up on any group. Now a group at full opacity
without a mask is inlined:
- a pass-through group's children composite straight onto the backdrop,
so an adjustment inside it reaches the layers below;
- an isolated (Normal) group of raster layers alone composites as its
children would, since Normal is associative.
An isolated group with an adjustment, which only sees the group's
content, keeps the RGB compositor. So do a translucent group and an
artboard. Nesting is bounded.
psd-tools' CMYK adjustment files keep their swatches in pass-through
groups. With #4948's fair comparison, curves_cmyk drops from 37% to 0.3%
bad pixels and levels_cmyk from 14% to 0.2%. The photoshop corpus is
unchanged at 165.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Modern brightness: knee at output 126, fit out to -95 and +150
psd-tools' brightnesscontrast_rgb.psd masks four Brightness/Contrast
layers to their own columns. Read through the exact contrast curve, it
gives Photoshop's brightness at -95, -31, +23 and +150, well past the
+/-50 samples the first fit had. The +/-50 curve extrapolated badly
there: 22 levels off at -95 and 25 at +150.
Every setting puts the knee at about the same output level, not input.
The roll-off's end slope falls as 1/(1 + 0.084 b). With the knee at
126/255 and that end slope, the cubic Hermite fits -95, -50, -31, +23,
+25 and +50 within 1.4 levels. Past about +110 no such cubic stays
below white, so there the roll-off is 1 - (1 - y0)(1 - u)^alpha, with
the slope continuous at the knee. That gives +150 within 6 levels,
where every curve family tried bottoms out. Lowered curves are still
the inverse of raised ones.
The photoshop corpus's six brightness-contrast oracles still pass.
psd-tools brightnesscontrast_rgb goes from 0.13 to 0.020 max error and
_grayscale from 0.098 to 0.016. The new measurements are pinned in
modern_brightness_matches_photoshop.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Lab: adjustment layers composite on the stored L, a and b
The compositor reads a Lab document's layers through display RGB, which
clips colours outside its gamut, so the adjustments Photoshop evaluates
on L, a and b (Invert, Posterize, Threshold, Brightness/Contrast, Levels
and Curves with Lab records) only saw the clipped colours: the photoshop
corpus's Lab oracles were up to 11 levels off on saturated colours.
compose::native_lab works a Lab stack of plain layers out on the stored
values instead, like the CMYK ink path: Normal mixes L, a and b
linearly, an opaque layer keeps them exactly, a layer mask scales the
mix, and adjustment layers run adjust::lab_stored_op on each pixel.
Stacks with such adjustments render for display this way; flattening a
plain Lab stack keeps its stored values (exports no longer clip).
Anything else (other blend modes, effects, other adjustments) keeps the
RGB compositor.
Photoshop oracles 145 -> 152: lab16 invert, posterize-4, levels, curves
and brightness-contrast-legacy; lab8 invert and posterize-4. lab8
levels and legacy B/C come within a level (their 8-bit rules are #5143
and #5132). psd-tools and io unchanged; every round trip unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Posterize works on the stored values (1 - ink)
Photoshop posterizes a CMYK channel's stored value, 1 - ink, not the
ink: with levels whose steps are uneven (3, 7, 13, 21) the two differ,
e.g. 3 levels take an ink of 85 to 128, where posterizing the ink drops
it to none. psd-tools posterize_cmyk.psd goes from 13.8 % of pixels off
to exact; psd-tools 241 -> 242, floor 242.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Modern Brightness/Contrast: Photoshop's brightness curve
The modern brightness curve was a line rolled off by a v^P term. It sat
up to 4 levels off Photoshop's in the highlights, and up to 2.3 levels
off for lowered brightness.
The curve is refit to the Photoshop ground truth we have:
- the ±50 samples from 4759d12b;
- the photoshop corpus's brightness +25 / contrast +30 files,
inverted through the (exact) contrast curve.
Raised, it is the line s*x (s = 1.375^(b/50)) up to a knee at 80/255,
then a cubic Hermite to white with end slope s^-6.6. Lowered, it is the
exact inverse of the curve raised by as much (the -50 samples confirm
this to within a level). Every measured point is within about a level.
For strong values the knee moves down so the curve stays monotone
across -150..150.
Brightness then contrast are evaluated through one 4096-entry table,
built in 55 us. The GPU uses that same table through the Levels/Curves
kernel instead of rebuilding the curves in WGSL, so GPU and CPU agree
by construction.
Photoshop oracles: rgb8/16/32 and gray8/16/32 brightness-contrast now
pass (145 -> 151 of 258). The floor goes 133 -> 151.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK: Brightness/Contrast curves each ink, as Photoshop does
Photoshop applies Brightness/Contrast in a CMYK document to each ink's
stored value, 1 - ink, black included. Both the modern and the legacy
curve do this; the cmyk8 brightness-contrast(-legacy) oracles match it
within a level on every pixel of the files' own inks. Raising
brightness takes ink away. We applied it to the screen colour instead.
- cmyk_channel_adjustment gains Brightness/Contrast. apply_in builds
the channel curve once (brightness_contrast_lut) and runs each
separated ink through it (ink_through).
- Destructive Image > Adjustments on a CMYK layer curves the stored
inks directly, so it is exact; a selection mixes them.
adjusted_channels now takes the per-pixel operation, and works out
luma only for CMYK Threshold, which compares it.
- The GPU leaves CMYK Brightness/Contrast to the CPU, like the other
ink rules.
cmyk8/16 brightness-contrast oracles, bad pixels:
- modern: 72% -> 48-57%, max error 0.19 -> 0.09
- legacy: 71% -> 51-53%
The rest is our separation of the RGB composite, not the rule.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK Threshold compares the inks' own luma, as Photoshop does
Photoshop's CMYK Threshold splits on the Rec. 601 luma of the inks' plain
complement, (1 - c)(1 - k) and so on, whatever the document's profile.
On the files' own inks this puts every pixel of cmyk8 and cmyk16
threshold-128.psd on the right side of level 128. The profile's RGB luma
missed 100.
stored_channels works the luma out from the inks itself (ink_luma), so it
no longer takes one; the engine's stored-channel path no longer converts
each pixel to RGB at all. Through the separated RGB composite the cmyk
threshold oracles drop from 1.6% to 0.2% bad pixels.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* CMYK legacy Brightness/Contrast pivots at 128 in ink
Photoshop's legacy contrast pivots at level 128 of the ink, which is 127
of 255 of the stored value (1 - ink) that the curve runs on, not 128 as
in RGB. On Photoshop's own 16-bit inks (cmyk16
brightness-contrast-legacy.psd) ours sat a constant 0.43 of a level
above. That is exactly the one-level pivot shift times contrast +30's
slope minus one; with the pivot moved, the inks match within 0.01.
brightness_contrast_lut is now ink_brightness_contrast_lut, since its
legacy curve is the CMYK one.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Modern brightness: knee at output 126, fit out to -95 and +150
psd-tools' brightnesscontrast_rgb.psd masks four Brightness/Contrast
layers to their own columns. Read through the exact contrast curve, it
gives Photoshop's brightness at -95, -31, +23 and +150, well past the
+/-50 samples the first fit had. The +/-50 curve extrapolated badly
there: 22 levels off at -95 and 25 at +150.
Every setting puts the knee at about the same output level, not input.
The roll-off's end slope falls as 1/(1 + 0.084 b). With the knee at
126/255 and that end slope, the cubic Hermite fits -95, -50, -31, +23,
+25 and +50 within 1.4 levels. Past about +110 no such cubic stays
below white, so there the roll-off is 1 - (1 - y0)(1 - u)^alpha, with
the slope continuous at the knee. That gives +150 within 6 levels,
where every curve family tried bottoms out. Lowered curves are still
the inverse of raised ones.
The photoshop corpus's six brightness-contrast oracles still pass.
psd-tools brightnesscontrast_rgb goes from 0.13 to 0.020 max error and
_grayscale from 0.098 to 0.016. The new measurements are pinned in
modern_brightness_matches_photoshop.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Lab: adjustment layers composite on the stored L, a and b
The compositor reads a Lab document's layers through display RGB, which
clips colours outside its gamut, so the adjustments Photoshop evaluates
on L, a and b (Invert, Posterize, Threshold, Brightness/Contrast, Levels
and Curves with Lab records) only saw the clipped colours: the photoshop
corpus's Lab oracles were up to 11 levels off on saturated colours.
compose::native_lab works a Lab stack of plain layers out on the stored
values instead, like the CMYK ink path: Normal mixes L, a and b
linearly, an opaque layer keeps them exactly, a layer mask scales the
mix, and adjustment layers run adjust::lab_stored_op on each pixel.
Stacks with such adjustments render for display this way; flattening a
plain Lab stack keeps its stored values (exports no longer clip).
Anything else (other blend modes, effects, other adjustments) keeps the
RGB compositor.
Photoshop oracles 145 -> 152: lab16 invert, posterize-4, levels, curves
and brightness-contrast-legacy; lab8 invert and posterize-4. lab8
levels and legacy B/C come within a level (their 8-bit rules are #5143
and #5132). psd-tools and io unchanged; every round trip unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Type and Content-Aware Move boxes: resize arrows follow a rotated view
The type box (paragraph text handles, Ctrl/⌘ transform corners) and
Content-Aware Move's Transform On Drop box chose their resize arrows with
type_transform::resize_cursor, which mirrored the direction for a flipped
view but ignored a rotated one: on a view turned 90°, the side handles
showed ↕ although they resize left and right on screen.
The helper now takes the view and maps the handle's direction through it
(ViewXform::map_vec: rotation, flip and pixel aspect) before rounding to the
nearest of egui's four resize arrows. It moves to the canvas cursors as
tool_cursor::resize, so the other tools with resize handles can use it too.
The new test, resize_follows_the_view, checks views turned 0°, 45° and 90°
and a flipped view, with the expected arrows written out by hand; it fails
when the view's rotation is ignored.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Free Transform, Warp, Artboard, guides, Crop: arrows follow the view
The remaining tools with resize arrows chose them without the view:
- Free Transform picked the arrow from the handle's index alone (corners 0
and 2 ↖↘, 1 and 3 ↗↙, edges 0 and 2 ↕, 1 and 3 ↔), so a rotated box showed
wrong arrows (all 8 handles at 45°, 90° and 135°), and so did a rotated or
flipped view.
- Warp's Split Vertically/Horizontally cursor was always ↔/↕, though the
split line follows the grid (turned when Warp starts on a rotated box) and
the view.
- The Artboard tool used a fixed table, ignoring a rotated or flipped view.
- Guides turn with a rotated view (#3277), but the cursor over them stayed ↔
for a vertical guide and ↕ for a horizontal one.
- Crop followed the frame's angle and a rotated view, but not a flipped one:
its corners showed the other diagonal.
Free Transform and the Artboard tool now get the canvas's view and use
tool_cursor::resize: Free Transform with the direction from the opposite
handle, the Artboard tool with the handle's outward direction.
rulers::guide_cursor uses it for the direction across a guide, and Warp's
split_cursor for the grid's own direction across the split line at the
pointer (split_cursor and quick_split_kind take the grid position, so the
cursor looks it up once). Crop keeps its own resize_icon, which now mirrors
the turned direction when the view is flipped. On an upright, unflipped view
every arrow is as before.
New tests, with the expected arrows written out by hand, fail on the old
code: resize_cursors_turn_with_the_box_and_the_view (Free Transform),
warp_split_cursors_turn_with_the_grid_and_the_view,
resize_cursors_turn_with_the_view (Artboard),
guide_cursors_turn_with_a_rotated_view, and flipped-view checks in Crop's
cursor_follows_the_frame.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Modern Brightness/Contrast: Photoshop's brightness curve
The modern brightness curve was a line rolled off by a v^P term. It sat
up to 4 levels off Photoshop's in the highlights, and up to 2.3 levels
off for lowered brightness.
The curve is refit to the Photoshop ground truth we have:
- the ±50 samples from 4759d12b;
- the photoshop corpus's brightness +25 / contrast +30 files,
inverted through the (exact) contrast curve.
Raised, it is the line s*x (s = 1.375^(b/50)) up to a knee at 80/255,
then a cubic Hermite to white with end slope s^-6.6. Lowered, it is the
exact inverse of the curve raised by as much (the -50 samples confirm
this to within a level). Every measured point is within about a level.
For strong values the knee moves down so the curve stays monotone
across -150..150.
Brightness then contrast are evaluated through one 4096-entry table,
built in 55 us. The GPU uses that same table through the Levels/Curves
kernel instead of rebuilding the curves in WGSL, so GPU and CPU agree
by construction.
Photoshop oracles: rgb8/16/32 and gray8/16/32 brightness-contrast now
pass (145 -> 151 of 258). The floor goes 133 -> 151.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Modern brightness: knee at output 126, fit out to -95 and +150
psd-tools' brightnesscontrast_rgb.psd masks four Brightness/Contrast
layers to their own columns. Read through the exact contrast curve, it
gives Photoshop's brightness at -95, -31, +23 and +150, well past the
+/-50 samples the first fit had. The +/-50 curve extrapolated badly
there: 22 levels off at -95 and 25 at +150.
Every setting puts the knee at about the same output level, not input.
The roll-off's end slope falls as 1/(1 + 0.084 b). With the knee at
126/255 and that end slope, the cubic Hermite fits -95, -50, -31, +23,
+25 and +50 within 1.4 levels. Past about +110 no such cubic stays
below white, so there the roll-off is 1 - (1 - y0)(1 - u)^alpha, with
the slope continuous at the knee. That gives +150 within 6 levels,
where every curve family tried bottoms out. Lowered curves are still
the inverse of raised ones.
The photoshop corpus's six brightness-contrast oracles still pass.
psd-tools brightnesscontrast_rgb goes from 0.13 to 0.020 max error and
_grayscale from 0.098 to 0.016. The new measurements are pinned in
modern_brightness_matches_photoshop.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Lab: adjustment layers composite on the stored L, a and b
The compositor reads a Lab document's layers through display RGB, which
clips colours outside its gamut, so the adjustments Photoshop evaluates
on L, a and b (Invert, Posterize, Threshold, Brightness/Contrast, Levels
and Curves with Lab records) only saw the clipped colours: the photoshop
corpus's Lab oracles were up to 11 levels off on saturated colours.
compose::native_lab works a Lab stack of plain layers out on the stored
values instead, like the CMYK ink path: Normal mixes L, a and b
linearly, an opaque layer keeps them exactly, a layer mask scales the
mix, and adjustment layers run adjust::lab_stored_op on each pixel.
Stacks with such adjustments render for display this way; flattening a
plain Lab stack keeps its stored values (exports no longer clip).
Anything else (other blend modes, effects, other adjustments) keeps the
RGB compositor.
Photoshop oracles 145 -> 152: lab16 invert, posterize-4, levels, curves
and brightness-contrast-legacy; lab8 invert and posterize-4. lab8
levels and legacy B/C come within a level (their 8-bit rules are #5143
and #5132). psd-tools and io unchanged; every round trip unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Lab: adjustment layers composite on the stored L, a and b
The compositor reads a Lab document's layers through display RGB, which
clips colours outside its gamut, so the adjustments Photoshop evaluates
on L, a and b (Invert, Posterize, Threshold, Brightness/Contrast, Levels
and Curves with Lab records) only saw the clipped colours: the photoshop
corpus's Lab oracles were up to 11 levels off on saturated colours.
compose::native_lab works a Lab stack of plain layers out on the stored
values instead, like the CMYK ink path: Normal mixes L, a and b
linearly, an opaque layer keeps them exactly, a layer mask scales the
mix, and adjustment layers run adjust::lab_stored_op on each pixel.
Stacks with such adjustments render for display this way; flattening a
plain Lab stack keeps its stored values (exports no longer clip).
Anything else (other blend modes, effects, other adjustments) keeps the
RGB compositor.
Photoshop oracles 145 -> 152: lab16 invert, posterize-4, levels, curves
and brightness-contrast-legacy; lab8 invert and posterize-4. lab8
levels and legacy B/C come within a level (their 8-bit rules are #5143
and #5132). psd-tools and io unchanged; every round trip unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Layers panel: Clip Thumbnails to Layer Bounds (#4281)
Add interface.clipThumbnailsToLayerBounds (off by default), toggled from the
Layers panel menu or Preferences > Interface. When on, a pixel layer's
thumbnail frames the layer's own pixels (cut to the canvas) instead of the
whole canvas, so small objects in a large document stay visible. Empty or
fully off-canvas layers keep the canvas frame.
The frame is computed only when the thumbnail is rebuilt (tile-cached
content bounds) and is part of the cache key, so toggling rebuilds it.
Closes#4281
* Regenerate docs/scorecard.md after merging main
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: badaaif <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Levels: whole levels at every step and Photoshop's shadow toe
In 8 bits Photoshop rounds each of the three Levels steps to a whole
level (half up): the input-range stretch (already modelled), the
midtone gamma and the output range. A narrow output range therefore
merges neighbouring gamma levels: psd-tools levels_rgb maps R 43..48 to
58 through a 0.47 channel gamma and the master's 48..255, where the
unrounded curve gave 56..59 (3 levels off in blue).
The gamma above 1 starts along a toe a*u - b*u^q of slope a = 2^gamma
that joins the power curve with the same value and slope (q = 2.2,
fitted). With the rounding, that is exact on 1116 of the 1118 levels
of gamma 1.16, 1.22, 1.3, 1.78 and 2.0 oracles; the soft-min it replaces
missed 52. Gamma below 1 stays the plain power curve (0.67 and 0.7:
every level exact).
Oracles: psd-tools levels_rgb 0.0118 -> 0.0039 (PASS); photoshop
rgb8/stack-levels-curves-hue-posterize 0.2000 -> 0.0000 (PASS); rgb8 and
gray8 levels.psd 0.0039 -> 0.0000; rgb16/gray16 0.0039 -> 0.0020. Floors:
photoshop 145 -> 146, psd-tools 236 -> 238 (main is at 237).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Empty commit to resync the pull request head
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Pen tool: Shift constrains segments and handles to 45° (TOOL-214-10, #3751)
Shift-click puts the next Pen anchor on the nearest 45-degree ray from the
previous anchor (at the pointer's projection onto it); Shift held while
dragging out direction handles turns them to 45-degree steps, read live
per move. The first anchor, closing on the first anchor and the last-anchor
break are unaffected; after Alt breaks the pair Shift constrains the
outgoing handle alone. The rubber band previews the constrained point.
Reuses stroke_constraint::snap45. Flips scorecard TOOL-214-10 to done.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Regenerate docs/scorecard.md after merging main
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Regenerate docs/scorecard.md after merging main
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Regenerate docs/scorecard.md after merging main
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Brush Settings: a response curve per dynamic control (#3294)
Every brush Dynamic (Size, Opacity, Flow, Roundness, Scatter, Count,
Depth, Foreground/Background...) now carries its own response curve for
its control (Fade, Pen Pressure, Pen Tilt, Stylus Wheel, Rotation),
applied after the global Preferences pen curve. It is part of the brush
JSON, so tools.setBrush sets it (e.g. shapeDynamics.size.curve), actions
record it and presets keep it; linear by default, so existing brushes
paint unchanged.
- paint: ResponseCurve, the global pen curve's monotone cubic moved down
from the engine as a fixed-size Copy type (no allocation per dab);
serialises as its point list. dynamics::dynamic_value applies it.
- engine: prefs::PressureCurve is now that type.
- ui-egui: each Control row gets a curve button (lit when the curve isn't
linear) that shows the same editor as Preferences > Tools.
- scorecard: TOOL-213-11 done.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n: Response curve for hu, tr
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs: regenerate scorecard after merging main
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Fix i18n after merge: Serbian Response curve, drop duplicate hu row; regenerate scorecard
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Regenerate docs/scorecard.md after merging main
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Brushes: import an image file as a brush tip (#3279)
`brush.presets.importImage` reads an image (path, or base64 data with
an optional fileName) through the normal importer, so any format and
bit depth File > Open reads works. Its composite becomes a sampled tip
the way Define Brush Preset reads a selection: darkness x alpha,
cropped to the painted extent, at most 2500 px a side. The bounds are
found band by band first, so a large mostly-white image never needs a
full-size buffer. The preset is named after the file (never replacing
an existing one) and becomes the current brush unless select is false.
Import Brushes... in the Brushes panel and the brush preset picker now
opens a picker for .abr and image files; images go to the new command,
.abr files to brush.presets.importAbr as before. Automation refuses an
ambient `path` for it like the other import commands.
Define Brush Preset and the new command share `sampled_brush`.
Scorecard TOOL-213-12 -> done.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Regenerate scorecard
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs: regenerate scorecard after merging main
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Regenerate docs/scorecard.md after merging main
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* fix(ui): remember the selected tool between launches
* fix(ui): preserve explicit Eraser selection during temporary tool overrides
Clear temporary pen and spring state for explicit toolbar, menu, preset and control selections. Keep spring-loaded key behavior; only permanent key choices clear the pen marker. Add exit/relaunch coverage for explicit Eraser with pen and spring overrides.
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Brandon Thomas <[email protected]>
* Batch: Include All Subfolders (AUTO-219-5, #3753)
File > Automate > Batch and File > Scripts > Image Processor take an
optional "includeSubfolders" (default false): a folder input also yields
the images in its subfolders, walked depth first in name order, bounded at
32 levels and 100000 files, with a seen-set of canonical folders against
symlink loops and the output folder left out. Each result keeps its
relative subfolder under the output folder, so same-named files in
different subfolders no longer collide. A non-boolean value is BadParams.
Both dialogs gain an "Include All Subfolders" checkbox (translated in
every complete catalog). Scorecard AUTO-219-5 stays partial (naming
template, progress/cancel, error log file); gaps.md G11 updated.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
* Batch: translate Include All Subfolders for Hungarian and Turkish
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The filter row's "Kind" was painted text. It is now a dropdown, as in
Photoshop: Kind keeps the five layer-kind toggles, Name lists layers whose
name contains the typed text (ignoring case), and Mode lists layers with
one blend mode (Pass Through included), filtering by Normal at first.
The criteria live in UiState (layer_filter_by, layer_filter_name,
layer_filter_mode; serde defaults, so saved settings still load) and are
drivable with ui.set {layerFilter: {by, kinds, name, mode}}, validated
before anything changes. ui-parity.md gets a partial row: Effect,
Attribute, Color, Smart Object, Selected and Artboard are still missing.
Co-authored-by: Brandon Thomas <[email protected]>
With a selection, ⌘T and the Move tool's Show Transform Controls framed the
layer's content clipped to the selection, so the box was smaller than the
marching ants wherever the layer did not fill the selection. The ants then
stayed where the selection started until the transform was applied.
- transform_frame / target_frame (engine): the box is the selection's own
bounds when a selection limits the transform. transform_bounds is unchanged,
so snapping and the command defaults are as before.
- The selection outline is drawn through the open box (homography, or the warp
in Warp mode), so it follows move, scale, rotate, skew, distort, perspective
and warp before the transform is applied. The whole outline is traced while
such a box is open, not only the part in view where it started.
- marching_ants_lines: ants along screen-space lines; non-finite lines are
skipped and very long ones are clipped to the view.
Part of #3402.
Claude-Session: https://claude.ai/code/session_015vKPoMXFhSvWmX5D9XM646
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The embedded and working profile names now sit on the choices they belong
to ("Use the embedded profile (Display P3)", "Convert to the working space
(sRGB IEC61966-2.1)") instead of two lines above them; the sentence naming
the file stays. The missing-profile prompt is unchanged.
"Remember this choice for future images" saves the answer as the Color
Settings policy for the file's mode (preserve, convert or off) and turns
Ask On Mismatch off, so files read from then on open without asking. Files
already waiting in the queue with a mismatch in the same mode get the same
answer (open_prompt::answer_remembering); a different question still asks.
The checkbox's tooltip says where to turn the prompt back on.
The catalogs drop the rows nothing uses any more ("Embedded: {embedded}" and
the two old choice labels) and add the four new strings in every language.
Co-authored-by: Brandon Thomas <[email protected]>
An action can take one of Photoshop's button colours, and the panel can switch to Button Mode so a click plays that action. The colour is stored on the action and is empty until it is set. The new labels are in every complete catalog.
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* feat(ui): Darken and Lighten Interface (Ctrl/Cmd+Shift+F1, F2)
Photoshop's Shift+Ctrl+F1 (⇧⌘F1 on macOS) darkens the interface by one of
its four brightnesses and Shift+Ctrl+F2 lightens it, stopping at the
darkest and lightest. PhotoCraft has the same four as themes (Pro, Pro
Medium Gray, Pro Light Gray, Pro Light), so view.darkenInterface and
view.lightenInterface step through them from the theme on screen and save
the choice like picking it in Preferences › Interface (setting the dark or
light theme and the matching appearance). From another theme they step
from the nearest brightness: Medium Gray for a dark one, Light Gray for a
light one. They work with no document open. Both labels are in every
catalogue.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* feat(i18n): Vietnamese strings for Darken and Lighten Interface
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* feat(i18n): Hungarian and Turkish strings for Darken and Lighten Interface
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n(sr): translate Darken and Lighten Interface
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The engine, renderer and PSD mapping already supported Smooth, Chisel Hard
and Chisel Soft, but the dialog's Bevel & Emboss page had no control for it.
Add the Technique dropdown under Style (Photoshop's order), a smooth default
for new bevels, and translations.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
* ACEScg: File Handling option opens colour-less EXR renders in the ACES working space
Our render pipelines write ACEScg (AP1 primaries, ACES white ~D60,
linear) and — like most — state no colour space in the EXR header, so
io's linear-sRGB default misrepresented them (saturated hues shifted).
- cms: AcesCg joins the built-in profiles, generated from the published
AP1 primaries and ACES white point with a linear TRC (id "aces-cg",
aliases acescg).
- engine: File Handling gets "Open color-less EXR as ACEScg". When on,
an EXR that io tagged with the linear-sRGB default opens as ACEScg
instead; embedded or EXIF-declared profiles stay untouched.
Translated in every catalog. ACES view transforms (OCIO-based display)
remain future work; this covers the document's source colours.
* scorecard: 26 of 158 (main's #3472 preference added one too)
* Regenerate docs/scorecard.md after merging main
---------
Co-authored-by: Brandon Thomas <[email protected]>
Clicking a snapshot now selects it (highlighted in place of the current
state) until the history moves on. With a snapshot selected the footer's
trash runs history.deleteSnapshot instead of history.deleteState, as in
Photoshop. Snapshot rows and past state rows can also be dragged onto
the trash to delete them.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
With always() the gate also ran when a newer push cancelled the run,
taking a runner slot only to report a red check on the superseded
commit. !cancelled() still runs it when a test leg fails (a skipped
required check would count as passing), but not in a cancelled run.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* raster: a shifted surface copies in one pass per row without pre-filled tiles
Surface::translated's misaligned path allocated every target tile through
Tile::filled (a per-pixel default fill when the default is nonzero)
before overwriting nearly all of it, and addressed spans through
per-row closures. The path now writes each stored row directly: content
spans copy as single memcpys crossing source tiles as needed, vacated
edges fill from one prebuilt default row, and filled() builds its
default rows in bulk. Same output (the read-back oracle covers every
depth, sign and tile crossing), cheaper for nonzero defaults and one
less pass over each tile.
* raster: a missing source tile in the shifted content reads the default pixel
The misaligned path skipped absent tiles inside the content rect,
leaving zeros where the surface's default pixel belongs (a white mask
with sparse tiles turned black after a move). The span now fills from
the prebuilt default row. echelon's snippet is a regression test, and a
randomized oracle compares the whole path against a per-pixel reference
gather across formats, default pixels, sparse tiles and tile-boundary
crossing shifts.
* ci: retrigger wasm32 (the full workspace wasm check is clean locally; the job failed before writing a log and the run is stuck queued)
* fix(windows): respect top and side taskbars in custom-frame maximize
* Keep work-area correction stable for narrow taskbar reservations
* style: apply rustfmt to taskbar window correction
* fix(window): repaint after custom-frame work-area correction
* Undo and redo Camera Raw adjustments from the keyboard.
Changing a slider and pressing Ctrl+Z did nothing, because the dialog blocked application shortcuts and kept no history of its own. Each finished edit, including one slider drag, is now one step. The control channel accepts ui.undo and ui.redo.
Refs #4100
* Format camera_raw_ui.rs (rustfmt)
---------
Co-authored-by: Brandon Thomas <[email protected]>
The canvas drew guides before transform_tool::draw_overlay, so the moving
pixels of a layer under Free Transform covered them. Split the overlay:
draw_content paints the preview (plain, styled or warped) and the layers
above it right after the document, under the grid and guides; draw_overlay
keeps the box, handles, reference point, warp grid and split guides over
them, as in Photoshop. The hand-pan drag path does the same.
Test: guides_draw_over_the_transform_preview_and_under_its_box checks the
painted order preview < guide < box; it failed before the fix.
* Layer styles: editing effects no longer re-enables a disabled effect list (#3648)
* Layer styles: follow #3648's exact dialog steps in a test
Preview off shows the document as saved (the in-dialog stroke disable is
a pending edit), so the state the issue reports is Photoshop's preview
semantics, not a resurrected effect: the live preview honours the
disable, and OK applies it.
* style: cargo fmt
On top of #5001. Two UI-thread stalls it leaves on a 16000 × 10000, 30-layer PSB
whose layers exceed the GPU memory budget:
- GpuCanvas::composite checked the memory budget for full refreshes only. Hiding
or showing a layer that covers most but not all of the canvas took the
synchronous GPU path, paging every layer in on the UI thread. Damage of
DEFER_PX or more now gets the same check, so it goes to the background redraw.
Small edits keep the GPU path without the check.
- The first draw after a background open scanned every layer's content bounds
on the UI thread (GPU planning). The open job now warms the shared tile bounds
cache before handing the document over (warm_layer_bounds, cancellable between
layers).
Co-authored-by: syncroot <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Interface: a preference to hide the Home screen's recent files (#4382)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* i18n: drop the hu/tr 'Enable animations' lines, #4362 adds them
Both PRs added the same two lines; whichever merged second would leave a duplicate key (bundled_catalogs_are_consistent rejects it).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs: regenerate scorecard after merging main
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Merge main; fix i18n (duplicate ja row, missing sr row) and regenerate scorecard
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Drop the duplicate hu row
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The Swatches panel used the shared preset browser grid, with a 4 px gap
between cells and a rounded, bordered card for each. Photoshop's panel
packs flat squares edge to edge. Place gets a `flush` flag: no gap, no
per-cell border (only the selected and hovered cells are outlined) and
square corners. Swatches sets it; gradients, patterns, styles and shapes
keep the gap, now the named GRID_GAP and GRID_INSET constants.
Test: swatches_touch_without_a_gap clicks 1 px past the first swatch and
expects the second one selected; with the gap that point was empty.
Quick Selection reused the Ellipse Marquee's dashed circle. It now has a
distinct glyph: Lucide's brush with a dashed selection arc around the
bristles (Photoshop's Quick Selection mark). A rasterized test asserts
the two icons read as distinct and that the arc actually renders.
Co-authored-by: highperfocused <[email protected]>
While the Info tab was visible, every frame ran
`view.colorSamplers.list` through the command dispatcher, compositing
each sampler point (up to 10) through the whole layer stack, with layer
effects building a 1-pixel effect-map entry each time.
sampler_readouts keeps the readouts for (document, revision). Adding,
moving or deleting a sampler and every pixel edit bump the revision,
so the readouts refresh exactly when they can change and are shared
between frames otherwise.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Hiding Edit › Menus (or Keyboard Shortcuts) left no way back into the
dialog, and Save, Save As, Exit and Preferences could disappear too.
edit.menus now refuses to hide them, the dialog greys their Visible box,
and the menus keep them even when a preferences file lists them.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The hex field's six-character limit dropped any paste into a full field
(every paste after the first) and any paste with a leading #. A pasted hex
colour now replaces the whole code.
Co-authored-by: kwesi <[email protected]>
#1322 lifted the number's slot 1 pt above the box's centre because digits
have no descenders. Measured on today's mono font at 4x, the digits' ink
already sits within 0.2 pt of centre without the lift; with it the ink is
0.9 pt high, the focus outline has 1 pt of gap above and 3 pt below, and the
digits start 1 pt above the unit. The slot is centred again.
Tests: value_field_digits_ink_is_centred_and_level_with_the_unit renders the
field (Pro and Studio) and checks the digits' ink is centred and starts level
with the "%"; value_field_number_slot_is_centred_in_its_box replaces the
1 pt-lift test, and the #2823 slim-box test now expects centring. Both new
assertions fail with the lift restored.
Co-authored-by: prekabreki <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
The options bar and the Properties panel's Character section both draw the
"type-font" family picker, and the Character one is drawn later. With its
menu closed it removed the shared hover key every frame, so the options bar's
hover never reached the canvas. Only the picker whose menu is open now clears
the key; follow_hover already drops it each frame once nothing writes it.
Test: options_bar_hover_previews_with_the_character_picker_also_shown runs
the whole app (Pro theme, both pickers on screen), hovers a font in the
options bar and checks the canvas shows it, then that leaving the row
restores it. It fails without the fix.
Co-authored-by: prekabreki <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Two monitors without ICC profiles (a common Linux desk, e.g. DP-4 and DP-5)
showed two identical "Display profile not used" cards at every launch. A
reading now posts one notice: one reason line per display, then the advice
line naming them all. Displays added by a later reading still get a card of
their own, and a reading with nothing new stays quiet, as before.
Claude-Session: https://claude.ai/code/session_01SvxNXnS4tcg6BHT9VG67t7
Co-authored-by: Claude Opus 5.5 <[email protected]>
The separate mip render targets work around reduced-preview corruption on Intel DX12 drivers (#892, verified on Intel UHD 620), but every DX12 adapter took them. Creating a scratch texture per level on every refresh costs about 1 ms of CPU time per brush dab on an NVIDIA DX12 adapter. Gate them to Intel under DX12; PHOTOCRAFT_SEPARATE_MIPS=1/0 forces either way. Part of #4515.
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Canvas: themes without a dot pattern draw no dots texture (#4542)
Switching the dark theme to Pro filled the empty workspace with red and
black blocks on the reporter's machine. The blocks repeat every 22 px,
the tile of the canvas-surround dots texture, so it is that texture
showing wrong contents.
Pro and ten other themes set canvas_dot to their canvas colour: they
have no dot pattern, yet still uploaded a texture for it and tiled it
over the whole pasteboard. dots() now returns None for them, so nothing
is uploaded or drawn. Studio, Studio Light and Solarized Dark, whose
dots differ from the canvas, are unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Contributors: credit Higgs32584 by GitHub profile name
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
The eyedroppers (the points/subtractPoints params, and clicks on the
canvas while the dialog is open) refused points outside the active
selection. Photoshop samples anywhere and only limits the result to the
selection, which Color Range already does; the refusal is gone.
On Windows with the automatic backend, GL adapters now rank after every
other adapter, WARP included. VMware SVGA3D over GL and GL-only old GPUs
drew a black window (#1788, #1647); WARP renders.
a_system_arabic_face_renders_both_samples took the first family whose
character map claims Latin and Arabic and expected its Arabic sample to
render. A font can map those characters to empty glyphs (for example
Black Mountain), so cargo test failed on machines that have one
installed. The test now renders every candidate and checks the first one
that actually draws an Arabic sample, and skips only if none does.
effect_maps hashed the whole layer (Debug of effects and vector mask, a walk
over every pixel and mask tile, every child of a group) on every call, i.e.
once per render tile per effect layer. The identity depends on the layer
only; memoise it in the render's Ctx for layers that are part of the
document being rendered (their addresses cannot be reused during the render).
Co-authored-by: Claude Opus 5.5 <[email protected]>
source_size_probe decoded every embedded file other than PSD/TIFF/PNG/JPEG
in full just to learn its size. photocraft_codecs::probe_size reads the
size from the same decoders' headers (WebP chunks, the image crate's
GIF/BMP/TGA/ICO/QOI/HDR headers, flat EXR part headers), with the same
limit checks and the EXIF quarter-turn swap, so it agrees with decode
wherever decode succeeds. Deep EXR, AVIF and anything else still decode.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Photoshop doesn't apply the shadows, midtones and highlights curves one after
another: per channel it builds a single Levels (black point from shadows,
white point from highlights, one midtone gamma 2^((midtones + (shadows +
highlights) / 2) / 100)), and a gamma > 1 starts with slope 2^gamma at black
instead of a pure power curve. Applying the tones in sequence let a lowered
shadows and a raised highlights slider compound their gammas (up to 85/255 off
on mixed settings), and the pure power curve over-brightened the shadows of
the issue's red +100 / magenta -100 repro by up to 16/255.
Measured against Photoshop on generated 16-bit ramps and pure colours:
worst error 0.8/255 (was 85/255), the issue's repro 0.02/255 (was 16.5).
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
* Record action layer selections by name and show their targets
* Use document pixel format in grouped-layer regression
* Document the named layer action UI with a synthetic capture
* PSD import: flat fallback respects the decode budget; checked sizes (#5200, #5202)
The flattened Bitmap/Indexed/Duotone path retried every composite_rgba8
error, including LimitExceeded, by allocating the refused RGBA buffer
(OOM abort on e.g. a 50000x50000 1-bit PSB). LimitExceeded now
propagates; the Multichannel fallback checks its output against the cap,
reserves fallibly, borrows the merged image, uses get for planes and
converts in TILE_SIZE bands. A failed composite now warns.
Checked usize arithmetic for alpha channel names, merged plane sizes and
the gradient map name length (overflowed on 32-bit targets).
Also: EXR part import drops a cache that never hit but cloned each
image; raw KRA tiles are borrowed; flat import detects the format once.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* PSD import: flat Background opacity check reads a band at a time (#5200)
flat_background read the whole surface into floats to see whether it is
opaque, undoing the banded conversion. Read TILE_SIZE row bands, stop at
the first transparent pixel, and drop the RGBA8 buffer before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
The layers were only in document.inspect, together with history,
channels and much more, so agents looked for a layer.* command that
didn't exist.
layer.list returns the layers top-down with id, name, kind, visibility,
bounds, selected and active; groups list their children, or with
"flat": true every layer is one row with its depth and parent group.
It takes the same optional "document" index as document.inspect, is not
journaled and bad params are errors.
Stretching a 1 px row or column faded out along the stretch: bicubic and
bilinear sampling read transparent pixels past the source edge, and with a
selection the edge was the whole layer's bounds (lifted pixels kept their
colour at zero alpha). Samples past the source rect now repeat the edge and
the outline is antialiased by coverage, and the lifted pixels' bounds are the
selected pixels'.
Fixes#3773
Co-authored-by: Fabio Lussi <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Smart Objects: inspect and Replace Contents show shared instances (#4075)
Placed layers that share a source have imported as one set of contents
since #2222, and Replace Contents changes every instance, but nothing
reported it: `linkGroup` is Link Layers, and `smartSource` can't tell
copies of one contents from independent copies of the same file.
document.inspect now gives each smart object `smartContents`: the
contents `id` its copies share and how many `instances` there are.
Replace Contents, Relink to File and Convert to Embedded/Linked list the
`instances` they changed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Test rendered caches after shared Smart Object replacement
---------
Co-authored-by: jonathnwxh <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Repeating a duplicate or paste of a 12 MP layer cost as much as the first time because the
GPU compositor's lower-stack cache stayed at the checkpoint of the first edit after it was
filled: an edit higher up only checked that the layers below that checkpoint were unchanged,
so every refresh re-ran all passes above it (in the #209 layered document, 87 of 110 passes
over the 12 MP damage rect; the duplicate itself takes 0.3 ms, its tiles are already shared).
Each cached chunk now remembers its own checkpoint. A render captures at the highest
checkpoint below the edit, also where it resumed from a lower cached one, so the next
refresh of that edit (its undo, redo or repetition) resumes just below it. Chunks a refresh
doesn't reach keep their lower prefix; a chunk moved by a partial capture holds only that
rectangle afterwards. Validity is unchanged: a chunk survives only while the layers below its
checkpoint are untouched.
Release perf_scenarios, layered group, Windows 11, RTX 5070 Ti, 7 interleaved runs:
P9 p50 22-29 -> 12-17 ms, P11 p50 37-41 -> 11-18 ms; P8 and P10 fall too; P1-P7, P12 unchanged.
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
The compositor kept every pattern texture it ever uploaded (RGBA32F,
16 bytes per pixel) until the device was lost: `forget_doc` and
`retain_docs` skipped them and no budget applied, so a long session
that opened and closed documents with pattern fills or pattern effects
only ever grew.
Each cached pattern now records the document that last painted with it
and the frame it last painted in:
- closing a document (`forget_doc`, `retain_docs`) frees its patterns;
- a pattern no frame has painted with for 240 frames is dropped;
- over PATTERN_BUDGET (256 MiB) the least recently painted go first,
never one the current frame paints with.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Offsets and counts stored in a profile reach u32::MAX, which is
usize::MAX on 32-bit targets such as wasm32. Several places added to or
multiplied them before any bounds check, so a malformed profile could
panic (overflow checks on) or read from a wrapped offset (checks off).
- curv: reject an over-long count before computing the element length.
- mAB/mBA: the matrix and CLUT read through one bounds-checked slice at
the stored offset instead of adding to the offset.
- mluc: the record offset is computed with checked arithmetic.
- XYZ: one bounds-checked slice instead of off + 4 and off + 8.
Co-authored-by: Claude Opus 5.5 <[email protected]>
The archive limits bound what one import extracts (256 MiB an entry, 1 GiB in
total), and one pixel layer is skipped past 64 megapixels, but nothing bounded
what the layers materialized out of that cost: each bitmap's declared
`BmpW`/`BmpH` sizes its RGBA output and its decode planes, with or without tile
data, so a small document with many layers could still allocate without end.
`Limits` gains `max_pixel_bytes` (1 GiB by default, the budget the other
importers allow decoded pixels). `Reader` admits a layer before `decode`
allocates it: the layer's RGBA joins the layers already materialized, and its
working planes are counted with them, since every earlier layer is still live
while this one decodes. Past the budget the import fails with `Error::Limit`,
which `photocraft-io` reports and falls back from to the embedded preview.
Tests: a synthetic 300-byte document declaring layers with no tiles at all - one
layer inside the budget imports, the second past it is rejected before
allocation, and a declared 2048x2048 layer is rejected under a 1 MiB budget.
The fuzz targets get a small pixel budget so mutations reach the new rejection
path.
Co-authored-by: DeepSeek <[email protected]>
* Linux: a virtual keyboard claiming BTN_TOOL_PEN no longer sends the window through Xwayland (#5284)
On Wayland the app opens its window through Xwayland when a pen is attached
(#639), and it took any input device whose key bitmap has BTN_TOOL_PEN for a
pen. A virtual keyboard made with uinput (RustDesk's, for one) claims every
key code, that one included, so the window went through Xwayland on machines
with no pen at all. On a fractionally scaled desktop (GNOME 46 at 125%) the
X11 window is drawn at the logical size and upscaled by the compositor, so
the UI came out blurred even though Auto UI Scale already followed the
system's 1.25 (#1237): native Wayland, where winit negotiates
wp_fractional_scale_v1 and renders at the physical 3840 px, was never used.
A pen now also needs ABS_X and ABS_Y in its absolute-axis bitmap, as
libinput asks of a tablet tool; a keyboard has no axes. Regression tests
with the virtual keyboard's own bitmaps, a pen beside it, and a pen display.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
* Docs: on a fractionally scaled Wayland desktop, Xwayland is blurred and native Wayland has no file drops (#5284)
Say where the trade-off is: a native Wayland window is sharp at 125%,
150% and 175% but takes no dropped files (#386); an Xwayland window takes
drops and pen input (#639) but is enlarged by the compositor. List the ways
to open through Xwayland and what counts as a pen.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Fable 5.1 <[email protected]>
* feat(commands): find the Elliptical Marquee by 'ellipse' or 'marquee' (#5060)
The Elliptical and Single Row/Column Marquee are options of select.rect
("ellipse":true, "line"), but the command-list filters only looked at
the id and the label ("Rectangular Selection"), so 'ellipse' and
'marquee' found nothing.
Commands can now carry search keywords. command.list / engine.commands
rows include them as "keywords", and the three filters (CLI commands
--filter, the engine.commands RPC and the MCP command_list) share one
matcher over id, label and keywords. select.rect gets the Photoshop
marquee names as keywords.
* Command list replies: document the optional keywords key
The reply-shape test from #3684 pins the exact keys of each row; select.rect
now also carries keywords, so the test and docs/control-protocol.md allow
that one optional string.
* Map .atn duplicate-into-new-document (Mk) to layer.duplicate
Photoshop records Layer > Duplicate Layer with Destination: New as a
Mk event (document class + layer Usng, Nm = new doc name, LyrN = copy
name), verified against a ScriptListener/XTensions export. Import maps
it to layer.duplicate {document: new}; export writes it back, so the
steps round-trip instead of becoming kept Photoshop steps that fail at
playback. A Dplc with a T document reference (another open doc) now
imports as a kept step rather than misplaying as a same-doc copy.
* Fix Rust formatting in ATN command tests
* Fix rustfmt formatting in ATN command tests
Photoshop's Gradient Map picks each pixel's place on the gradient by
0.3 R + 0.59 G + 0.11 B, the luminosity of the blend modes. We used the
Rec. 601 grey (0.299, 0.587, 0.114). That is a few thousandths off, but
it is enough to land two levels away on steep stretches of a gradient.
The photoshop corpus's rgb8 and rgb16 gradient-map.psd now pass (they
were 2 levels off on 0.8-1.7% of pixels). Sampling the gradient through
a 256- or 1024-entry table didn't help; the weights did.
CPU (compose::adjust) and GPU (the shader's lum) both change. Lab
documents keep their Gradient Map driven by L, and grey pixels map as
before.
Photoshop oracles 145 -> 147 of 258; floor 145 -> 147.
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Brandon Thomas <[email protected]>
The canvas|gpu|mip|type_tool name filter missed ui-egui unit tests that
build a Harness with a real wgpu device: background_redraw's GPU tests,
panel_docking's floating_group_titles_are_opaque_in_studio_themes and
swap_arrow's rendered_glyph_has_a_left_head_and_a_lower_right_head. Name
them in both overrides, so they share the 4-thread cap and get the CI
retries. The other .wgpu() harnesses only render when a fixture env var
is set, or are #[ignore]d.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
monitor_macos is harness = false, so nothing answered cargo-nextest's
discovery: `--list --format terse` ran the whole test and nextest failed
to parse its output on macOS. List the one test, and nothing for
`--ignored`, as mac_menu_appkit already does; `monitor_macos --exact`
still runs the body.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
selected_style reads the font engine through try_shared, which returns None while the system
font scan is still running. Under cargo test the test passed only because an earlier test in the
same process had already finished the scan; run on its own (nextest runs every test that way, and
`cargo test -- --exact` does too) it saw no engine and got "Regular" instead of "SemiBold". It now
waits for the scan with photocraft_text::shared(), as every_theme_shows_the_type_options does.
On main, alone, it failed before this and passes after, under both cargo test and nextest.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Splitting the Windows tests renamed their check from `test (windows-latest)` to one per leg. No
status check is required on main today, but if one is added under the old name it would wait
forever. A small job under that name now fails unless the platform pick and the whole test matrix
passed (`needs` gives one result for all legs, so a failing Linux leg fails it too). It checks the
result instead of exiting 0 under `if: always()`, which would pass with every leg red. Suggested
by Zach in review.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The test runs without a harness (AppKit needs the main thread), so nothing answered the `--list`
that cargo-nextest sends each test binary: on macOS the binary ran the whole test as its listing
and nextest failed on the output. It now lists itself in libtest's terse format, and nothing for
`--ignored`; a run passes `--exact <name>`, which falls through to the test as before. Found by
Zach running `cargo nextest run --workspace` on a Mac.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The integration tests run as one leg again. On the Windows runner (EPYC 7763, 3 runs) that leg
took 14.3-15.3 min, against 10.5-10.8 for each half, so the slowest leg is now the integration
one at ~15 min instead of the unit one at ~12.8. In exchange each PR that runs Windows takes two
Windows runners instead of three and ~7 fewer runner-minutes: every leg pays its own setup
(~1 min) and builds the workspace libraries its tests link against (~4 min even from sccache).
That matters more while PRs queue for the account's concurrent-job limit.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Comments and docs still said CI runs cargo test, that the craft-fonts tests run a second time in
the Linux job, and that CI's debuginfo matches the dev profile everywhere. development.md also
notes that test builds compile ui-egui unoptimized.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
nextest --partition splits the integration tests in two, so Windows runs three legs: on the
Windows runner (EPYC 7763, 2 runs) the integration halves took 9.8 and 10.8 min against 13.8 for
one, leaving the unit leg (~12.8) as the longest.
A ci nextest profile retries the GPU tests: on WARP, under load, one hit wgpu's poll timeout once
in ~40 runs while measuring. A test that passes on retry is reported as FLAKY.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
rust-cache keeps only the dependencies, so every run recompiled the workspace crates. sccache
caches their libraries (test executables can't be cached and still link). Warm, on the Windows
runner (EPYC 7763, 2-3 runs), the unit leg built in 7.1 min instead of 11.5 and the integration
leg in 4.3 instead of 8.8. Only main writes (PRs use SCCACHE_GHA_RW_MODE=READ_ONLY); a full set
of entries was ~90 MB.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Unit tests (--lib --bins, plus doctests) and integration tests (--test '*', plus Clippy and the
layering check) now run on two Windows runners at once. The feature selection is the same, so
both legs restore one dependency cache; the integration leg saves it, so Clippy's metadata stays
in it. On the Windows runner (EPYC 7763) the job took 28.3 min; the legs take 18.2 and 18.3,
running the same 6095 tests between them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Its library and unit tests were the biggest test-build units after engine's (12.6 of ~63 compile
minutes on the Windows CI runner), and its tests run no heavy pixel code of their own. On the
Windows runner (EPYC 7763, 3 runs vs 7): build 18.9 -> 15.4 min, test run 10.0 -> 10.5. Only test
builds change; cargo run and cargo build use the dev profile.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
They ran after the main tests in the Linux test job and added ~10 min to its critical path (and
were why that job ran out of disk). Now a two-leg job restores the test job's cache read-only
(shared-key: test) and saves nothing, so the cache budget (#2688) doesn't grow.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
With the binaries running in parallel, running the rest after a failure costs little, and a red
main no longer hides other failures behind the first one.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Writing PDBs for ~200 test binaries took 2.2 of ~21 build minutes on the Windows runner (3 runs
each on the same CPU model). A panic still reports its file and line; Linux keeps
line-tables-only for backtraces with lines.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
cargo test runs the ~200 test binaries one after another; on Windows that took 14 of the job's
37 minutes, 5 of them in gpu's parity tests on the software adapter. nextest runs them in parallel
with the GPU group from .config/nextest.toml. The examples were built and linked but never run;
Clippy's --all-targets still checks them. Doctests get their own step.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-11 01:41:45 +03:00
487 changed files with 39841 additions and 3673 deletions
if [ "$EVENT" != pull_request ] || [ "$FULL" = true ]; then
echo "os=$all" >> "$GITHUB_OUTPUT"
echo "legs=$all" >> "$GITHUB_OUTPUT"
echo "web=true" >> "$GITHUB_OUTPUT"
exit 0
fi
# The browser build (a fat-LTO wasm build) runs only when something it compiles or
# packages changed: the web app and its packaging, crate sources and manifests (the web
# build compiles nearly every crate), the workspace manifests and this workflow.
if git diff --name-only "$BASE" HEAD | grep -Eq '^(apps/photocraft-web/|packaging/web/|packaging/env\.sh$|crates/[^/]+/(src/|Cargo\.toml$|build\.rs$)|Cargo\.(toml|lock)$|rust-toolchain(\.toml)?$|\.github/workflows/ci\.yml$)'; then
if git diff --name-only "$BASE" HEAD | grep -Eq '^(apps/photocraft-web(-boot)?/|packaging/web/|packaging/env\.sh$|crates/[^/]+/(src/|Cargo\.toml$|build\.rs$)|Cargo\.(toml|lock)$|rust-toolchain(\.toml)?$|\.github/workflows/ci\.yml$)'; then
echo "web=true" >> "$GITHUB_OUTPUT"
else
echo "web=false" >> "$GITHUB_OUTPUT"
fi
if git diff --name-only "$BASE" HEAD | grep -Eq '^(apps/|crates/(ui-egui|text|gpu|tablet|heif)/|Cargo\.lock$|(.*/)?build\.rs$|\.github/workflows/ci\.yml$)'; then
# Intentionally tracked at @main (not SHA-pinned) so org-wide CI improvements, security fixes, and build optimizations propagate without PR churn across all craftapp repos.
| `crates/codecs/src/codecs/xpm/x11_colors.rs` | X11 `rgb.txt` color-name lookup table (normalized X.Org RGB values) | X.Org Foundation and X11 contributors | <https://cgit.freedesktop.org/xorg/app/rgb/tree/rgb.txt> | MIT (X.Org app/rgb) |
| `crates/ui-egui/src/i18n/el.tsv` | Greek UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `crates/ui-egui/src/i18n/de.tsv` | German UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `crates/ui-egui/src/i18n/es.tsv` | Spanish UI translations | @gorevyoneticisi; AI contributor: OpenAI Codex | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `crates/ui-egui/src/i18n/it.tsv` | Italian UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `crates/ui-egui/src/i18n/ko.tsv` | Korean UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `crates/ui-egui/src/i18n/nl.tsv` | Dutch UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
@@ -26,7 +27,7 @@ next to them.
| `crates/ui-egui/src/i18n/sr.tsv` | Serbian (Latin, Ekavian) UI translations | PhotoCraft contributors | Original translations of PhotoCraft's English labels | MIT OR Apache-2.0, [`LICENSE-translations.txt`](crates/ui-egui/src/i18n/LICENSE-translations.txt) |
| `assets/fonts/Inter-Regular.ttf`, `Inter-Medium.ttf`, `Inter-SemiBold.ttf` | Inter 4.001 (UI font) | The Inter Project Authors (Rasmus Andersson) | <https://github.com/rsms/inter> | SIL OFL 1.1, [`assets/fonts/OFL-Inter.txt`](assets/fonts/OFL-Inter.txt) |
| `assets/icons/*.svg` except `slice-knife.svg`, `sampler.svg`, `artboard.svg`, `perspective-crop.svg`, `frame-placeholder.svg`, `gradient.svg`, `type-mask.svg`, `type-mask-vertical.svg`, `clip-below.svg`, `clip-release.svg`, `layer-mask.svg`, `adjustment-layer.svg`, `brush-replace.svg`, `eraser-background.svg`, `eraser-magic.svg`, `lasso-magnetic.svg`, `direct-select.svg`, `convert-point.svg`, `pen-add-anchor.svg`, `pen-delete-anchor.svg`, `pen-freeform.svg`, `pen-curvature.svg`, `bandage-sparkle.svg`, `painting-symmetry.svg`, `symmetry-*.svg` and `marquee-{row,column}.svg` (113 files; `star-filled.svg` is `star.svg` with a solid fill) | Lucide icons | Lucide Icons and Contributors | <https://github.com/lucide-icons/lucide> (`icons/<name>.svg`; `align-*` are now named `text-align-*` upstream) | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/*.svg` except `slice-knife.svg`, `sampler.svg`, `artboard.svg`, `perspective-crop.svg`, `frame-placeholder.svg`, `gradient.svg`, `type-mask.svg`, `type-mask-vertical.svg`, `clip-below.svg`, `clip-release.svg`, `layer-mask.svg`, `adjustment-layer.svg`, `brush-replace.svg`,`quick-selection.svg`,`eraser-background.svg`, `eraser-magic.svg`, `lasso-magnetic.svg`, `direct-select.svg`, `convert-point.svg`, `pen-add-anchor.svg`, `pen-delete-anchor.svg`, `pen-freeform.svg`, `pen-curvature.svg`, `bandage-sparkle.svg`, `painting-symmetry.svg`, `symmetry-*.svg` and `marquee-{row,column}.svg` (113 files; `star-filled.svg` is `star.svg` with a solid fill) | Lucide icons | Lucide Icons and Contributors | <https://github.com/lucide-icons/lucide> (`icons/<name>.svg`; `align-*` are now named `text-align-*` upstream) | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/{check,chevron-down,chevron-left,chevron-right,chevron-up,chevrons-left,chevrons-right,circle,clock,compass,info,link,link-2,lock,minus,monitor,moon,move,navigation,plus,search,square,trash,triangle,type,x,zoom-in}.svg` | Lucide icons derived from Feather (subset of the row above) | Cole Bemis (Feather), Lucide Contributors | <https://github.com/feathericons/feather> via Lucide | MIT (Feather) and ISC (Lucide), [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/slice-knife.svg` | Slice tool glyph | PhotoCraft contributors | Original work | MIT OR Apache-2.0 |
| `assets/icons/sampler.svg` | Color Sampler tool glyph: a target reticle marking a sampled pixel | PhotoCraft contributors | Original work | MIT OR Apache-2.0 |
@@ -41,6 +42,7 @@ next to them.
| `assets/icons/gradient.svg` | Gradient tool glyph: a rounded rectangle filled with a left-to-right linear ramp, drawn on Lucide's 24 px grid and stroke | PhotoCraft contributors | Original work | MIT OR Apache-2.0 |
| `assets/icons/eraser-background.svg`, `assets/icons/eraser-magic.svg` | Background Eraser and Magic Eraser tool glyphs: Lucide `eraser` plus added marks | Lucide Icons and Contributors; marks by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/eraser.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/brush-replace.svg` | Color Replacement tool glyph: Lucide `brush` plus two swap arrows | Lucide Icons and Contributors; arrows by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/brush.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/quick-selection.svg` | Quick Selection tool glyph: Lucide `brush` plus a dashed selection arc around the bristles | Lucide Icons and Contributors; arc by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/brush.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/bandage-sparkle.svg` | Remove tool glyph: Lucide `bandage` moved down, plus a sparkle mark | Lucide Icons and Contributors; mark by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/bandage.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/direct-select.svg` | Direct Selection tool glyph: Lucide `mouse-pointer-2` plus an anchor square | Lucide Icons and Contributors; square by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/mouse-pointer-2.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
| `assets/icons/pen-freeform.svg`, `assets/icons/pen-curvature.svg` | Freeform Pen and Curvature Pen tool glyphs: Lucide `pen-tool` scaled down, plus a freehand squiggle or a curve through two points | Lucide Icons and Contributors; marks by PhotoCraft contributors | Derived from <https://github.com/lucide-icons/lucide> `icons/pen-tool.svg` | ISC, [`assets/icons/LICENSE-lucide.txt`](assets/icons/LICENSE-lucide.txt) |
@@ -55,6 +57,7 @@ next to them.
| `crates/ui-egui/src/theme.rs` (Solarized Dark palette) | Solarized colour scheme used by the Solarized Dark theme | Ethan Schoonover | <https://ethanschoonover.com/solarized/> | MIT, [`LICENSE-solarized.txt`](crates/ui-egui/src/LICENSE-solarized.txt) |
| `crates/ui-egui/src/theme.rs` (Breeze Light and Breeze Dark themes) | The Breeze Light and Breeze Dark themes are original colour values on PhotoCraft's Studio layout, chosen to look close to KDE Plasma's Breeze (cool blue-greys, blue accent, 5 px frame radius) so the app blends into a Plasma desktop. None of Breeze's colour values, files or code are copied; the values are a few levels away from its schemes and tuned for text contrast | PhotoCraft contributors | Inspired by the look of KDE Breeze, <https://invent.kde.org/plasma/breeze> (its colour schemes are LGPL-2.0-or-later) | Original work |
| `crates/ui-egui/src/theme.rs` (Soft Dark and Soft Light themes) | Original colour values and radii for two low-contrast themes on PhotoCraft's Studio layout; no third-party palette, file or code is used | PhotoCraft contributors | Original | Original work |
| `crates/algo/data/libfacedetection.bin` (face detection weights, read by `crates/algo/src/faces.rs`, whose network code is a port of the library's) | libfacedetection CNN (YuNet design) weights and inference code, weights converted to half precision | Shiqi Yu | [ShiqiYu/libfacedetection](https://github.com/ShiqiYu/libfacedetection) (`src/facedetectcnn-data.cpp`, `facedetectcnn.cpp`, `facedetectcnn-model.cpp`) | BSD-3-Clause, [`LICENSE-libfacedetection.txt`](crates/algo/data/LICENSE-libfacedetection.txt) |
### Optional build input: craft-fonts (not files in this repo)
@@ -112,6 +115,6 @@ next to the files:
| `corpus/exr/` (5 files, 2.3 MB) | The deep OpenEXR test images `11`, `42`, `64`, `multivariate` and `objectid.deep.exr` | Contributors to the OpenEXR Project | [openexr `src/test/bin/test_images`](https://github.com/AcademySoftwareFoundation/openexr/tree/main/src/test/bin/test_images) | BSD-3-Clause |
| `corpus/affinity/` (60 files) | Public Affinity documents and PNG references: `vector-art/` (9 documents), `afdesignload/` (9), `jac21/` (2), `asset-store-template/` (1), and `affinity-samples/` (18 CC0 documents plus exported references) | samuel-etver; Nick Beeuwsaert; Jeremy Cantu; DAIKI (eviltwo); satoshoe-dev | [vector-art `simple/`](https://github.com/samuel-etver/vector-art/tree/255f8add3c8f0740196e22bd59502b811b532f0b/simple), [AFDesignLoad `testDesigns/`](https://github.com/NickBeeuwsaert/AFDesignLoad/tree/a18dd50a7079fb861a28835eeefa0d015f03f4a1/testDesigns), [Jac21/Branding `Logos/JC/DesignerFiles/Affinity`](https://github.com/Jac21/Branding/tree/57ae3f45bf4637f6927c0b07de25b922c55f944f/Logos/JC/DesignerFiles/Affinity), [AssetStoreTemplate `AssetStoreTemplate/`](https://github.com/eviltwo/AssetStoreTemplate/tree/f671981ee89d15526285b3cc87b4393d3defe644/AssetStoreTemplate), [affinity-samples](https://github.com/satoshoe-dev/affinity-samples/tree/96aca6bd86bbe7a677f1cf0630bd832d23943be8) | CC0-1.0 (vector-art and affinity-samples); MIT, Copyright (c) 2015 Nick Beeuwsaert; MIT, Copyright (c) 2018 Jeremy Cantu; MIT, Copyright (c) 2024 DAIKI |
| `corpus/pngsuite/` | PngSuite | Willem van Schaik | <http://www.schaik.com/pngsuite/> | Public domain |
| `corpus/pixls/` (7 files, 226 MB) | Real camera raws, one per decode path: Canon PowerShot SX50 HS (CR2 and its DNG conversion), Nikon D3 uncompressed NEF, Sony DSC-RX0 cRAW ARW, Panasonic DC-G9 high-res RW2, Olympus E-1 GRBG ORF, plus the E-5 packed ORF as the documented known-unsupported case. Per file (each CC0, per its database entry): [IMG_4059.CR2](https://raw.pixls.us/data/Canon/PowerShot%20SX50%20HS/IMG_4059.CR2), [CRW_4061.DNG](https://raw.pixls.us/data/Canon/PowerShot%20SX50%20HS/CRW_4061.DNG), [JD1_8203.NEF](https://raw.pixls.us/data/Nikon/D3/JD1_8203.NEF), [DSC00009.ARW](https://raw.pixls.us/data/Sony/DSC-RX0/DSC00009.ARW), [P1000475.RW2](https://raw.pixls.us/data/Panasonic/DC-G9/P1000475.RW2), [E_1__C106743_gredos.ORF](https://raw.pixls.us/data/Olympus/E-1/E_1__C106743_gredos.ORF), [_7061961_copy.ORF](https://raw.pixls.us/data/Olympus/E-5/_7061961_copy.ORF) | The photographers of the pixls.us raw samples database | [raw.pixls.us](https://raw.pixls.us) `data/`, fetched per [xtask/src/corpus_pins.rs](xtask/src/corpus_pins.rs) | CC0 / public domain, per file as linked |
| `corpus/pixls/` (11 files) | Real camera raws, one per decode path: Canon PowerShot SX50 HS (CR2 and its DNG conversion), Nikon D3 uncompressed NEF and Z6 paired 12/14-bit lossy and packed NEFs, Sony DSC-RX0 cRAW ARW, Panasonic DC-G9 high-res RW2, Olympus E-1 GRBG ORF, plus the E-5 packed ORF as the documented known-unsupported case. Per file (each CC0, per its database entry): [IMG_4059.CR2](https://raw.pixls.us/data/Canon/PowerShot%20SX50%20HS/IMG_4059.CR2), [CRW_4061.DNG](https://raw.pixls.us/data/Canon/PowerShot%20SX50%20HS/CRW_4061.DNG), [JD1_8203.NEF](https://raw.pixls.us/data/Nikon/D3/JD1_8203.NEF), [DSC00009.ARW](https://raw.pixls.us/data/Sony/DSC-RX0/DSC00009.ARW), [P1000475.RW2](https://raw.pixls.us/data/Panasonic/DC-G9/P1000475.RW2), [E_1__C106743_gredos.ORF](https://raw.pixls.us/data/Olympus/E-1/E_1__C106743_gredos.ORF), [_7061961_copy.ORF](https://raw.pixls.us/data/Olympus/E-5/_7061961_copy.ORF), [DSC_0750.NEF](https://raw.pixls.us/data/Nikon/Z%206/DSC_0750.NEF), [DSC_0751.NEF](https://raw.pixls.us/data/Nikon/Z%206/DSC_0751.NEF), [DSC_0753.NEF](https://raw.pixls.us/data/Nikon/Z%206/DSC_0753.NEF), [DSC_0754.NEF](https://raw.pixls.us/data/Nikon/Z%206/DSC_0754.NEF) | The photographers of the pixls.us raw samples database | [raw.pixls.us](https://raw.pixls.us) `data/`, fetched per [xtask/src/corpus_pins.rs](xtask/src/corpus_pins.rs) | CC0 / public domain, per file as linked |
Files copied into `corpus/` by hand (tiff, exr, raw) must be MIT, BSD or CC0.
@@ -16,6 +16,7 @@ Inter font (assets/fonts/Inter-*.ttf) — SIL Open Font License 1.1. See assets/
JetBrains Mono font (assets/fonts/JetBrainsMono-Regular.ttf) — SIL Open Font License 1.1. See assets/fonts/OFL-JetBrainsMono.txt.
Lucide icons (assets/icons/*.svg, except slice-knife.svg) — ISC; icons derived from Feather also MIT. See assets/icons/LICENSE-lucide.txt.
SCOWL English word list (assets/dict/en_US-scowl-50.txt.gz) — SCOWL license. See assets/dict/LICENSE-SCOWL.txt.
libfacedetection face detection weights and network (crates/algo/data/libfacedetection.bin, crates/algo/src/faces/cnn.rs) — BSD-3-Clause, Copyright (c) 2015-2019, Shiqi Yu, all rights reserved. See crates/algo/data/LICENSE-libfacedetection.txt.
Builds made with craft-fonts (CRAFT_FONTS_DIR; all official releases) also embed the fonts listed in
https://github.com/storytold/craft-fonts/blob/main/ATTRIBUTION.md (today BIZ UDPGothic, Shippori Mincho and
@@ -277,7 +277,7 @@ flatpak run ai.storyteller.photocraft
The AppImage needs no install: the first run registers its launcher icon and menu entry in `~/.local/share` so the dock shows PhotoCraft's icon on Wayland. Set `PHOTOCRAFT_NO_DESKTOP_INTEGRATION=1` to skip that, and see [`docs/releasing.md`](docs/releasing.md) › Linux to undo it.
On a Wayland session, files dropped on the window don't open yet: winit 0.30, the windowing library under egui, has no Wayland drag and drop (#386). Use File › Open, or copy the image in your file manager and paste it with Ctrl+V. To drag and drop, start PhotoCraft under XWayland: `WAYLAND_DISPLAY= photocraft`, `WAYLAND_DISPLAY= ./photocraft-<version>-linux-x86_64.AppImage`, or `flatpak run --nosocket=wayland --socket=x11 ai.storyteller.photocraft`. On Wayland, PhotoCraft shows the command for your install in a notice until you dismiss it.
On a Wayland session, files dropped on the window don't open yet: winit 0.30, the windowing library under egui, has no Wayland drag and drop (#386). Use File › Open, or copy the image in your file manager and paste it with Ctrl+V. To drag and drop, start PhotoCraft under XWayland: `WAYLAND_DISPLAY= photocraft`, `WAYLAND_DISPLAY= ./photocraft-<version>-linux-x86_64.AppImage`, or `flatpak run --nosocket=wayland --socket=x11 ai.storyteller.photocraft`. On Wayland, PhotoCraft shows the command for your install in a notice until you dismiss it. Preferences › Performance › Linux display server = X11 does the same on every launch. With a fractional display scale (125%, 150%, 175%) an XWayland window is drawn at a lower resolution and enlarged by the compositor, so the interface is blurred there; native Wayland stays sharp (#1072).
On macOS, the command-line tool comes as `photocraft-cli-<version>-macos-universal.zip`. The binary is signed with the same Developer ID as the app and notarized by Apple. A bare binary can't carry a stapled notarization ticket the way the DMG does, so the first time you run it macOS checks the notarization online. You can confirm it yourself:
@@ -300,6 +300,19 @@ Maintainers: [`docs/releasing.md`](docs/releasing.md) explains how releases are
> [!IMPORTANT]
> **Status:** PhotoCraft is in early alpha, and we want to be straight about where it stands: much of Photoshop's feature surface exists in some form, but **it is not yet a Photoshop replacement for daily professional work**. The biggest gaps are AI/generative features, 15 missing tools, depth in typography and pro workflows, and plug-in compatibility. Every Photoshop menu item is wired to a command ([`docs/parity-checklist.md`](docs/parity-checklist.md)), but that measures wiring, not behaviour. The honest, dimension-by-dimension picture and where we're going next are in the [roadmap](ROADMAP.md) and [parity assessment](docs/target-app-parity.md). Expect rough edges, and please file issues (include your OS, document size, layer count and a screenshot). You can also tell us what broke on [Discord](https://discord.gg/artcraft).
## Community packages
These installers are maintained by community members, not by the PhotoCraft team.
On Gentoo, the [::snakebyte overlay](https://github.com/switch87/snakebyte-overlay) packages the Linux release as `media-gfx/photocraft-bin`:
web_sys::console::error_2(&JsValue::from_str("photocraft-web-boot: could not start PhotoCraft:"),&error);
letSome(&wait)=RETRY_AFTER.get(attempt)else{
say("PhotoCraft couldn't be downloaded. Check your connection, then try again: ",true);
return;
};
say("Loading PhotoCraft\u{2026} (the download failed, trying again)",false);
sleep(wait).await;
attempt+=1;
}
});
}
}
#[cfg(target_arch = "wasm32")]
fnmain(){
boot::start();
}
#[cfg(not(target_arch = "wasm32"))]
fnmain(){
eprintln!("photocraft-web-boot only runs in the browser: packaging/web/package.sh builds it with `trunk build --release` in apps/photocraft-web-boot");
@@ -38,7 +38,7 @@ let out = encode(&img, Format::Tiff, &EncodeOptions::default())?;
| WebP | yes | yes | U8 | RGB, RGBA | yes | yes | yes | yes | no | no | yes (VP8, when `webp_lossless` is off) | `image-webp` / built-in VP8 |
| GIF | yes | yes | U8 | RGBA | 1-bit | no | no | no | no | no | yes (256-colour palette) | `image` |
| BMP | yes | yes | U8 | RGB, RGBA | yes | no | no | no | no | no | no | `image` |
| TGA | yes | yes | U8 | Gray, GrayA, RGB, RGBA | yes | no | no | no | no | no | no | `image` |
| TGA | yes | yes | U8 | Gray, GrayA, RGB, RGBA | yes | no | no | no | no | no | no | `image` (read) / built in (RLE write, one scanline per packet) |
| ICO | yes | yes | U8 | RGBA (at most 256×256) | yes | no | no | no | no | no | no | `image` |
| DDS | yes | **no** | U8 | Gray, GrayA, RGB, RGBA | yes | no | no | no | no | no | n/a | built in |
| Netpbm (PBM/PGM/PPM/PAM/PFM) | yes | yes | U8, U16, F32 (float: PFM gray/RGB only) | all six (PAM `CMYK`/`CMYK_ALPHA`) | yes (PAM) | no | no | no | no | no | no | built in |
@@ -47,6 +47,7 @@ let out = encode(&img, Format::Tiff, &EncodeOptions::default())?;
| Radiance HDR | yes | yes | F32 | RGB | no | no | no | no | no | no | yes (RGBE) | `image` |
| FITS | yes | yes | U8, U16, F32 | Gray, RGB | no | no | no | no | no | yes (header cards) | no | built in |
| DPX | yes | yes | U8, U16 (10/12-bit code values read and written exactly through U16) | Gray, RGB, RGBA | yes | no | no | no | no | DPX header fields only (creator, project, copyright, creation time, transfer, colorimetric, frame position, frame rate, time code) | no | built in |
| Cineon | yes | **no** | U16 (10-bit code values) | Gray, RGB | no | no | no | no | no | no | n/a | built in |
| AVIF | **no** | only with feature `avif` | U8 | RGB, RGBA | yes | no | no | no | no | no | yes | `image` + `ravif` |
| HEIF/HEIC | with feature `heif` | **no** | U8, U16 (10/12-bit decodes to U16) | RGB, RGBA | yes (auxiliary alpha) | yes (`colr` prof) | yes | yes | no | no | n/a | `heic-rs` |
@@ -93,7 +94,15 @@ embedded table (and a PNG's transparent entry) so `photocraft-io` can open them
big-endian, one unencoded element, at `EncodeOptions::dpx_bits` (8/10/12/16; by default 8 for
U8 and 16 otherwise), packing 1 for 10 and 12 bits; gray with alpha is written as luma.
Other descriptors, signed or run-length-encoded data and other bit depths return
`CodecError::Unsupported` naming the field. Cineon (`.cin`) is not supported.
`CodecError::Unsupported` naming the field.
* **Cineon (read-only).** Kodak's 10-bit logarithmic film-scan format (the one DPX grew from),
read from the same Kodak header family: the generic 1024-byte header, big-endian, one (Gray)
or three (RGB) channels, three 10-bit samples packed into each 32-bit word, lines on a word
boundary. The 10-bit code values open as U16 `round(c × 65535 / 1023)`; the file's black/white
reference points are a log-conversion choice and are not applied. A depth other than 10 bits or
a channel count other than 1/3 returns `CodecError::Unsupported`; truncation errors rather than
panicking. There is no Cineon writer (`ASYMMETRIC_EXCEPTIONS`); save as DPX, TIFF or EXR.
* **Camera raw files** (DNG, CR2, NEF, ARW… which are TIFF-structured) are recognised and refused
with `CodecError::Unsupported`: they are sensor data, not flat images. `photocraft-raw` decodes
"{what} of a {tile_w}x{tile_h} deep tile does not decode at the clipped size, and at the full tile size it needs {need} bytes with its scratch copy, exceeding max_alloc {}",
limits.max_alloc
))
})?,
};
// The table is cumulative within each row of the block (a tile restarts at every row);
// `prefix` is cumulative over the whole block, and each pixel's own count is the difference.
Loaded 100 of 487 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Hamza Naveed
