On wasm32 a near-1 GiB .abr arrives as base64 and is copied several times
before decoding; with a 1 GiB sample budget the import could pass 4 GiB and
abort the tab instead of failing cleanly. Pick the decoder budget by pointer
width (like MAX_DECODED_BYTES) and derive the engine's file cap from it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Import Brushes refused .abr files over 512 MiB: the engine's preset file cap
and the .abr decoder's sample budget were both 512 MiB. Raise both to 1 GiB
for brush files only; gradient, swatch, action and image-tip imports keep the
512 MiB cap. A compile-time assert keeps the reader and decoder caps equal.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
photocraft_psd::fixtures (feature testgen) generates the FILE-216-2
fixtures at test time: a 20 MP layered PSD and a 150-layer A4 layout
(groups, masks, hidden layers, drop shadows, known merged colours), a
2.02 GB PSD just under Photoshop's cap and the 4-5 GB PSBs, streamed row
by row. io tests/large_files.rs opens, checks, saves and re-opens each
(the 2 GB PSD and 4.38 GB PSB are opt-in). The ui-egui large_file_bench
times opening the big two, wired to P20 and a new P60.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* raw: fit non-DNG colour to the camera's embedded JPEG
* raw look: clippy (as_chunks) and rustfmt after merging main
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Opening the preset store decoded every imported tip at full resolution, and
each tip then lived twice: in the session's library and in the store's
snapshot of the last sync. A generated 600-brush library (16-bit tips,
800-2000 px, 2.27 GB decoded) cost about 4.5 GB after import and again after
every restart.
Library presets now hold their bitmaps as StoredTile references with a small
saved preview (at most 96 px): photocraft-paint gains TipShape::Stored and
Pattern::Stored (rendered with the full bitmap when loaded, else the preview),
and the store writes each tip's size and preview with its group. Opening the
store reads previews only; once a group is written its presets drop their full
bitmaps. A tip is decoded when its brush is used (picking a preset, a stroke
naming one, exporting) into a 256 MB least-recently-used cache. A loaded stored
tip serialises as `sampled`, so tool presets, actions and brush.get keep the
same JSON. Stores written before this change have no previews: their tips are
decoded once at the next start, eight at a time, and the groups rewritten. A
missing or corrupt tip file is now reported when the brush is used.
Measured with crates/engine/examples/abr_memory.rs on the 600-brush library
from crates/io/examples/gen_abr_library.rs (Windows, release):
after import 4565 -> 38 MB working set, after restart 4551 -> 20 MB,
after using 5 brushes 4560 -> 37 MB; store open 1.47 s -> 0.05 s.
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* PSD plane shuffles in bands; the document histogram computes off the UI thread
interleave/deinterleave walked every pixel of a 20 MP image on one
thread during PSD open and save. Both now split the rows into one band
per core (the file's own par_map/bands helpers) and merge the band
results in order: 20 MP RGB u16 interleave 152 -> 101 ms, deinterleave
461 -> 160 ms (2.9x); f32 rgba 329/730 -> 220/396 ms. Behaviour is
pinned by a parity test against the sequential reference at odd pixel
counts (band boundaries), all depths, with a missing plane in the mix.
The Histogram panel computed its payload inline in the frame every
250 ms while stale - a 41.5 ms UI-thread stall measured on a 24 MP
document, on top of brush work. compute_histograms now runs on a
worker thread; the panel keeps showing the previous histogram with the
existing cached-data warning until the result lands, and a stale
result cannot overwrite a newer revision. Pinned by a panel test that
drives one egui frame and asserts the job is scheduled, delivers off
thread and respects revisions.
* Histogram panel: compute inline when no thread can be spawned; drop the benchmark test
std::thread spawn fails on wasm32 (and can fail natively), which left
hist_job waiting on a disconnected receiver forever. Without a worker the
histogram is now computed inline and delivered through the same channel, a
disconnected job is cleared, and the perf span is measured in the worker.
Removes the dead t0 and the wall-clock benchmark test; adds a dead-worker test.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
The Photopea screenshots, their ATTRIBUTION row and the example that made
them documented the bug for review; the unit tests in crates/text cover
the behaviour, so they are not needed in the tree.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UHgFEHEsefoQoBParH47JA
* PSD: decode planes up to 8 GiB, so large PSBs keep their pixels (#375)
MAX_DECODED_BYTES capped a single decode at 2 GiB. Real PSBs go past
that in one plane set: the merged image of a 30000x30000 RGB file is
2.7 GB, a 33000x33000 16-bit layer channel 2.2 GB. Opening such a file
gave a document with no layers (flattened files) or layers whose
channels were "treated as empty" (layered files). Saving hit the same
cap on the encode side and silently fell back to uncompressed data.
The cap is now 8 GiB on 64-bit targets and 2 GiB elsewhere, the budget
photocraft-codecs already uses for Limits::max_alloc. It stays a
backstop: the decoders bound their output by their input on their own
(RLE expands at most 64x, Raw needs every byte present). The Raw and
RLE output buffers are reserved with try_reserve_exact, so a machine
without that much memory gets an error instead of an abort.
generated with Claude Code
* PSD import: a channel that can't be decoded leaves its layer empty (#375)
The warning says the channel is "treated as empty", but the importer
filled a missing transparency channel with opaque alpha, so a layer
whose channels all failed became an opaque black rectangle over the
document. A channel that is present but undecodable now leaves the
whole layer empty (transparent), as the warning says.
tests/large_psb.rs: the damaged-channel case (PSD and PSB), and an
opt-in release test that opens a flattened 29000x29000 PSB whose
merged image decodes to 2.52 GB (16 s, 11.6 GB peak here).
generated with Claude Code
* Large-file fixtures: generate and check 4-5 GB PSBs (#375, #216)
photocraft-psd example big_psb writes five PSBs straight from the spec,
independent of this crate's writer and its limits: flattened and
layered, raw (4.4 and 4.8 GB files, data past 4 GiB) and RLE (small
files whose planes decode past 2 GiB, up to a 33000x33000 16-bit
layer). Every sample follows one formula, so photocraft-io example
big_psb_check opens a file through photocraft_io::import and checks
pixels without a reference; with SAVE set it saves as PSB and checks
the re-opened file.
Release build, this machine: all five open with correct pixels and
survive a PSB save and re-open (5-14 s per import). Before the 8 GiB
cap, three of the five opened empty or black.
Scorecard FILE-216-2 missing -> partial: no 2 GB PSD, 150-layer or
20 MP fixture yet, and nothing runs in CI.
generated with Claude Code
* psd: ZIP channels reserve fallibly and read in bounded chunks
A small ZIP channel declaring a huge decoded size could make
zip_decompress try a multi-GiB with_capacity/read_to_end and abort on
allocation failure. Reserve with try_reserve_exact/try_reserve, read
through take(expected) in 1 MiB chunks, and report LimitExceeded
instead. Tests: absurd declared sizes, past MAX_DECODED_BYTES, garbage.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* psd: .abr brush (v1/v2/v6+) and .grd gradient (v5) parsers and writers
Clean-room from public format descriptions: legacy computed/sampled brushes,
8BIM-sectioned samp/patt/desc files, 8/16-bit raw and PackBits tips. Bounded
(MAX_EDGE, MAX_TOTAL_BYTES), damaged brushes are skipped with warnings, hostile
input returns Err. Adds an abr fuzz target.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* paint: 36 grouped built-in brush presets and new procedural tips
General / Dry Media / Wet Media / Special Effects groups (pressure size/opacity
variants, airbrush, pencils, charcoal, pastel, crayon, rake, watercolour, ink,
oil, foliage, grass, stipple, sparkle, halftone…). Leaf, grass, charcoal,
sponge, star and rake tips are generated in code. BrushPreset gains a group.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* io: map .abr brushes onto BrushSettings
Tip shape, shape dynamics, scattering, texture (embedded patt patterns),
dual brush, colour dynamics, transfer, pose, noise, wet edges, build-up and
captured tool options; unmapped settings are reported as warnings. Includes
bench_abr (100-brush import, 5000-dab stroke).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* engine: brush.presets.importAbr and gradient.presets.importGrd
Path or base64 data, group per file (re-import replaces), unique names next to
the built-ins; graceful-failure tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ui: real stylus pressure/tilt, grouped Brushes panel, Import Brushes
Pen force from egui Touch events (Windows WM_POINTER via winit), Pointer Events
pressure/tilt/twist on the web, tilt via ui.pointer; mouse stays 1.0. .abr/.grd
open from File > Open, drag-and-drop, Brushes panel and Preset Manager. Preset
previews keyed on a cheap signature instead of the full brush JSON.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
CI's `cargo fmt --all -- --check` has failed on every run because the
repo had no rustfmt config and was never formatted. Use the same settings
as EffectCraft/FilmCraft (max_width 160, use_small_heuristics Max), which
also keeps the churn small. Formatting only; no code changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aiongg