Profile::content_hash is a fixed-key 64-bit SipHash of the ICC bytes, so a
crafted profile can share another profile's hash (about 2^32 work). Two
places took an equal hash as proof of an identical profile:
- transform::cached keyed its process-wide cache on the two hashes and the
options, so a colliding profile was handed the other profile's cached
transform and converted with the wrong colours.
- Profile::same_colors returned true on equal hashes, so a colliding
profile was taken for the other one and a needed conversion was skipped.
Add Profile::same_bytes, which compares the ICC bytes (an Arc pointer check
first, then the hash, then the bytes). same_colors uses it instead of the
hash. Each transform cache entry now keeps the ICC bytes of its two
profiles; a hit counts only when they match, and a colliding pair gets its
own transform without replacing the cached entry. The transform is built
after the cache lock is released, as before.
The regression tests give Display P3 the content hash of sRGB (a stand-in
for a found collision): before the fix, same_colors called them the same
and the cache returned sRGB's transform for P3; now both see P3's colours.
Fixes#8613
Co-authored-by: John Turnbach <jturnbach@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
- gpu: the test modules share one GPU_LOCK, and timing and vector_mask
take it too, so one process drives one wgpu device at a time.
- ui-egui: the wgpu tests in pattern_preview_canvas, adjust_preview_gpu,
marquee_autoscroll and layout_perf take gpu_lock().
- affinity compound_limits: the child process selects its test by the
module path (compound_limits::<name>); with --exact, the bare name
matches no test and the child passes without running the check.
- xtask scorecard reads crates/io/tests/it/corpus.rs.
- Docs and comments name the new test paths and the --test it filters.
connection() scaled the PCS by src.media_white() / dst.media_white() with
no check on the divisor, so a v4 profile whose wtpt tag holds zeros (parsed
and stored verbatim) gave a NaN scale, and every colour through a direct
absolute transform or a paper-simulating proof came out NaN. The transform
is now refused with CmsError::Invalid when either white has a zero,
negative or non-finite component; the other intents never divide by it and
are unchanged.
Co-authored-by: wstock <1394687+wstock@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Opening any file Photoshop tagged with sRGB IEC61966-2.1 asked
"Embedded Profile Mismatch: sRGB IEC61966-2.1 vs sRGB IEC61966-2.1":
open_policy compared a hash of the ICC bytes, and Photoshop's v2 profile
(1024-entry tables) and the built-in v4 sRGB (parametric curves) are the
same colours in different bytes. 32-bit documents in Photoshop's linear
sRGB missed the stay-linear exception for the same reason.
Add Profile::same_colors: same colour space and every sample of a device
grid within dE76 0.5 through both (re-encoding costs 0.025; moving the
red primary 2% costs 1.0). Colours are compared rather than round-tripped
because CMYK -> PCS -> CMYK is not an identity even for one profile.
Of the 75 profile-tagged PSDs in the test corpora, all 75 prompted on
open; now 6 do (Adobe RGB, Gray Gamma 2.2, SWOP), each a real mismatch.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>