fix: preserve decoded image depth and indexed transparency - #276
Open
yeagoo wants to merge 3 commits into
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The Q8 WASM engine can retain a 16-bit input depth tag even though its decoded channels are 8-bit. Writing that tag back produces unnecessarily high-depth output without recovering precision. Conversely, a low-bit palette PNG reports its index depth, which does not describe the precision of its RGB/alpha channels: a semi-transparent alpha value of 128 can become 85 in TIFF output. Indexed PNGs with transparency can also produce an unreadable PSD.
Bound output depth by
Quantum.depth, with an 8-bit minimum for decoded channel values except valid packed TIFF samples, and expand indexed-alpha images to true-color alpha when writing PSD. The upper bound follows the engine's Quantum depth instead of hard-coding Q8. For low-depth TIFF/TIF output, inspect the actual channel precision before promoting the source depth. Valid 1/2/4-bit gray samples remain packed; palette colors and semi-transparent alpha that need 8 bits are still promoted. The precision scan only runs when writing TIFF from an image reporting fewer than 8 bits.Validation
Real WASM round trips cover 16-bit-tagged PNG, low-bit indexed alpha and nontrivial palette colors across PNG, TIFF/TIF, JXL, JP2, PSD, PPM and WebP as applicable. Additional grayscale tests check 1/2/4-bit packed samples through both TIFF extensions. Pixel assertions compare against decoded Q8 inputs, not the original file's full 16-bit precision. AVIF's separate encoder compatibility fix is not required by this branch's tests.
bun run test: 28 passed, including five shared behavior tests.bun run build: passed with a local.envbased on.env.example.bun run check: the same 7 pre-existing missing Node type errors as upstream; no additional errors. Full-repository lint has existing failures outside this change; changed files pass.Patch structure
This branch is based on upstream
cc7b5a54d5e9c797b377db47b9bdfbb561707783and includes shared test harness commitc69e961, which moves the existingmagickConvertfunction unchanged into a utility imported by the worker. Tests invoke that exact production function with the installed WASM. No new package dependency is introduced.Each repair branch can be reviewed independently. After the shared harness lands, the remaining branches will need rebasing onto upstream main to resolve overlapping edits in
magick-convert.ts.Shares the behavior-preserving test harness with #275.