Expect OverflowError for dimensions that exceed sys.maxsize, since
PyArg_ParseTuple rejects them before the PNG buffer validation runs.
Continue to expect ValueError for invalid dimensions and buffer lengths
that reach the validation code.
- Parse in-band resize reports robustly: require the 48; prefix, accept
extra parameters and reject negative values instead of panicking
- Handle bracketed paste in parse_input_from_terminal with a dedicated
scan for the end of paste marker so pasted escape bytes are never lost
- Guard against NaN in scroll_by_pixels and scroll_to_absolute, clamp
negative absolute scroll targets, let infinite fractional scrolls clamp
and propagate overflow errors from PyLong_AsDouble
- Empty disk cache entries no longer claim offset 0 in the cache file and
defrag only moves entries actually written to disk
- Reject zero and too large block sizes in rsync signature headers
- Only replace the encryptor tag once it has been read successfully
- Back off on all accept() errors that leave the connection pending, not
just resource exhaustion, so the talk thread cannot spin on a readable
listener
- Use an atomic for the test only accept failure injection instead of
taking the talk mutex on every accept
- Encrypt disk cache entries in chunks of up to 1MB to reduce the number
of write syscalls for large images
- Have GlobalWatchers.set_extra() invalidate the cached watchers
resize_screen_buffers() moves the storage of the source history buffer into
the result, both when reusing unchanged lines and in the same size fast path.
Replace the global used to opt out of this with an explicit HistoryResizeMode
argument so callers that must keep their history (HistoryBuf.rewrap) state it,
and make the same size fast path copy rather than steal in that mode. Document
that nothing may fail after the history has been consumed.
These two code points have Grapheme_Cluster_Break=SpacingMark but a
width of one, the only such code points in Unicode. Grapheme
segmentation (GB9a) adds them to the previous cell, so "จำ" occupied a
single cell in kitty while wcwidth() based programs count two. As a
result the SARA AA glyph was clipped, only the nikhahit was visible and
the cursor drifted right by one cell per AM.
Now, when such a code point combines into a plain one cell wide cell,
the cell is widened to two cells, using the same mechanism as U+FE0F
does for narrow emoji. The C and Go wcswidth() implementations are
updated to match so that kitty's own width calculations agree with the
screen.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
NUL is anyway not legal in paths and it causes session files saved with
such paths to fail
We dont remove other control codes since they are actually legal in
paths.
Fixes#10321
Implements modern wide gamut color formats with CSS Color Module Level 4
gamut mapping, addressing PR feedback with Go implementation, performance
benchmarks, and reorganized documentation.
Features:
- OKLCH (perceptually uniform color space)
- CIE LAB (device-independent color space)
- CSS Color 4 compliant gamut mapping algorithm
- Inline comment support in color config parsing
Addressing PR Feedback:
1. Go Implementation (tools/utils/style/):
- Complete OKLCH and LAB parsing with gamut mapping
- Matches Python implementation structure
- Comprehensive test suite (all tests passing)
- Performance benchmarks showing acceptable overhead
2. Performance Benchmarks:
- OKLCH: ~4.6 µs/op
- LAB: ~1.5 µs/op
- 10 mixed colors: ~13 µs total
- Typical config (50 colors): <0.5ms startup impact
3. Documentation Reorganization:
- Moved detailed color docs to docs/wide-gamut-colors.rst
- Configuration docs now link to separate documentation
- Reduces size of main configuration documentation
Gamut Mapping:
- Binary search chroma reduction from CSS Color Module Level 4
- Preserves lightness and hue while reducing chroma for out-of-gamut colors
- Uses deltaE OK (JND threshold: 0.02) for perceptual difference
- Ensures graceful degradation on sRGB displays
Python Implementation:
- parse_oklch(): OKLCH color parsing with gamut mapping
- parse_lab(): CIE LAB parsing with gamut mapping via OKLCH conversion
- lab_to_oklch(): LAB to OKLCH conversion for consistent gamut mapping
- oklch_to_srgb_gamut_map(): CSS Color 4 gamut mapping algorithm
- srgb_to_oklab(): Reverse conversion for deltaE calculations
- deltaE_ok(): Perceptual color difference in OKLab space
Go Implementation:
- colorspaces.go: All color space conversions and gamut mapping
- wrapper.go: ParseColor() updated to support OKLCH and LAB
- Comprehensive test coverage with benchmarks
- Matches Python implementation behavior
Robustness:
- NaN and infinity validation in all color parsing functions
- Defense-in-depth with validation at parsing and gamut mapping levels
- Returns None/error for invalid input (consistent error handling)
- Validates before clamping operations to prevent NaN propagation
Files changed:
- Python: kitty/rgb.py, kitty_tests/datatypes.py (+250 lines)
- Go: tools/utils/style/colorspaces.go, wrapper.go (+350 lines, tests)
- Docs: docs/wide-gamut-colors.rst (moved from inline)
- Config: kitty/options/definition.py (simplified, links to docs)
References:
- CSS Color Module Level 4: https://www.w3.org/TR/css-color-4/
- OKLCH Color Space: https://bottosson.github.io/posts/oklab/🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
This was needed to fix various corner cases when doing blending of colors
in linear space. The new architecture has the same performance as the
old in the common case of opaque rendering with no UI layers or images.
In the case of only positive z-index images there is a performance
decrease as the OS Window is now rendered to a offscreen texture and
then blitted to screen. However, in the future when we move to Vulkan or
I can figure out how to get Wayland to accept buffers with colors in
linear space, this performance penalty can be removed. The performance
penalty was not significant on my system but this is highly GPU
dependent. Modern GPUs are supposedly optimised for rendering to
offscreen buffers, so we will see. The awrit project might be a good
test case.
Now either we have 1-shot rendering for the case of opaque with only ext
or all the various pieces are rendered in successive draw calls into an
offscreen buffer that is blitted to the output buffer after all drawing
is done.
Fixes#8869
Now we do the rewrap of history and line buffers together. This is
faster and deals with multiline chars split between the two buffers
correctly. Also, considerably simpler code.