feat(Alert): cascade color to inner Avatar - #29
Open
IgorShevchik wants to merge 1 commit into
Open
Conversation
This was referenced May 11, 2026
IgorShevchik
force-pushed
the
feat/alert-avatar-color
branch
from
May 13, 2026 06:59
6c9f2f5 to
c13b748
Compare
The existing top-level `color` prop now also tints the inner `B24Avatar` by default. An explicit `avatar.color` still wins, since `v-bind="props.avatar"` is applied after `:color="props.color"`. `Alert.variants.color` is a superset of `AvatarProps['color']` (legacy `'default'` etc.), so the binding casts to the Avatar color type at the template site. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
IgorShevchik
force-pushed
the
feat/alert-avatar-color
branch
from
May 13, 2026 07:06
c13b748 to
d2502cf
Compare
IgorShevchik
pushed a commit
that referenced
this pull request
Aug 13, 2026
…code points Code points are not what a reader sees. A flag is two regional indicators, a family emoji several joined by ZWJ, `कि` a consonant plus a vowel sign. Cutting inside one produces a *different* character rather than a broken one — 🇺🇸 truncated by a single code point re-pairs into 🇸🇺, a different country, with nothing to signal the loss. Both boundaries snap off the straddled cluster: the mark insertion in `generateHighlightedText` and the truncation in `truncateHTMLFromStart`. Chosen on measurement. Segmenting the whole string costs 455 µs at 979 characters and 52 ms at 100k. A fixed ±64 window is constant-time but wrong: a run of flags is longer than the window, so segmentation starts mid-run and re-pairs the indicators — reproducing the exact bug, 28 disagreements in 1044 probes. `Segments.containing()` is correct, so it ships, behind three guards: a `< U+0300` screen, which keeps ASCII and Latin-1 off the segmenter but does nothing for Cyrillic or CJK, both above the floor; one segmenter view per value, built lazily; and an 8192-character ceiling, past which only the surrogate snap applies and every multi-code-point cluster loses protection. Two defects found in review are folded in. A region left empty by the ordering clamps emitted a bare `<mark></mark>` with the highlight lost. A region past the end of the value bypassed that guard entirely, because `substring()` clamps its own arguments while the comparison did not — it emitted the same empty mark, froze the cursor past every later region, and the trailing mark then drove truncation into eating the text before it. `indices` are sorted before use. That is a no-op for the search path, since Fuse merges and sorts them itself, but `highlight()` is a published export and `CommandPaletteGroup.postFilter` lets a caller supply its own matches; out-of-order boundaries make `containing()` re-walk a long run of regional indicators from its start, costing as much as the per-boundary segmenter this avoids. `truncateHTMLFromStart` counts retained code units and slices once, because the snap needs an index to move. Verified equivalent over 400,078 fuzzed inputs, including strings that themselves begin with `...`. One fixture trap is recorded in the tests: UAX #29 glues a run of bare emoji modifiers into a single cluster, so `'\u{1F3FF}'.repeat(20)` is one character rather than twenty. Closes #364 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LWWrBHgfqGSbeU3V6UuMF8
IgorShevchik
pushed a commit
that referenced
this pull request
Aug 13, 2026
…code points Code points are not what a reader sees. A flag is two regional indicators, a family emoji several joined by ZWJ, `कि` a consonant plus a vowel sign. Cutting inside one produces a *different* character rather than a broken one — 🇺🇸 truncated by a single code point re-pairs into 🇸🇺, a different country, with nothing to signal the loss. Both boundaries snap off the straddled cluster: the mark insertion in `generateHighlightedText` and the truncation in `truncateHTMLFromStart`. Chosen on measurement. Segmenting the whole string costs 455 µs at 979 characters and 52 ms at 100k. A fixed ±64 window is constant-time but wrong: a run of flags is longer than the window, so segmentation starts mid-run and re-pairs the indicators — reproducing the exact bug, 28 disagreements in 1044 probes. `Segments.containing()` is correct, so it ships, behind three guards: a `< U+0300` screen, which keeps ASCII and Latin-1 off the segmenter but does nothing for Cyrillic or CJK, both above the floor; one segmenter view per value, built lazily; and an 8192-character ceiling, past which only the surrogate snap applies and every multi-code-point cluster loses protection. Two defects found in review are folded in. A region left empty by the ordering clamps emitted a bare `<mark></mark>` with the highlight lost. A region past the end of the value bypassed that guard entirely, because `substring()` clamps its own arguments while the comparison did not — it emitted the same empty mark, froze the cursor past every later region, and the trailing mark then drove truncation into eating the text before it. `indices` are sorted before use. That is a no-op for the search path, since Fuse merges and sorts them itself, but `highlight()` is a published export and `CommandPaletteGroup.postFilter` lets a caller supply its own matches; out-of-order boundaries make `containing()` re-walk a long run of regional indicators from its start, costing as much as the per-boundary segmenter this avoids. `truncateHTMLFromStart` counts retained code units and slices once, because the snap needs an index to move. Verified equivalent over 400,078 fuzzed inputs, including strings that themselves begin with `...`. One fixture trap is recorded in the tests: UAX #29 glues a run of bare emoji modifiers into a single cluster, so `'\u{1F3FF}'.repeat(20)` is one character rather than twenty. Closes #364 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LWWrBHgfqGSbeU3V6UuMF8
IgorShevchik
pushed a commit
that referenced
this pull request
Aug 13, 2026
…code points Code points are not what a reader sees. A flag is two regional indicators, a family emoji several joined by ZWJ, `कि` a consonant plus a vowel sign. Cutting inside one produces a *different* character rather than a broken one — 🇺🇸 truncated by a single code point re-pairs into 🇸🇺, a different country, with nothing to signal the loss. Both boundaries snap off the straddled cluster: the mark insertion in `generateHighlightedText` and the truncation in `truncateHTMLFromStart`. Chosen on measurement. Segmenting the whole string costs 455 µs at 979 characters and 52 ms at 100k. A fixed ±64 window is constant-time but wrong: a run of flags is longer than the window, so segmentation starts mid-run and re-pairs the indicators — reproducing the exact bug, 28 disagreements in 1044 probes. `Segments.containing()` is correct, so it ships, behind three guards: a `< U+0300` screen, which keeps ASCII and Latin-1 off the segmenter but does nothing for Cyrillic or CJK, both above the floor; one segmenter view per value, built lazily; and an 8192-character ceiling, past which only the surrogate snap applies and every multi-code-point cluster loses protection. Two defects found in review are folded in. A region left empty by the ordering clamps emitted a bare `<mark></mark>` with the highlight lost. A region past the end of the value bypassed that guard entirely, because `substring()` clamps its own arguments while the comparison did not — it emitted the same empty mark, froze the cursor past every later region, and the trailing mark then drove truncation into eating the text before it. `indices` are sorted before use. That is a no-op for the search path, since Fuse merges and sorts them itself, but `highlight()` is a published export and `CommandPaletteGroup.postFilter` lets a caller supply its own matches; out-of-order boundaries make `containing()` re-walk a long run of regional indicators from its start, costing as much as the per-boundary segmenter this avoids. `truncateHTMLFromStart` counts retained code units and slices once, because the snap needs an index to move. Verified equivalent over 400,078 fuzzed inputs, including strings that themselves begin with `...`. One fixture trap is recorded in the tests: UAX #29 glues a run of bare emoji modifiers into a single cluster, so `'\u{1F3FF}'.repeat(20)` is one character rather than twenty. Closes #364 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LWWrBHgfqGSbeU3V6UuMF8
IgorShevchik
pushed a commit
that referenced
this pull request
Aug 13, 2026
…code points Code points are not what a reader sees. A flag is two regional indicators, a family emoji several joined by ZWJ, `कि` a consonant plus a vowel sign. Cutting inside one produces a *different* character rather than a broken one — 🇺🇸 truncated by a single code point re-pairs into 🇸🇺, a different country, with nothing to signal the loss. Both boundaries snap off the straddled cluster: the mark insertion in `generateHighlightedText` and the truncation in `truncateHTMLFromStart`. Chosen on measurement. Segmenting the whole string costs 455 µs at 979 characters and 52 ms at 100k. A fixed ±64 window is constant-time but wrong: a run of flags is longer than the window, so segmentation starts mid-run and re-pairs the indicators — reproducing the exact bug, 28 disagreements in 1044 probes. `Segments.containing()` is correct, so it ships, behind three guards: a `< U+0300` screen, which keeps ASCII and Latin-1 off the segmenter but does nothing for Cyrillic or CJK, both above the floor; one segmenter view per value, built lazily; and an 8192-character ceiling, past which only the surrogate snap applies and every multi-code-point cluster loses protection. Two defects found in review are folded in. A region left empty by the ordering clamps emitted a bare `<mark></mark>` with the highlight lost. A region past the end of the value bypassed that guard entirely, because `substring()` clamps its own arguments while the comparison did not — it emitted the same empty mark, froze the cursor past every later region, and the trailing mark then drove truncation into eating the text before it. `indices` are sorted before use. That is a no-op for the search path, since Fuse merges and sorts them itself, but `highlight()` is a published export and `CommandPaletteGroup.postFilter` lets a caller supply its own matches; out-of-order boundaries make `containing()` re-walk a long run of regional indicators from its start, costing as much as the per-boundary segmenter this avoids. `truncateHTMLFromStart` counts retained code units and slices once, because the snap needs an index to move. Verified equivalent over 400,078 fuzzed inputs, including strings that themselves begin with `...`. One fixture trap is recorded in the tests: UAX #29 glues a run of bare emoji modifiers into a single cluster, so `'\u{1F3FF}'.repeat(20)` is one character rather than twenty. Closes #364 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LWWrBHgfqGSbeU3V6UuMF8
IgorShevchik
pushed a commit
that referenced
this pull request
Aug 13, 2026
…code points Code points are not what a reader sees. A flag is two regional indicators, a family emoji several joined by ZWJ, `कि` a consonant plus a vowel sign. Cutting inside one produces a *different* character rather than a broken one — 🇺🇸 truncated by a single code point re-pairs into 🇸🇺, a different country, with nothing to signal the loss. Both boundaries snap off the straddled cluster: the mark insertion in `generateHighlightedText` and the truncation in `truncateHTMLFromStart`. Chosen on measurement. Segmenting the whole string costs 455 µs at 979 characters and 52 ms at 100k. A fixed ±64 window is constant-time but wrong: a run of flags is longer than the window, so segmentation starts mid-run and re-pairs the indicators — reproducing the exact bug, 28 disagreements in 1044 probes. `Segments.containing()` is correct, so it ships, behind three guards: a `< U+0300` screen, which keeps ASCII and Latin-1 off the segmenter but does nothing for Cyrillic or CJK, both above the floor; one segmenter view per value, built lazily; and an 8192-character ceiling, past which only the surrogate snap applies and every multi-code-point cluster loses protection. Two defects found in review are folded in. A region left empty by the ordering clamps emitted a bare `<mark></mark>` with the highlight lost. A region past the end of the value bypassed that guard entirely, because `substring()` clamps its own arguments while the comparison did not — it emitted the same empty mark, froze the cursor past every later region, and the trailing mark then drove truncation into eating the text before it. `indices` are sorted before use. That is a no-op for the search path, since Fuse merges and sorts them itself, but `highlight()` is a published export and `CommandPaletteGroup.postFilter` lets a caller supply its own matches; out-of-order boundaries make `containing()` re-walk a long run of regional indicators from its start, costing as much as the per-boundary segmenter this avoids. `truncateHTMLFromStart` counts retained code units and slices once, because the snap needs an index to move. Verified equivalent over 400,078 fuzzed inputs, including strings that themselves begin with `...`. One fixture trap is recorded in the tests: UAX #29 glues a run of bare emoji modifiers into a single cluster, so `'\u{1F3FF}'.repeat(20)` is one character rather than twenty. Closes #364 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LWWrBHgfqGSbeU3V6UuMF8
8 tasks
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
1 participant
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.
Summary
colorprop onB24Alertnow also tints the innerB24Avatarby default.avatar.colorstill wins, sincev-bind="props.avatar"is applied after:color="props.color".Alert.variants.coloris a superset ofAvatarProps['color'](it includes legacy entries like'default'), so the binding casts to the Avatar color type at the template site — no API change, no theme change.Follow-up to #24 (Avatar/AvatarGroup
color), #27 (User color), #28 (ChatMessage color).Test plan
pnpm run lintpassespnpm run typecheckpasses (root + docs + both playgrounds)pnpm run testpasses — newrenders with color and avatar correctlysnapshot witnesses the cascade (style-filled-successon the Avatar root when Alert isair-primary-success)avatar.coloroverrides the cascade (set:avatar="{ src: '...', color: 'air-tertiary-accent' }"whilecolor="air-primary-alert"onB24Alert— avatar should be tertiary-accent)🤖 Generated with Claude Code