Text sizing: span width (w), packed sub-cell runs, TextSizing converter, wide-glyph guarantee - #74
Merged
Merged
Conversation
…ured, packed; TextSizing string form + converter The 'w' key is the width in cells of the WHOLE text of one sequence — 'all the text in that escape code must be rendered in s·w cells' (spec re-read; maintainer-verified against kitty, 2026-09-12) — not the per-cluster advance the 2026-08-02 'unsupported by decision' stance assumed. It is now emitted and measured: TextSizing.Width (0–7) pins one sequence's width (TextSizing.FixedWidth — the wide-glyph guarantee); TextSizing.Packed has the writer compute w per sequence as the packed natural width, ⌈natural·n/d⌉ cells, chunking the text on cluster boundaries so no sequence exceeds the spec's seven cells (TextSizingPacking) — several half-size glyphs per cell for read-only sub/superscripts (TextSizing.Superscript/Subscript). SpanSize/SpanColumns give the whole-cell footprint layout, the caret and the buffer see; SizedTextFragment measures with it. IsSupported gates w on the Width capability (kitty negotiates Width with Scale). The fraction rule is the spec's strict d > n. TextSizing gains a string form and a TypeConverter (XAML and options): named tokens (Normal, Double, 2x…7x, Superscript, Subscript [n/d], Packed, Top/Bottom/Center, Left/Right/HCenter, a bare fraction) and the OSC metadata keys (s= w= n= d= v= h= p=), mixed freely; ToString is the canonical metadata form and round-trips. Gallery: packed samples on terminals that negotiate w. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013qbF5ru6uh5w7fqhP3yoPE
…red pieces in the formatter ScaledGlyphMetrics is fraction-aware: a packed cluster's own advance rounds up to whole cells (so a formatter's wrap decisions stay conservative — a packed run breaks a touch early, never late), while StringWidth is the exact packed span footprint. A metrics that MeasuresSpans has TextFormatter re-measure each piece it emits, so the width recorded for a piece is the footprint its SizedTextFragment claims: 'H' + a packed superscript '12' + 'O' lays out three cells wide. True sub-cell layout (a caret between two glyphs in one cell) stays out of scope — TextPresenter is untouched; this serves read-only text. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013qbF5ru6uh5w7fqhP3yoPE
… emitted as OSC 66 spans pinned to their cell count On a terminal that honors the 'w' key, FrameRenderer emits a WideLeft glyph as a w=2 sequence and an East-Asian- Ambiguous single as w=1: the terminal renders exactly the footprint the buffer allotted whatever its own width tables say, the cursor advance is known, and the two defenses (the two-space pre-paint + CUP-back for untrusted wide glyphs; the neighbor-first + skip for ambiguous glyphs, with its distinct-neighbors limitation) are bypassed there. Plain ASCII and empty cells stay plain. Gated on TextSizingCapabilities.Width (kitty negotiates it with Scale). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013qbF5ru6uh5w7fqhP3yoPE
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.
Kitty text sizing: the
wkey is a span width (maintainer-verified 2026-09-12), so it is now emitted, measured and packed; plus sub-cell measuring for read-only text and the wide-glyph guarantee in the renderer. Three commits, each shippable on its own.1. Span width, packing, string form.
TextSizing.Width(0–7) is the width in cells of the whole text of one sequence and is emitted asw.TextSizing.Packedhas the writer computewper sequence as the packed natural width, ⌈natural·n/d⌉ cells, chunking on cluster boundaries so no sequence exceeds the spec's seven cells (TextSizingPacking) — that's how several half-size glyphs share a cell (TextSizing.Superscript/Subscript).SpanSize/SpanColumnsgive the whole-cell footprint layout, the caret and the buffer see;SizedTextFragmentmeasures with it;IsSupportedgateswon theWidthcapability. The fraction rule is the spec's strictd > n.TextSizingalso gainsParse/TryParse/ToStringand a[TypeConverter]for XAML and options: named tokens (Normal,Double,2x…7x,Superscript,Subscript [n/d],Packed,Top/Bottom/Center,Left/Right/HCenter, a bare fraction) and the OSC keys (s= w= n= d= v= h= p=), mixed freely;ToStringis the canonical metadata form and round-trips.2. Sub-cell measuring.
ScaledGlyphMetricsis fraction-aware: a packed cluster's own advance rounds up (wrap decisions stay conservative),StringWidthis the exact packed span, and aMeasuresSpansmetrics hasTextFormatterre-measure each emitted piece —H+ a packed superscript12+Olays out three cells wide, matching the fragment it paints.TextPresenteris untouched; this serves read-only text.3. Width guarantee. On a
w-capable terminalFrameRendereremits a wide glyph as aw=2sequence and an East-Asian-Ambiguous glyph asw=1, so the terminal renders exactly the buffer's footprint whatever its width tables say, the cursor advance is known, and both defenses (the pre-paint + CUP-back, the neighbor-first + skip with its distinct-neighbors limitation) are bypassed there.Gallery: packed superscript/subscript/double samples on terminals that negotiate
w— worth an eyeball in kitty, especially thew=1ambiguous path under an ambiguous-as-wide setting.Tests: Core 1178, Rendering 1876, UI 3962, Bars 314, Xaml 647, Interactivity 47, Drawing 707 — all green.
🤖 Generated with Claude Code
https://claude.ai/code/session_013qbF5ru6uh5w7fqhP3yoPE