[krispya]

Superseded by merged #299. Main 9022bb40 preserves existing public signatures and supplies current desired placement/caret/selection reads and independent split() copies synchronously without source or whole-root publication. Exact-head review, all five CI jobs and glyph Labs passed (2 faster / 0 slower / 10 neutral). This PR's publish-on-read implementation remains held and must not land alongside #299. #247 remains open; rope draft #292 is held.

---

Mounted Text measurement, caret, selection and split reads publish pending layout through the ordinary engine-wide batch before answering. Publication-before-copy is preserved, and the next unchanged draw reuses the accepted result.

Updated onto current main, including readGlyphs(), split() and the Wasm fallback follow-up. Combined local verification passes: all package builds, source/public/Labs types, 111 focused tests, 277 package tests, 864 complete Node tests, both migration fixture suites, 17 benchmark workflow tests and docs validation.

Performance gate remains open. The previous full Labs run at 961c852 timed out after 90 minutes: artifact preparation took seconds; the 30-case read-publication matrix took 28m 24s on baseline and more than 41m 59s on candidate before cancellation, without a candidate report. A prior comparison reproduced large regressions when editing and reading each label in turn in a 1,000-label root. #247 must resolve that repeated root publication cost. No current-head no-regression claim is made.

The selected benchmark:read-publication suite covers the changed reads, batched and interleaved schedules, and cold load/split copies. This focused exact-head CI comparison is requested to complete the workload report within the existing job budget; cold/full confirmation remains required before release. Hold merge until regression evidence is resolved.

[github-actions]

size-limit report 📦

Surface	gzip	Surface	gzip
Core JS	88.2 kB (0%)	Inter font · Bitmap	558.2 kB (0%)
Shaper Wasm	536.1 kB (0%)	Inter font · MTSDF	8.17 MB (0%)
Three.js adapter JS	136.4 kB (+0.16%)	Inter font · Slug	618.3 kB (0%)
React adapter JS	136.1 kB (+0.18%)	Font Awesome icons · Bitmap	449.9 kB (0%)
Vue adapter JS	135.7 kB (+0.18%)	Font Awesome icons · MTSDF	8.76 MB (0%)
R3F hello-world app JS	742.5 kB (+0.03%)	Font Awesome icons · Slug	657.9 kB (0%)
Tres playground app JS	566.5 kB (+0.04%)	Bitmap baker JS	4.9 kB (0%)
Font validator JS	144.6 kB (0%)	Bitmap baker Wasm	231.1 kB (0%)
Runtime bake host JS	6.1 kB (0%)	MTSDF baker JS	5.6 kB (0%)
Runtime bake Worker JS	13.5 kB (0%)	MTSDF baker Wasm	215.5 kB (0%)
Font baker JS	2.6 kB (0%)	Slug baker JS	4.5 kB (0%)
Font baker Wasm	389.0 kB (0%)	Slug baker Wasm	183.7 kB (0%)

[github-actions]

Docs report 📚

Every changed package is attested at this head. A reviewer verifies each claim after merge.

Package	Concept	Attestation	Status
@pmndrs/glyph-benchmarks	packages/benchmarks.md · edited	Combine both old-canary adapters after their prototype binding; migrate read-publication split calls and indexed type k…	✅
@pmndrs/glyph	packages/glyph.md · edited	Reviewed the combined commit-on-read source after merging origin/main #258. Preserved publication-before-split, readGly…	✅

✅ attested · ⚠️ stale · ❌ missing

[github-actions]

Package performance: read-publication suite

0 faster · 4 slower · 12 neutral · 2 skipped

Baseline 0.1.0-canary-2ab37fda-20261008 → candidate 0.1.0 (ef316833)

                                                faster ◀       ┊       ▶ slower    Δ p50      p  baseline   candidate
- scene 1000 each: 100 rolling tickers                         ┊              ▶  +2553.8%  <.001  █▁▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁▁▅█
- scene 1000 each: edit 50 lines and carets                    ┊              ▶  +1194.6%  <.001  █▁▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁▁█▄
- scene 100 each: 100 rolling tickers                          ┊              ▶  +456.8%  <.001  █▂▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁▇█▂
- scene 100 each: edit 50 lines and carets                     ┊              ▶  +349.6%  <.001  █▁▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁█▂▁
  scene 1: click to place the caret                           ├┼●─┤                +4.3%   .130  ▁▃▆█▆▄▄▃▂▂ ▁▂▅█▅▂▅▆▂▂
  scene 100 batched: 100 rolling tickers                       ├●┤                 +3.7%   .021  ▅█▆▅▄▃▃▄▃▂ ▃█▄▃▆▄▄▂▃▁
  scene 1000 batched: edit 50 lines and carets                 ├●                  +2.3%   .130  ▂▅▅▇█▇▅▃▄▃ ▁▂▄▇█▇▇▄▄▂
  edit a label and draw without reading layout                ├┼●┤                 +2.2%   .382  ▃█▇█▃▃▂▄▅▃ ▄▅█▇▆▃▅▅▄▁
  scene 1: type in a text field                               ├●┤                  +0.6%   .959  ▂▇█▆▇▆▇▅▄▁ ▂▇▅▅██▅▅▅▃
  scene 100 batched: edit 50 lines and carets                ├─●┤                  -0.2%  1.000  ▂▇█▆██▄▄▁▁ ▂▅█▅▆▅▅▄▂▂
  scene 1000: 100 tickers without layout reads                ├●                   -0.5%   .505  ▂▇██▆▆▅▂▂▂ ▃█▇▇▆▄▃▁▁▁
  edit a label and read its glyphs                           ├─●┤                  -0.9%   .959  ▃██▅▃▄▃▃▃▂ ▃█▆▃▃▄▄▃▁▁
  scene 1000: click to place the caret                       ├─●─┤                 -1.0%  1.000  ▁▅▆█▆▄▃▂▁▁ ▂▅▇█▅▄▂▂▁▁
  scene 100: click to place the caret                        ├●┼┤                  -2.1%   .574  ▂█▅▄▂▂▃▃▂▂ ▃██▅▂▂▃▃▅▁
  scene 100: 100 tickers without layout reads                ├●┼─┤                 -2.1%   .645  ▄█▆▃▄▃▄▄▃▁ ▄█▄▄▃▃▂▄▂▁
  scene 100: type in a text field                            ├●┤                   -4.0%   .130  ▃▇▇█▇▃▅▂▃▂ ▃█▇▅▅▃▃▂▂▁
                                                └──────────────┴──────────────┘
                                                -50%           0           +50%

status	bench	baseline	candidate	Δ p50	p	95% CI
🔴 slower	scene 1000 each: 100 rolling tickers	31.23ms	828.85ms	+2553.8%	<.001	+2437.3..+2594.8%
🔴 slower	scene 1000 each: edit 50 lines and carets	24.78ms	320.79ms	+1194.6%	<.001	+1165.8..+1211.7%
🔴 slower	scene 100 each: 100 rolling tickers	18.50ms	102.99ms	+456.8%	<.001	+446.0..+472.3%
🔴 slower	scene 100 each: edit 50 lines and carets	16.15ms	72.59ms	+349.6%	<.001	+336.5..+363.1%

12 neutral, 2 skipped

status	bench	baseline	candidate	Δ p50	p	95% CI
neutral	edit a label and read its glyphs	2.03ms	2.02ms	-0.9%	.959	-5.8..+3.9%
neutral	edit a label and draw without reading layout	1.60ms	1.63ms	+2.2%	.382	-2.4..+7.4%
neutral	scene 1: type in a text field	2.27ms	2.28ms	+0.6%	.959	-2.6..+4.5%
neutral	scene 1: click to place the caret	293.31µs	305.94µs	+4.3%	.130	-1.9..+9.5%
neutral	scene 100: type in a text field	3.09ms	2.97ms	-4.0%	.130	-7.0..+0.9%
neutral	scene 100: click to place the caret	412.13µs	403.59µs	-2.1%	.574	-6.8..+4.5%
neutral	scene 100 batched: edit 50 lines and carets	16.13ms	16.09ms	-0.2%	1.000	-5.9..+3.0%
neutral	scene 100 batched: 100 rolling tickers	18.44ms	19.13ms	+3.7%	.021	+0.3..+6.8%
neutral	scene 100: 100 tickers without layout reads	6.22ms	6.09ms	-2.1%	.645	-6.9..+5.7%
neutral	scene 1000: click to place the caret	1.44ms	1.43ms	-1.0%	1.000	-6.0..+6.7%
neutral	scene 1000 batched: edit 50 lines and carets	25.08ms	25.67ms	+2.3%	.130	-0.4..+4.4%
neutral	scene 1000: 100 tickers without layout reads	15.65ms	15.57ms	-0.5%	.505	-4.6..+0.9%
skipped: clock-confounded: neutral→slower · 2.42→2.62	scene 1000: type in a text field
skipped: clock-confounded: neutral→slower · 2.43→2.59	scene 1000 batched: 100 rolling tickers