perf(textkit): improve line slicing performance

[wojtekmaj]

This improves textkit line slicing performance for long paragraphs without changing line-breaking behavior.

Changes:

• use binary search in glyphIndexAt instead of scanning backward through the full glyph index array
• count ligature offsets by scanning the local repeated-index cluster instead of slicing/filtering the whole prefix
• build breakLines with push instead of copying the accumulated line array on every breakpoint

Measured in a downstream PDF generator with a single long left-aligned body paragraph:

Baseline, 200k chars: ~31.3s wall, ~43.5s CPU, ~3.38GB maxRSS
With these textkit changes: ~17.0s wall, ~26.4s CPU, ~3.33GB maxRSS

Together with #3421, this improves performance in the same downstream PDF generator by:

Baseline, 50k chars: ~2.4s wall, ~3.0s CPU, ~563MB maxRSS
With #3420 + #3421: ~713ms wall, ~1.11s CPU, ~291MB maxRSS

Baseline, 100k chars: ~8.1s wall, ~10.1s CPU, ~1.18GB maxRSS
With #3420 + #3421: ~1.21s wall, ~1.73s CPU, ~423MB maxRSS

Baseline, 200k chars: ~31.3s wall, ~43.5s CPU, ~3.38GB maxRSS
With #3420 + #3421: ~2.28s wall, ~2.96s CPU, ~513MB maxRSS

The memory profile is mostly unchanged with this PR alone because the dominant heap allocation comes from the Knuth-Plass active-node search, but the CPU/wall-time improvement is still measurable.

[changeset-bot]

🦋 Changeset detected

Latest commit: a2b98ae

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 9 packages

Name	Type
@react-pdf/textkit	Patch
@react-pdf/layout	Patch
@react-pdf/render	Patch
@react-pdf/renderer	Patch
@react-pdf/math	Patch
@react-pdf/mermaid	Patch
next-14	Patch
next-15	Patch
@react-pdf/vite-example	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[Copilot]

Pull request overview

This PR improves @react-pdf/textkit line slicing performance for long paragraphs while preserving existing line-breaking behavior.

Changes:

• Optimize ligature offset calculation by scanning only the local repeated-index cluster in offset.
• Replace the linear backward scan in glyphIndexAt with a binary search over glyphIndices.
• Avoid repeated array copying in breakLines by accumulating lines via push.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
packages/textkit/src/run/offset.ts	Reduces ligature offset computation from prefix slicing/filtering to a bounded backward scan.
packages/textkit/src/run/glyphIndexAt.ts	Uses binary search to find the last glyph mapping for a given string index.
packages/textkit/src/engines/linebreaker/index.ts	Builds the lines array via mutation (push) instead of repeated spreading during accumulation.
.changeset/smart-maps-pay.md	Adds a patch changeset describing the performance improvement.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.