Skip to content

Skip per-visit work that a visit can't need - #189

Merged
joeldrapper merged 1 commit into
mainfrom
claude/project-thread-7y3nfe
Oct 8, 2026
Merged

joeldrapper merged 1 commit into
mainfrom
claude/project-thread-7y3nfe

Conversation

@joeldrapper

Copy link
Copy Markdown
Collaborator

Requested by Joel · project thread

Before: every visited element built its sibling lists and ran every matching pass, even when its only child was a text node on both sides. Every target looked up its enclosing select before checking whether there was a live element to claim, and every inserted node searched for selects and forms even when it had no element children.

After: those visits skip that work. Editing every row of a 1000-item list is 30–41% faster, and no benchmark scenario got slower.

The matching result doesn't change. An element whose only child is text on both sides (both whitespace or both not) morphs the text directly, which is the pair the passes would pick anyway.

How: visitChildNodes has a fast path for a single text child. #canClaim and #claimMovableElement only look up the select once they have a claimable live element. clearImplicitSelection, #uncheckRadiosNamingFormsIn and #uncheckRadiosInFormsIn return early for leaves.

scenario (Chromium, lower quartile of 15 rounds) main this PR
keyed 1000 edit all 15.2ms 10.7ms
unkeyed 1000 edit all 12.6ms 7.4ms
keyed 1000 replace all −26%
links 500 text change −31%
keyed 1000 shuffle −17%

Coverage is 100%. A deep fuzz run passed locally: 5,000 matching, 2,000 preserve, 5,000 tree, 3,000 move and 5,000 form-state seeds.

🤖 Generated with Claude Code

https://claude.ai/code/session_01VJNvudrTrB2rfJaLWKRMsc


Generated by Claude Code

…nd leaf searches

An element whose only child is text on both sides now morphs that text
directly instead of building the sibling lists and running every matching
pass. Claiming a movable element only walks up to its select once there's
a live element to claim, and inserting a node skips the select and form
searches when it has no element children.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VJNvudrTrB2rfJaLWKRMsc
@joeldrapper
joeldrapper marked this pull request as ready for review October 7, 2026 16:35
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 7, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-07T16:36:54.213539Z 78aa85f Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@joeldrapper
joeldrapper merged commit 7dc1b33 into main Oct 8, 2026
12 of 14 checks passed
@joeldrapper
joeldrapper deleted the claude/project-thread-7y3nfe branch October 8, 2026 07:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants