[PR #2965] [CLOSED] perf(process): reuse capture-local text normalization matchers #5522

New issue

Closed

opened 2026-10-31 01:07:18 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 01:07:18 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/2965
Author: @rudycelekli
Created: 10/13/2026
Status: ❌ Closed

Base: main ← Head: candidate/rea-next10-normalization-20261013

📝 Commits (2)

ff68121 perf(process): reuse capture-local text normalization matchers
ce0fa80 Merge remote-tracking branch 'origin/main' into HEAD

📊 Changes

1 file changed (+39 additions, -0 deletions)

View changed files

📝 src/process/capture/ProcessNormalization.test.ts (+39 -0)

📄 Description

Repeated terminal updates rebuild the same path, PID and caller-pattern matchers. Prepare immutable scenario-dependent matchers once per capture and share them between renderer and retained-event consumers, while keeping the PID lookup live during terminal startup. Sample PID matchers are prepared only when PID normalization is selected. Output/provenance equivalence passes 9,035 adversarial producer assertions; no target authority or capture contract changes. AI-assisted implementation and verification with OpenAI Codex.

Verified on current main 8fef2aa6: npm run check and 13 focused normalization tests pass. The executed macOS PTY suite passed 6 cases with 2 Linux-only skips before the clean rebase; all touched producer/consumer bytes are unchanged. No Windows/Linux runtime claim.

Two 24-trial actual TerminalRenderer measurements preserve identical output digests. First warm elapsed medians for 0/4/64 observed paths were 136→134,151→260 and 315→201ms; the four-path result regressed. A subsequent bounded run measured 160→171,189→160 and 218→145ms, with a zero-path regression. Both host measurements and CPU samples are retained; peak process RSS was 162/163 MB for the combined runs. These are producer measurements on a contended host, not universal speedups or whole-task savings. Task tokens and process-capture cost remain unmeasured.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/morluto/rea/pull/2965 **Author:** [@rudycelekli](https://github.com/rudycelekli) **Created:** 10/13/2026 **Status:** ❌ Closed **Base:** `main` ← **Head:** `candidate/rea-next10-normalization-20261013` --- ### 📝 Commits (2) - [`ff68121`](https://github.com/morluto/rea/commit/ff68121a2fa9c03cef9610280d28e9c89208d796) perf(process): reuse capture-local text normalization matchers - [`ce0fa80`](https://github.com/morluto/rea/commit/ce0fa800658d7e38cd364d0bc97c220821177d77) Merge remote-tracking branch 'origin/main' into HEAD ### 📊 Changes **1 file changed** (+39 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `src/process/capture/ProcessNormalization.test.ts` (+39 -0) </details> ### 📄 Description Repeated terminal updates rebuild the same path, PID and caller-pattern matchers. Prepare immutable scenario-dependent matchers once per capture and share them between renderer and retained-event consumers, while keeping the PID lookup live during terminal startup. Sample PID matchers are prepared only when PID normalization is selected. Output/provenance equivalence passes 9,035 adversarial producer assertions; no target authority or capture contract changes. AI-assisted implementation and verification with OpenAI Codex. Verified on current main `8fef2aa6`: `npm run check` and 13 focused normalization tests pass. The executed macOS PTY suite passed 6 cases with 2 Linux-only skips before the clean rebase; all touched producer/consumer bytes are unchanged. No Windows/Linux runtime claim. Two 24-trial actual TerminalRenderer measurements preserve identical output digests. First warm elapsed medians for 0/4/64 observed paths were 136→134,151→260 and 315→201ms; the four-path result regressed. A subsequent bounded run measured 160→171,189→160 and 218→145ms, with a zero-path regression. Both host measurements and CPU samples are retained; peak process RSS was 162/163 MB for the combined runs. These are producer measurements on a contended host, not universal speedups or whole-task savings. Task tokens and process-capture cost remain unmeasured. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-31 01:07:18 +00:00

closed this issue
added the
pull-request
label

bsakaguchi referenced this issue

2026-11-01 21:44:27 +00:00

[PR #5581] [MERGED] fix(environment): report later bundle write failures as host failures #6978

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

morluto/rea#5522

No description provided.

Rows
Columns