Repository navigation
Integrate three open upstream mods PRs, rebased onto current main - #3
Conversation
…entries' mtimeMs; the test fakes answer them
… to the next edit, under test Applied from anthropics/claude-code PR 94847 (squashed; all 7 commits). Two conflicts with newer main (a doc comment in hooks/register.ts and a test title in tests/register.test.ts) were resolved by taking the PR's side.
…nd and reads the entry as the telemetry types give it Applied from anthropics/claude-code PR 97688 (squashed; both commits). Merged with newer main: the README event list, the fixtures index and the test file were combined, and the new collector test answers settings.read like its neighbours, since main now hooks plugin.register.
![chatgpt-codex-connector[bot]](/web-api/iframe-proxy?url=https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a701c730d6
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".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| tree: { cwd: string; toplevel: string }, | ||
| ): boolean { | ||
| const top = normalPathOf(tree.toplevel) | ||
| const absolute = Git.isAbsolutePath(path) ? path : `${tree.cwd}/${path}` |
There was a problem hiding this comment.
On Windows, a valid drive-rooted path such as \temp\note.md (and a UNC path) does not satisfy Git.isAbsolutePath, so it is concatenated with cwd and normalized as if it were under the repository. If the working tree already has a session diff—for example, from a preceding Bash command—an external Write using one of these paths can incorrectly trigger the pane's auto-open and repository fetch. Treat Windows rooted/UNC paths as absolute before applying the working-tree check.
Useful? React with 👍 / 👎.
3 participants
Summary
Applies three open pull requests from
anthropics/claude-codethat touchmods/, merged onto currentmain. All three were opened against an oldermainand conflicted or failed to typecheck, so each needed a small merge by hand. Original authorship is kept on every commit.process.runtruncation flags andmtimeMson list entries in the declarationsmain;tests/register.test.tsbuilt anFsEntrywithout the now-requiredmtimeMs, so it gotmtimeMs: 0hooks/register.tsand one test title conflicted; both resolved by taking the PR's sidesec-default: collector records continue past the user tiersettings.read, becausemainhooksplugin.registersince the PR was openedNot applied: upstream PR 97334 (
sec-default: rows a conversation keeps continue past the user tier). It overlaps with newermainwork in the same mod: both sides created differenttests/fixtures/rewording.tsfiles, and the README and test file conflict in several places. That needs its author to rebase, becausesec-defaultis a security policy mod and I did not want to guess at the merge.Review notes
sec-defaultis a security policy mod. The only behaviour change is 4 lines inhooks/register.tsthat passtelemetry.logcollector records on to theappendtier, the same pattern assettings.read. Worth a human read.main's doc comment inmods/diff/hooks/register.tswith the PR's wording. The tests pass, but whether the comment still matchesmain's intent is a judgement call.Test plan
Run as the repo's
mod-tests.ymldoes (tsc5.9.3 andclaude plugin test, CLI 2.1.287):tsc -p mods/tsconfig.json: no errorsmods/agents-md: 5 passmods/diff: 234 pass (210 onmain, +24)mods/sec-default: 58 pass (57 onmain, +1)mods/telemetry: 34 passMod testsworkflow runs against the CLI from npm and may behave differently🤖 Generated with Claude Code
https://claude.ai/code/session_01LoLAv3aLaRJocsRMcjFbpv
Generated by Claude Code