Repository navigation
Integration batch H (#1413, #1376, #1406) - #1428
Merged
Merged
Conversation
…esidual) Refs #1273 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Claude proxy addon now rotates proxy-events.jsonl into proxy-events.1.jsonl at 512 MiB. Right after a rotation the live file holds a few lines, so a bundle would miss the recent traffic a report is about. When the live file is under the 5 MiB cap, the rest is filled from the end of the previous generation, cut at a line boundary. Refs #1273 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…wnership Refs #1273 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…og (#1273) claude-code-headless#64 (merged, f52fc82): the addon rotates proxy-events.jsonl at 512 MiB keeping one previous generation, each generation stamped with a header; the app's tail holds a FileHandle on the generation it reads and reports generations deleted unread as a `transport-gap`. A live Claude run is now bounded instead of growing for the life of the session. No lockfile resync: the app consumes the package from source and its runtime deps are root deps; the bump changes only the package's devDependencies. The plan records the header design, the accepted crash limitation, and that #1332 (W1) owns the rotation-safe bundle reader. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… retention (review) Review of #1376 (a, b, c): - ClaudeSession listened to the proxy's `event` channel only, so the `transport-gap` claude-code-headless#64 emits for generations deleted unread never left the package. attachProxyServer subscribes both and re-emits `proxy-transport-gap`; SessionManager records a claude.proxy_transport_gap incident and re-emits it with the session id. The normal event wiring is now pinned (deleting it passed every test before). - collectProxyRunDirs counts a run holding only proxy-events.1.jsonl. - sslkeylog.log is a separate unbounded file (and TLS secrets on disk): filed as #1380; the PR narrows to Refs #1273. Mutations killed: session ignoring transport-gap, session ignoring event, manager ignoring the gap, retention ignoring `.1`. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… marker is #1381 (q87) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…inst its successor (focused review) Focused review of #1376 (c): the ownership guard on the gap subscription was unpinned (deleting it passed). Also corrects the plan: generation 0 has no header, and PROXY_EVENTS_ROTATE_BYTES=0 disables the bound. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…1187) The flaky ['/'] hit was the host Agent Code app's LanePortWatcher probing the harness's listener with GET / (Node fetch) because the suite ran inside an agent lane; under load it landed before the egress assertion. The ZodError is the harness's own forgery probe, and 'No handler registered' is normal disposal noise. Only requests for the probed path count as egress now. Reproduced 2/10 under load before, 0/10 after; a guard-removal mutant still fails. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…other request, add a positive control (#1406 review b) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…lank lines Reviewer C (round 2): Decisions/Tests/Delivery still stated the pre-review scope (a fully bounded run, no app tests, pointer-only). They now match the review appendix: the events file is bounded, the run directory is not (#1380), and the app-side wiring and its tests are listed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…1250 row 12) loadOlderHistory now answers loaded/skipped/failed; TileLeaf shows a fixed, coalesced pane toast on failed. hasOlderHistory stays set so the next scroll to the top retries. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…view a, b) Also records in the plan that e905104 carried the loader fix (an unreadable transcript now rejects an older page instead of paging it as empty). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…; sync the plan (#1406 review c) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… target pane (#1413 review c) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#1413 verification b) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…1413 verification a) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner
Author
|
Batch disposition (B6, pre-merge): members at GATE PASS (member mode): #1413 fd5c2ef (manager-verified a/b at the cap), #1376 69adefb (manager-verified c), #1406 55527cd. Base is 6938e61. Residuals are on each member's disposition. Not included: #1423 (CI pending), #1392 (reviewer c round 2 FIX), #1417/#1420/#1424 (held). |
Owner
Author
Juliusolsson05
marked this pull request as ready for review
September 27, 2026 13:44
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Integration batch H (owner-approved merge mode). Built after
git fetchfrom origin/main 6938e61 (batch G); each member merged --no-ff.Every member had GATE PASS (--member). Merged after green CI on the exact head, the disposition, and a recorded non-member merge-gate PASS (see the gate-record comment).
🤖 Generated with Claude Code