Repository navigation
t27b-lab: take up to two lane requests per master run (Closes #8586) - #8588
Merged
Merged
Conversation
The lab ran every new master commit first and then one lane request, so a queue of four waited up to four master runs (1.5-2 h per receipt) while master took a merge every few minutes. After each master run the lab now takes up to two admitted requests. One line edited, net 0 hand-written lines (#8236). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag
enabled auto-merge (squash)
October 10, 2026 13:18
Contributor
This was referenced Oct 10, 2026
Open
gHashTag
added a commit
that referenced
this pull request
Oct 10, 2026
…s not dropped (Closes #8639) (#8643) On 2026-10-10 at 14:09Z, t27b-lab's request queue held 4 of 4 with four lane agents asking for post-merge receipts, so a fifth would have been dropped as QUEUE_FULL. Since #8588 a master run is followed by up to two requests, and a request expires after 6 h, so eight drain in about four master cycles. .t27 only: the constant, its comment and the tests; bootstrap/gen/rust/verified/corpus_receipt.rs regenerated, the seal re-saved (32 of 32). On the t27c lab, admit gives ADMIT/0 for index 7, QUEUE_FULL/1 for index 8, WAIT_ORIGIN/2 and EXPIRED/1 as before. Part of the efficiency plan (see #8637). Co-authored-by: Claude <claude@anthropic.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag
added a commit
that referenced
this pull request
Oct 10, 2026
… master, not the master last run (Closes #8785) (#8786) lane_request() took the first-parent line of the master the lab LAST RAN, so a squash commit that landed after that run read NOT_ON_ORIGIN until the next master run, waited as WAIT_ORIGIN and was dropped once older than the grace (6a7b774 on 2026-10-10). It now fetches origin's REF into refs/remotes/origin/<REF> (the checkout does not move) and reads that ref's first-parent line, plus the ls-remote heads. A failed fetch leaves the ref at an older master, never at a lane sha, so on origin can only shrink. lab.py: 6 lines rewritten, net 0. Requests still skip the fuzzer (#8268) and up to two run per master run (#8588). corpus_receipt.t27: the comments that said the lab judges on-origin against the master it last ran are corrected, and a test pins the verdicts for a squash newer than the last master run. request_verdict and request_exit do not change. gen-rust regenerated and the seal re-saved (33 of 33) with master's t27c on t27b-lab, after a control run reproduced master's file and seal exactly. Co-authored-by: Claude <claude@anthropic.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
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
2 participants
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.
Closes #8586
Part of epic #8326 (why the loop idles, how to speed up).
The lab runs every new master commit first (15-22 min) and then ONE lane request. Master takes a merge every few minutes, so a queue of 4 requests waits up to 4 master runs: about 1.5-2 h per receipt. On 2026-10-09 to 2026-10-10 a lane agent waited over an hour for one receipt, and #8027's post-merge receipt took more than 2 h.
Change: after each master run, take up to two admitted requests instead of one. One line in contrib/railway/t27b-lab/lab.py, net 0 hand-written lines (#8236). Master runs still come first; their latest.json lags by at most one extra request run.
🤖 Generated with Claude Code