Skip to content

gunbc serve: hand peer_user only on the unix-socket listener (fixes srv1 roadmap 500) - #12627

Merged
gunbai-bot[bot] merged 5 commits into
mainfrom
session/jolly-owl-158-peerfix
Sep 30, 2026
Merged

gunbai-bot[bot] merged 5 commits into
mainfrom
session/jolly-owl-158-peerfix

Conversation

@gunbai-bot

@gunbai-bot gunbai-bot Bot commented Sep 29, 2026

Copy link
Copy Markdown
Contributor

This fixes the srv1 roadmap outage. After dashboard_deploy run 36584004074 at 178fb76c71, gunbc-roadmap.service answered HTTP 500 on every request. A rollback to 361db9017b is in progress separately.

Cause. #12482 made gunbc serve hand a new named argument, peer_user, to every handler. Named arguments must match the handler's parameters exactly, so an argument the handler doesn't declare is a call-contract refusal (no parameter named 'peer_user'), not a value that gets dropped. The roadmap handler doesn't declare it, so every request failed. The seed comment claiming otherwise was never executed.

Fix. serve_handler_args builds the argument list and appends peer_user only when the listener attests a peer, which means --unix-socket. TCP handlers get exactly the argument set they had before #12482. The fabric door's socket handler still receives peer_user.

Evidence

🤖 Generated with Claude Code

Brian Searls and others added 4 commits September 29, 2026 15:21
…admap 500 incident)

#12482 passed peer_user to every served handler; an undeclared named argument is a call-contract
refusal, so gunbc-roadmap answered HTTP 500 on every request after dashboard_deploy run 36584004074.
serve_handler_args now appends peer_user only when the listener attests a peer. Seed test
a_tcp_handler_is_not_handed_peer_user pins the argument names per listener; a live gunbc serve of a
handler without peer_user answers 200 (500 on the defective seed). Files the recurring failure mode.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…repair it

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ath stands (review 72786)

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@gunbai-bot

gunbai-bot Bot commented Sep 29, 2026

Copy link
Copy Markdown
Contributor Author

Review 72786 addressed in 711c851: the row now states RUNG NOW 1 (mitigatable). The seed test is a v1-compiler --lib unit test that blocks no merge while gunbc.rung_drop rust_unit_tests_off_the_merge_path stands, so it's local diligence, not a mechanism. The row says rung 2 needs that drop retired, or the required per-serve-unit argument/signature claim it already names as the next trigger. — sent from jolly-owl-158

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@gunbai-bot
gunbai-bot Bot added this pull request to the merge queue Sep 29, 2026
Merged via the queue into main with commit 8e0c1fa Sep 30, 2026
5 checks passed
@gunbai-bot
gunbai-bot Bot deleted the session/jolly-owl-158-peerfix branch September 30, 2026 00:47
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.

0 participants