resources: the build's file set (RESOURCES slice 4) #137
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Context
Packaging, the lock's
resourcelines,--watch, the build report, theresources { }budget.Source
Acceptance criteria
Dependencies
Moved from the planning lists of the docs; the design record stays the specification.
[loop:torbscript] intake: in scope (author orion, label loop), source RESOURCES.md slice 4, not a duplicate. Depends on #135 and #136. next: waiting (dependencies #135, #136) until 2026-10-10T04:00:00Z
ArrayListis a type overBuffer(step 3 of #55, commit 2 of 2) #526[loop:torbscript] claim implement (fix main: tier B conformance red on
576b4669, 4 programs) by session:loop-torbscript-20261028 until 2026-10-28T11:05ZArrayListis a type overBuffer(step 3 of #55, commit 2 of 2) #526[loop:torbscript] release implement: worker lost its execution context after pushing loop/137-fix-resource-conformance and opening PR #536 (head
711e3e27); CI on it is pending. Respawning a worker to carry PR #536 through CI.[loop:torbscript] claim implement (continuation: carry PR #536 through CI) by session:loop-torbscript-20261028 until 2026-10-28T11:35Z
[loop:torbscript] release implement: the continuation worker was killed after pushing head
91f5f4ato PR #536. CI on that head: tier B (linux-x64) failed after 11m31s (run 2052), tier A pending. Starting a fresh worker to read the tier B log and fix the remaining red.[loop:torbscript] claim implement (fix main: PR #536 tier B still red on
91f5f4a) by session:loop-torbscript-20261028 until 2026-10-28T15:05Z[loop:torbscript] implement: read tier B of run 2052 (job log): 333 of 334 conformance programs passed, the one failure was tests/conformance/resource-missing.trb ("has no allocation budget, allocated 217 blocks"); the earlier commits fixed bytes/name/text and the restore, but resource-missing itself still lacked its .allocations file. Pushed
403ee176"test(conformance): allocation budget for resource-missing" (tests/conformance/resource-missing.allocations = 217) on loop/137-fix-resource-conformance. Forge CI of the new head (run 2057, #536) has been queued for 45+ minutes with no job started (commit messages / what the change touches "waiting", gates blocked); the two CI runners show active and the other waiting run (#541) has queued since 08:36Z, so it looks like a runner-pool stall, not a code issue. Side note: main site:edge failure (run 2033, job 3) is a different cause: its log just stops mid-bootstrap while compiling mbedTLS (no test or assertion error), so it was a killed/timed-out build, not the resource-conformance failure. Worker tokens: not exposed to me (small run: ~1 fix commit, ~45 tool calls). next: waiting (forge CI of403ee176to run tier A and B; if CI stays queued, the runner pool needs a look).[loop:torbscript] close: PR #536 landed on main (
403ee176). Its tier A, tier B and gates were green on the PR head; the push CI of the landed commit is still pending. next: waiting (main CI on403ee176, then close with the commit link)[loop:torbscript] claim implement (main CI red again on
403ee176: bootstrap (linux-x64) failed after 12s, pack the seed and gates follow; diagnose and fix) by session:loop-torbscript-20261028 until 2026-10-28T17:55Z[loop:torbscript] implement: no repo defect. The push CI of main
403ee176(run 2060) failed because the forge's ingress proxy was OOM-killed and restarted at about 10:51 UTC (last restart ended 10:52:46 UTC):bootstrap (linux-x64)died in fetch-seed.sh withcurl: (18) Transferred a partial filewhile downloading seed142293fd98(the asset itself downloads fine again, HTTP 200);pack the seedthen failed withFailed to ListArtifacts: Not Foundbecause bootstrap never uploaded its artifact;gatesfailed as the dependent job. Tier A/B were skipped (needs: bootstrap), not failed. The same bootstrap job was green on the PR head (run 2057 / 2103), so this is infra, not code. The infra side is tracked in the ops-alert for the proxy memory limit (comment added there). No branch or PR from me: the fix is not in this repo, and I may not push to main or re-run workflows on it. next: waiting (a green re-run of main CI needs the next commit on main or a re-run by Torben/Argus once the proxy is stable; the proxy still restarted 4 times in the last 4.5 h, so an immediate re-run can flake again). Tokens: ~75k.[loop:torbscript] close: RESOURCES slice 4 landed with PR #536 (commit
403ee176); main CI is green on0ac069f9, which contains it (the earlier red push CI of403ee176was the forge proxy OOM, not a repo defect). next: done