fix(forge): make the live-Linear oracle deterministic + gate it on the forge surface (RIG-2909) - #721
Open
rigel-mintaka wants to merge 1 commit into
Open
Conversation
|
Compass engineering docs preview: https://compass-forge-rig-2909-forge.compass-eng-docs.pages.dev Deployed from |
…e forge surface (RIG-2909) The forge-oracle gate intermittently failed on TestLiveLinearListIssues: its live-Linear setup CreateIssue timed out awaiting response headers (a transient api.linear.app latency blip), failing the whole required gate on unrelated PRs. Two deterministic fixes, no retries=N (rule://no-retries): 1. Every live GitHub create is already wrapped in the bounded, ctx-aware one-shot createWithBackoff; none of the five Linear creates were. Extend createWithBackoff to also re-issue once on a transient network timeout (net.Error.Timeout / context.DeadlineExceeded through the client-timeout chain) alongside the existing GitHub secondary-rate-limit case, and wrap the five Linear creates in it. A persistent outage still fails loud after one re-issue; it never runs on the skip path. 2. Make the live oracle path-affected. It is the expensive third-party verification layered on the untagged golden-replay battery (leg 1), which already asserts the client contract on every PR with zero network. Narrow the forge trigger (ci-matrix FORGE_PATH_RE + the ci.yml in-step gate) to the forge contract surface go/internal/forge/** and drop the bare-ci.yml over-trigger that ran the whole live oracle on docs/CI-only PRs (e.g. #703). Oracle-wiring changes stay covered by the push/schedule full sweep. Refs RIG-2909 Co-authored-by: Matt Wilkinson <matt@rigel.build>
rigel-mintaka
force-pushed
the
compass-forge/rig-2909-forge-oracle-flake
branch
from
August 28, 2026 17:08
efec521 to
f716868
Compare
rigel-mintaka
marked this pull request as ready for review
August 28, 2026 17:16
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
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.
The forge-oracle gate intermittently failed on TestLiveLinearListIssues:
its live-Linear setup CreateIssue timed out awaiting response headers (a
transient api.linear.app latency blip), failing the whole required gate on
unrelated PRs. Two deterministic fixes, no retries=N (rule://no-retries):
Every live GitHub create is already wrapped in the bounded, ctx-aware
one-shot createWithBackoff; none of the five Linear creates were. Extend
createWithBackoff to also re-issue once on a transient network timeout
(net.Error.Timeout / context.DeadlineExceeded through the client-timeout
chain) alongside the existing GitHub secondary-rate-limit case, and wrap
the five Linear creates in it. A persistent outage still fails loud after
one re-issue; it never runs on the skip path.
Make the live oracle path-affected. It is the expensive third-party
verification layered on the untagged golden-replay battery (leg 1), which
already asserts the client contract on every PR with zero network. Narrow
the forge trigger (ci-matrix FORGE_PATH_RE + the ci.yml in-step gate) to
the forge contract surface go/internal/forge/** and drop the bare-ci.yml
over-trigger that ran the whole live oracle on docs/CI-only PRs (e.g.
docs(compass-gtk4): T6 prose sweep + GTK4-default ledger encode (RIG-2820) #703). Oracle-wiring changes stay covered by the push/schedule full sweep.
Refs RIG-2909
Co-authored-by: Matt Wilkinson matt@rigel.build