Add a ci-complete gate job and bound job runtimes - #65
Open
Gary Hsu (bghgary) wants to merge 2 commits into
Open
Add a ci-complete gate job and bound job runtimes#65Gary Hsu (bghgary) wants to merge 2 commits into
Gary Hsu (bghgary) wants to merge 2 commits into
Conversation
Bound apt's retries and transfer timeouts so a degraded in-datacenter mirror fails fast instead of stalling silently for hours, and cap every job with timeout-minutes so a future stall is a fast diagnosable failure rather than a 6h timeout. Install clang-14 alone; clang++-14 is not a package name, so apt fell through to regex matching and pulled in 20 packages where 2 were meant. ci-complete aggregates the four jobs behind one stable check name that branch protection can require, leaving the matrix-generated names free to change. if: always() is required or the job is skipped when a dependency fails, and a skipped required check satisfies the requirement. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
The runner image deliberately sets APT::Acquire::Retries 10; overriding it to 1 trades resilience against ordinary transient failures for a faster fail in one pathological case. actions/runner-images#14596 fixes the root cause image-side by demoting the azure mirror, which a workflow cannot do. timeout-minutes already bounds the hang, and it bounds every other hang too. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Gary Hsu (bghgary)
changed the title
Gary Hsu (bghgary)
marked this pull request as ready for review
August 19, 2026 23:49
There was a problem hiding this comment.
Pull request overview
Adds a stable “aggregate” GitHub Actions check (ci-complete) that can be required by branch protection without being impacted by matrix job name churn, and bounds CI job runtimes to avoid indefinite hangs that would otherwise block merges once checks are required.
Changes:
- Add
timeout-minutesto existing Windows/macOS/iOS/Linux jobs to bound runtime. - Refactor Linux compiler matrix to install compiler toolchains via a dedicated
packageskey (avoids treatingclang++-14as an apt package). - Add a
ci-completegate job that always runs and fails unless all dependent jobs reportsuccess.
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Branimir Karadžić (bkaradzic-microsoft)
approved these changes
Aug 20, 2026
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.
[Created by Copilot on behalf of Gary Hsu (@bghgary)]
Context
masterrequires no status checks, so an approval alone can merge a PR whose CI is red or never ran. Requiring the current checks as-is is not viable: all 14 names are matrix-generated, so any matrix edit strands a required check that never reports and leaves PRs blocked with no visible cause.ci-completeis one stable name to require instead.timeout-minutesis a precondition for that, not a separate cleanup. Three Linux jobs hung for ~6h on 19 August with no output at all before the 360-minute default timeout (actions/runner-images#14594). Once a check is required, an unbounded hang is not a slow build — it is an unmergeable PR with no failure to act on.Worth a look
if: always()onci-completeis load-bearing. Without it the job is skipped when a dependency fails, and a skipped required check satisfies the requirement — inverting the gate. Two workflows differing only in that line producedskippedagainstfailure.needs.*.resultmust equalsuccess— so a result value we have not anticipated fails closed rather than passing through.clang++-14is not a package name. apt fell through to regex matching and installed 20 packages instead of 2, includingclang-14-docandclang-14-examples;clang-14alone ships theclang++-14binary.cxxstill feeds CMake, so the install list moves to a separatepackageskey.APT::Acquire::Retries "10"; overriding it would trade resilience against ordinary transient failures for a faster fail in one case, and the root cause is fixed image-side by actions/runner-images#14596.Follow-up
Once
ci-completehas reported here at least once, it can be added tomaster's required status checks.