← Back to issue list

ci: test starflow run-spread action in spread workflows

View original Github issue

Metadata

Project
snapcraft
Number
#6545
Type
pull request
State
open
Author
mr-cal-bot
Labels
PR: Run Manual Spread
Created
Updated
Closed
—

Current evaluation

No evaluation has been recorded for this issue yet.

Issue body

Test PR for the new [run-spread](https://github.com/canonical/starflow/tree/work/junit-for-spread/run-spread) action in starflow (branch `work/junit-for-spread`), which runs spread and publishes the results as JUnit XML artifacts. ## Changes - **spread.yml**: both integration jobs now use `canonical/starflow/run-spread@work/junit-for-spread` instead of calling `spread` directly. The Ubuntu Pro token check is split into its own gating step so the action step stays declarative. - **spread-manual.yaml**: all seven jobs (kernel, initrd, remote-build, colcon, extensions, snapd-bases, squid-proxy) converted, each with a unique `junit-name`. - Every spread step is followed by an `if: always()` artifact upload of `results/junit*.xml`. - **spread-docs.yaml** is not converted: its `spread openstack:...docs...` invocation is a placeholder. ## What to look for - JUnit XML artifacts (`test-results-*`) attached to failed and successful runs. - Failing tasks carry spread's error message and the `-----` error block as failure text. - Aborted/never-started tasks (e.g. after an environment prepare failure) show up as skipped cases named from `spread -list`. - `spread -list` failures degrade to a warning, not a job failure. - On this fork PR the secret-gated paths (Ubuntu Pro, store, UA) are skipped by design; the core spread runs still exercise the action. Do not merge: this exists to exercise the action against real spread workflows. Once the starflow side lands, re-pin to `canonical/starflow/run-spread@main` or close.

Evaluation history

No evaluation history available.

Update history

Date Change
updated
updated
updated
created

Related issues

No embedding available for this issue yet. Run scripts/run_llm.py evaluate to enable similarity search.