mirror of
https://github.com/mattermost/mattermost-mobile.git
synced 2026-11-04 09:58:57 +00:00
3e1331248c
* ci(e2e): say whether each main-run failure is new, still failing, flaky or a setup problem, from the triage verdict The channel message judged failures from their error text, beside the E2E triage jobs that judge the same failures from TSIO history, and the two disagreed: on main run 40764686027 it listed tests that passed on retry as needing a look. The message now waits for triage and reads each run's findings. On main, a test that passed every earlier main run and fails now is new, with the changes merged since the last main run named (TSIO's previous run, GitHub's compare); a run where most tests failed at once is called out first. Runs with no triage verdict keep the error-text judgement. * ci(e2e): judge from the error text each failed test triage named no verdict for, and say how many were not checked * ci(e2e): count each test once per run, so a test that failed and then passed its retry is flaky, not a failure * ci(e2e): a main-run failure triage calls flaky or a test problem is not blamed on new changes; a same-commit re-run says nothing was merged; a long PR list says it is partial * Fix TSIO current report selection Co-authored-by: yasserfaraazkhan <15979783+yasserfaraazkhan@users.noreply.github.com> * Fix lint comment spacing Co-authored-by: yasserfaraazkhan <15979783+yasserfaraazkhan@users.noreply.github.com> * ci(e2e): update the TSIO upload action, so a failed screenshot upload no longer drops a shard's results * ci(e2e): an earlier attempt of the same main run is not the previous run, so a re-run still lists the changes merged since the last run * ci(e2e): drop the note from the TSIO upload action pin comment * ci(e2e): a re-run replaces the first attempt's triage findings, so the channel message never posts stale verdicts * ci(e2e): the main-run message says in a few lines which failures may be product bugs and which are not, and posts nothing when none may be * ci(e2e): main and CMT call a failure a product bug only when it fails again on a fresh device, and clear what the error text proves is not the app * ci(e2e): a re-run hands Android Detox and Maestro results to the channel message too * ci(e2e): CMT re-runs start after a failed lane; a spec that differs from main, or a feature the server has off, is not a product bug * ci(e2e): the CMT message plans the re-run as the planner job does, leaving out tests that differ from main * ci(e2e): one channel post per run, and a test's name can never mention the channel * ci(e2e): a re-run keeps the run's own report and runs only the failed flows * ci(e2e): a PR run posts to the channel only when something failed * ci(e2e): an E2E/Override PR keeps its statuses on a new push; a few simpler lines in the channel notify * ci(e2e): exclude isolated report-problem flows from reruns and prevent label/action concurrency collisions <!-- coderabbit-ai-generated:v1 --> * ci(e2e): pass support paths to E2E triage * ci(e2e): pin TSIO report-upload to release-0.16.0 (main@3212731) Brings the TSIO action pin up to the latest release, which includes mattermost-test-system-io#129, #130 and #124. The report-upload action itself is unchanged between the two pins. * ci(e2e): re-run E2E jobs once when a lane is red only on infrastructure Triage (rerun-attempts: 1, with actions: read to read the failed jobs) asks for one re-run when a lane is red only because jobs died or reports never arrived; a build or test-server failure holds it. e2e-tests-rerun.yml carries the re-run out once the run finishes. CMT keeps its own re-run of failed specs. --------- |
||
|---|---|---|
| .. | ||
| await-results-service | ||
| bandwidth-throttling | ||
| build-android-maestro-apk | ||
| compute-e2e-build-fingerprint | ||
| e2e-override-label | ||
| e2e-override-status | ||
| e2e-triage | ||
| e2e-verified-status | ||
| generate-specs | ||
| prepare-android-build | ||
| prepare-ios-build | ||
| prepare-low-bandwidth | ||
| prepare-mobile-build | ||
| prepare-node-deps | ||
| provision-e2e-servers | ||
| s3-build-cache | ||
| s3-spec-durations | ||
| setup-intune | ||
| setup-intune-ssh-key | ||
| setup-ssh-key | ||
| start-proxy | ||
| test | ||
| test-coverage | ||
| update-commit-status | ||