Confusing dryrun failures with no review? |
||
Issue descriptionCQ Dryrun builds will fail, if there is not a code review present, but the main purpose of the dryrun is to see if it would pass after being reviewed. This makes the failures confusing. So.... could we not enforce the review test for dryrun builds? An example failure: https://ci.chromium.org/p/infra-internal/builders/luci.infra-internal.try/build-internal-presubmit/b8942732168658640656
,
Jun 26 2018
As the one who actually implemented this, it already WAI: CQ sets special "dry_run=True" property for dry run builds, presubmit recipe passes it along to canned presubmit checks, which in turn reaches OWNERS check which turns error into warning. Let's see the end of the referenced builder log: This is a dry run, but these failures would be reported on commit: Missing LGTM from someone other than dgarrett@google.com ** Presubmit ERRORS ** /b/s/w/ir/cache/builder/build_internal_presubmit/build_internal/../build/tests/masters_test.py (73.44s) failed 1 masters were expected but not found: master.chromeos.chrome Indeed, missing OWNERS approval isn't the reason for failure. It's the fact that master name wasn't removed from masters_test.py :)
,
Jul 2
Ah... sorry. ;> |
||
►
Sign in to add a comment |
||
Comment 1 by qyears...@chromium.org
, Jun 25 2018