RI-000: keep manual trigger only for e2e #4995
Conversation
ArtemHoruzhenko
requested review from
KrumTy,
pawelangelow,
dantovska,
pd-redis and
valkirilov
as code owners
| @@ -1,13 +1,6 @@ | |||
| name: ✅ E2E Tests | |||
| run-name: > | |||
| E2E • ${{ github.event_name == 'pull_request_review' | |||
There was a problem hiding this comment.
we don't need this code, basically. there are checks github.event_name == 'pull_request_review' that will never satisfy
| target: build_linux_appimage_x64 | ||
| debug: ${{ inputs.debug || false }} | ||
|
|
||
| builds-complete: |
There was a problem hiding this comment.
I see your point of removing this, as a dependency for both the tests-e2e-playwright and e2e-appimage-tests jobs, because each of them depends on a separate build step.
But in the end, we don't have a way to trigger only one of these steps. As I understand, they will always run together, and if one of them fails, it means the whole pipeline goes red (yes, we can re-run them individually, but overall it means "tests didn't pass").
I believe this is why @KrumTy made this builds-complete step, to ensure that we have a proper setup for running the tests and save some time and resources, in case the build fails. I'm not against either of the approaches. I'm just pointing this out since it's relatively new code that we’re reverting now.
There was a problem hiding this comment.
I reverted a lot of new code :)
I understand your point but it shouldn't work this way. We shouldn't wait until not real dependency will be build before starting tests. I think we might separate desktop and web tests in the future. For now IMO it is enough to keep as is but no "builds-complete"
No description provided.