Adds a limited integration test in the runner which verifies that the embedded cache server starts up, can be written to by an action, and can be read by a subsequent action. This is a solid base foundation for future nearly-end-to-end tests.
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- other
- [PR](https://code.forgejo.org/forgejo/runner/pulls/889): <!--number 889 --><!--line 0 --><!--description dGVzdDogYWRkIGFuIGludGVncmF0aW9uIHRlc3QgZm9yIGVtYmVkZGVkIHJ1bm5lciBjYWNoZQ==-->test: add an integration test for embedded runner cache<!--description-->
<!--end release-notes-assistant-->
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/889
Reviewed-by: earl-warren <earl-warren@noreply.code.forgejo.org>
Co-authored-by: Mathieu Fenniak <mathieu@fenniak.net>
Co-committed-by: Mathieu Fenniak <mathieu@fenniak.net>
When run in parallel they will race to populate the Go cache in a different way (GOPROXY is not configured the same) and that will lead to transient errors when one job tries to use the cache that is populated differently by the other job.
The chain of `needs` is adjusted accordingly and the required checks too.
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- other
- [PR](https://code.forgejo.org/forgejo/runner/pulls/894): <!--number 894 --><!--line 0 --><!--description Y2hvcmU6IG1lcmdlIEFDVCBhbmQgcnVubmVyIHVuaXQgJiBpbnRlZ3JhdGlvbiB0ZXN0cyB0b2dldGhlcg==-->chore: merge ACT and runner unit & integration tests together<!--description-->
<!--end release-notes-assistant-->
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/894
Reviewed-by: Mathieu Fenniak <mfenniak@noreply.code.forgejo.org>
Co-authored-by: Earl Warren <contact@earl-warren.org>
Co-committed-by: Earl Warren <contact@earl-warren.org>
The `vars.ROLE` in each repository was set as follows:
- https://code.forgejo.org/forgejo/runner => forgejo-coding
- https://code.forgejo.org/forgejo-integration/runner => forgejo-integration
- https://invisible.forgejo.org/forgejo/runner => forgejo-release
It reflects how Forgejo roles are dispatched. Update all workflows to
use the ROLE to determine if it is relevant for them to run. It is
more straightforward than the previous approach based on the name of
the owner.
There were cases where the logic was wrong and workflows ran
needlessly, for instance because the runner release is now published
from an organization that is forgejo and not forgejo-release.
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- other
- [PR](https://code.forgejo.org/forgejo/runner/pulls/790): <!--number 790 --><!--line 0 --><!--description Y2hvcmUoY2xlYW51cCk6IGRpc3BhdGNoIHdvcmtmbG93cyBhY2NvcmRpbmcgdG8gdGhlIHJvbGUgb2YgdGhlIHJlcG9zaXRvcnk=-->chore(cleanup): dispatch workflows according to the role of the repository<!--description-->
<!--end release-notes-assistant-->
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/790
Reviewed-by: Michael Kriese <michael.kriese@gmx.de>
Co-authored-by: Earl Warren <contact@earl-warren.org>
Co-committed-by: Earl Warren <contact@earl-warren.org>
It matters because /bin/s6-svscan changed to /usr/bin/s6-svscan from v10 to v11 and it would fail in a confusing way if the admin tries to upgrade from v9 to v11.
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/542
Co-authored-by: Earl Warren <contact@earl-warren.org>
Co-committed-by: Earl Warren <contact@earl-warren.org>
