[work-dsl2] fix(ci): move the workflows from .gitea to .forgejo so they run #21

Merged

eden merged 1 commit from work-dsl2-awx-debian-ee-move-the-gitea-workflows into main

2026-10-26 19:53:27 +00:00

eden commented

2026-10-26 19:53:15 +00:00

Owner

Forgejo resolves workflows from the first directory it finds, in the order .forgejo/workflows > .gitea/workflows > .github/workflows (forgejo#9203, and the adoption trap in forgejo#14244). This repo gained .forgejo/workflows/close-bead.yml on 2026-09-26, which silently stopped build.yml and renovate.yaml in .gitea/workflows from ever being scheduled. No error, no status, nothing in the Actions tab.

Evidence from the run history:

last build run: 2026-09-24 (a pushed image)
last renovate schedule run: 2026-09-26 05:00
close-bead.yml landed: 2026-09-26 23:55
since then, only pull_request close-bead runs, including the merges of #18 and #19 on 2026-09-27, which never produced an image

So the execution environments have been frozen at the 2026-09-24 build for a month, and the community.hashi_vault, hvac and Fractal collection changes merged since are not in any published image.

This PR moves both workflows to .forgejo/workflows unchanged; README.md and .dockerignore follow the new path. Image tags already use ${{ github.repository }}, so publishing stays owner-relative. Merging this should make the build fire on main again, which is the verification.

Bead work-dsl2, epic work-thsk. Small routine CI change, author self-merge.

Forgejo resolves workflows from the first directory it finds, in the order `.forgejo/workflows` > `.gitea/workflows` > `.github/workflows` (forgejo#9203, and the adoption trap in forgejo#14244). This repo gained `.forgejo/workflows/close-bead.yml` on 2026-09-26, which silently stopped `build.yml` and `renovate.yaml` in `.gitea/workflows` from ever being scheduled. No error, no status, nothing in the Actions tab. Evidence from the run history: - last `build` run: 2026-09-24 (a pushed image) - last `renovate` schedule run: 2026-09-26 05:00 - `close-bead.yml` landed: 2026-09-26 23:55 - since then, only `pull_request` close-bead runs, including the merges of #18 and #19 on 2026-09-27, which never produced an image So the execution environments have been frozen at the 2026-09-24 build for a month, and the `community.hashi_vault`, `hvac` and Fractal collection changes merged since are not in any published image. This PR moves both workflows to `.forgejo/workflows` unchanged; `README.md` and `.dockerignore` follow the new path. Image tags already use `${{ github.repository }}`, so publishing stays owner-relative. Merging this should make the build fire on main again, which is the verification. Bead work-dsl2, epic work-thsk. Small routine CI change, author self-merge.

eden added 1 commit

2026-10-26 19:53:15 +00:00

[work-dsl2] fix(ci): move the workflows from .gitea to .forgejo so they run

Close bead / close (pull_request) Successful in 7s

Details

0f79c610cc

Forgejo resolves workflows from the first directory it finds, in the order
.forgejo/workflows > .gitea/workflows > .github/workflows (forgejo#9203).
Once .forgejo/workflows/close-bead.yml landed on 2026-09-26, the two workflows
still sitting in .gitea/workflows stopped being scheduled, silently: no error,
no status, nothing in the Actions tab.

Evidence: last build run 2026-09-24, last renovate schedule 2026-09-26 05:00,
and only pull_request close-bead runs since, including the merges of #18 and
#19 on 2026-09-27 which therefore never produced an image.

Both workflows move to .forgejo/workflows unchanged. README and .dockerignore
follow the new path. The image tags already use ${{ github.repository }}, so
publishing stays owner-relative.

eden merged commit 65d47b2c75 into main

2026-10-26 19:53:27 +00:00

eden deleted branch work-dsl2-awx-debian-ee-move-the-gitea-workflows

2026-10-26 19:53:27 +00:00

eden referenced this pull request from a commit

2026-10-26 19:53:28 +00:00

[work-dsl2] fix(ci): move the workflows from .gitea to .forgejo so they run (#21)