Set up a CI runner and prove the workflow runs #20
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Context
.forgejo/workflows/ci.ymlhas two jobs:fedora:44container, runsbuild-aux/ci.sh(fmt, clippy with-D warnings, build, tests, release build),then
meson setup/compile/test/install;build-aux/terminalko.spec, uploading both as artifacts.Neither has ever run. The repository has no runner configured, so the file is
untested YAML. The steps themselves are known to work: they are the same commands
used locally, and the rpm they build has been built by hand.
Work
forgejo-runner), with the Dockerbackend so the
fedora:44container in the workflow works. On this Forgejoinstance that is a settings page plus a token, and a runner somewhere that can
reach it.
show up in a container: no display (already accounted for — the smoke test is
deliberately not in CI), no
podman/docker/toolbox(the discovery tests arewritten to tolerate absent runtimes, and
a_resize_reaches_a_terminal_nested_in_anotherskips without
systemd-run— verify both actually skip rather than fail), anddnf installtime, which is most of the run.~/.cargoandtargetbetween runs, or the run will bedominated by rebuilding gtk4-rs every time.
Acceptance criteria
failing, and the skip is visible in the log.
README.mdmentions where CI runs, so the next person knows it is real.Notes
build-aux/smoke.sh) needs a display and is intentionally out ofCI. If a runner with a compositor ever exists, adding it is worthwhile: it is the
test that catches template and dialog mistakes.