ci(release): move :latest to the highest pre-release while no stable release exists

Only pre-releases (v0.1.0-rc.1, rc.2) are published, and :latest followed the highest stable release only, so the
Unraid template and the compose file, which pull :latest, found no image while :nightly was pushed every night.

The new Chronos.Tools command release-channels reads the published releases (stable and pre-releases), ignores drafts
and tags that are not SemVer, counts the tag itself as published and compares by SemVer 2.0 precedence. It prints
latest (the public releases repository: highest stable release only) and image_latest (:latest: the highest stable
release, or the highest pre-release while none is stable). Re-publishing an older tag never rolls either back.
release.yml uses it in place of curl|grep|sort -V; the image job pushes :latest on image_latest. README, the public
README, the compose file and ARCHITECTURE explain :latest, :<version> and :nightly.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Alby96andClaude Opus 5.5 committed 2026-10-26 21:45:57 +01:00
1 parent 5209579631
commit 3679e80ebb
8 files changed
+343 -31

No files matched your search

+4 -4
View File
@@ -16,7 +16,7 @@ Companion to `docs/BRIEF.md` (scope, behaviour) and `docs/ui/DESIGN.md` (looks).
| `src/Chronos.App` | WPF exe `Chronos.exe` | Desktop client |
| `src/Chronos.Uploader` | WPF exe | Uploader (tray, watch folders, resumable queue) |
| `src/Chronos.Setup` | WPF exe | In-house installer/uninstaller (TPK1 payload) |
| `src/Chronos.Tools` | net10.0 exe | CLI: `make-icon`, `i18n` (check, export, missing, import), `xaml-lint`, `package-guard`, `changelog`, `pack-installer`, `selftest`, `diagnose`, `fixtures`, `screenshots` (diff, accept), `zip`, `gitea-release`, `postgres-ci`, `pe-check` |
| `src/Chronos.Tools` | net10.0 exe | CLI: `make-icon`, `i18n` (check, export, missing, import), `xaml-lint`, `package-guard`, `changelog`, `pack-installer`, `selftest`, `diagnose`, `fixtures`, `screenshots` (diff, accept), `zip`, `gitea-release`, `release-channels`, `postgres-ci`, `pe-check` |
| `tests/Chronos.Testing` | net10.0 | In-house test framework (`[Test]`, `[Theory]`+`[Case]`, `[Category]`, `[RequiresEnvironment]`, `[WindowsOnly]`, `[Serial]`, `Assert`, `TestContext`, JUnit) |
| `tests/Chronos.*.Tests` | exe | One test program per area; `dotnet run --project tests/X -- --junit out.xml` |
@@ -240,8 +240,8 @@ run as server jobs whose progress comes from the events.
building it) and the 100,000-frame rebuild from sidecars (400 sequences of 250 real photos, every column compared,
< 120 s), each on SQLite and PostgreSQL. `.gitea/workflows/perf.yml` runs them weekly against PostgreSQL 17.
- `Chronos.Tools.Tests` covers the tools whose output CI relies on (fixtures warning lines and exit codes, the
screenshot diff, the PostgreSQL preparation, `pe-check`, `zip`, the Gitea release upload) and the test runner's
outcomes.
screenshot diff, the PostgreSQL preparation, `pe-check`, `zip`, the Gitea release upload, `release-channels`) and the
test runner's outcomes.
- Visual regression: the CI job `windows` (or a run on a Windows PC) compares the `--capture` screenshots with
`tests/visual/baselines` (`Chronos.Tools screenshots diff`; its README says how baselines are added). Once any
baseline is committed, a screen without one fails too; before that the step warns that it compared nothing. The
@@ -280,7 +280,7 @@ dotnet format whitespace Chronos.slnx --verify-no-changes
| Workflow | Trigger | Does |
|---|---|---|
| `ci.yml` | push to `main`, pull requests, manual | Linux build and tests, guards, i18n, XAML lint, format, fixtures (warnings when missing), the Windows programs cross-compiled for win-x64, packed and checked (`pe-check`), PostgreSQL 16/17 (every authentication method, TLS), Docker smoke test, YAML check of the workflows, issue forms and compose file; with a Windows runner (`WINDOWS_RUNNER`): Windows tests, screenshots and their pixel diff, installer test (install, run, upgrade, uninstall) |
| `release.yml` | tag `v*` (or manual re-publish) | Tests, self-contained win-x64 publish cross-compiled on Linux, `pe-check`, optional signing (osslsigncode), installers, portable and server zips, SHA256SUMS, release notes, Gitea release (`gitea-release`), then the linux/amd64 image in the Gitea container registry (`:latest` only for the newest stable release); a separate releases repository is updated only when `RELEASES_REPO` names one |
| `release.yml` | tag `v*` (or manual re-publish) | Tests, self-contained win-x64 publish cross-compiled on Linux, `pe-check`, optional signing (osslsigncode), installers, portable and server zips, SHA256SUMS, release notes, Gitea release (`gitea-release`), then the linux/amd64 image in the Gitea container registry (`:latest` for the highest stable release, or the highest pre-release while none is stable: `release-channels`); a separate releases repository is updated only when `RELEASES_REPO` names one, for the highest stable release |
| `docker.yml` | nightly, manual | `gitea.encke-hake.ts.net/balbosoft/chronos-server:nightly` from `main` |
| `perf.yml` | weekly, manual | Performance gates (SQLite, PostgreSQL 17) with timings at the end of the log |