Extract from #34531
## Move Commit status state to a standalone package
Move the state from `structs` to `commitstatus` package. It also
introduce `CommitStatusStates` so that the combine function could be
used from UI and API logic.
## Combined commit status Changed
This PR will follow Github's combined commit status. Before this PR,
every commit status could be a combined one.
According to
https://docs.github.com/en/rest/commits/statuses?apiVersion=2022-11-28#get-the-combined-status-for-a-specific-reference
> Additionally, a combined state is returned. The state is one of:
> failure if any of the contexts report as error or failure
> pending if there are no statuses or a context is pending
> success if the latest status for all contexts is success
This PR will follow that rule and remove the `NoBetterThan` logic. This
also fixes the inconsistent between UI and API. In the API convert
package, it has implemented this which is different from the UI. It also
fixed the missing `URL` and `CommitURL` in the API.
## `CalcCommitStatus` return nil if there is no commit statuses
The behavior of `CalcCommitStatus` is changed. If the parameter commit
statuses is empty, it will return nil. The reference places should check
the returned value themselves.
1. add `GetSiteCookieFlashMessage` to help to parse flash message
2. clarify `handleRepoHomeFeed` logic
3. remove unnecessary global variables, use `sync.OnceValue` instead
4. add some tests for `IsUsableUsername` and `IsUsableRepoName`
By some CI fine tunes (`run tests`), SQLite & MSSQL could complete
in about 12~13 minutes (before > 14), MySQL could complete in 18 minutes
(before: about 23 or even > 30)
Major changes:
1. use tmpfs for MySQL storage
1. run `make test-mysql` instead of `make integration-test-coverage`
because the code coverage is not really used at the moment.
1. refactor testlogger to make it more reliable and be able to report
stuck stacktrace
1. do not requeue failed items when a queue is being flushed (failed
items would keep failing and make flush uncompleted)
1. reduce the file sizes for testing
1. use math ChaCha20 random data instead of crypot/rand (for testing
purpose only)
1. no need to `DeleteRepository` in `TestLinguist`
1. other related refactoring to make code easier to maintain
Gitea instance keeps reporting a lot of errors like "LFS SSH transfer connection denied, pure SSH protocol is disabled". When starting debugging the problem, there are more problems found. Try to address most of them:
* avoid unnecessary server side error logs (change `fail()` to not log them)
* figure out the broken tests/user2/lfs.git (added comments)
* avoid `migratePushMirrors` failure when a repository doesn't exist (ignore them)
* avoid "Authorization" (internal&lfs) header conflicts, remove the tricky "swapAuth" and use "X-Gitea-Internal-Auth"
* make internal token comparing constant time (it wasn't a serous problem because in a real world it's nearly impossible to timing-attack the token, but good to fix and backport)
* avoid duplicate routers (introduce AddOwnerRepoGitLFSRoutes)
* avoid "internal (private)" routes using session/web context (they should use private context)
* fix incorrect "path" usages (use "filepath")
* fix incorrect mocked route point handling (need to check func nil correctly)
* split some tests from "git general tests" to "git misc tests" (to keep "git_general_test.go" simple)
Still no correct result for Git LFS SSH tests. So the code is kept there
(`tests/integration/git_lfs_ssh_test.go`) and a FIXME explains the details.
