Backport #32560 by @lunny
PushMirrors only be used in the repository setting page. So it should
not be loaded on every repository page.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #32531 by @lunny
WebAuthn should behave the same way as TOTP. When enabled, basic auth
with username/password should need to WebAuthn auth, otherwise returned
401.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #32186 by @lunny
Since there is a status column in the database, the transaction is
unnecessary when downloading an archive. The transaction is blocking
database operations, especially with SQLite.
Replace #27563
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #32424 by @Zettat123
Resolve#32232
Users can disable the "Actions" unit for all mirror repos by running
```
gitea doctor check --run disable-mirror-actions-unit --fix
```
Co-authored-by: Zettat123 <zettat123@gmail.com>
Backport #32432 by @kemzeb
Resolves#32371.
#31970 should have just showed the commit summary, but
`strings.SplitN()` was misused such that we did not perform any
splitting at all and just used the message. This was not caught in the
unit test made in that PR since the test commit summary was > 50 (which
truncated away the commit description).
This snapshot resolves this and adds another unit test to ensure that we
only show the commit summary.
Co-authored-by: Kemal Zebari <60799661+kemzeb@users.noreply.github.com>
Backport #32075 by @lunny
After migrating a repository with pull request, the branch is missed and
after the pull request merged, the branch cannot be deleted.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31946 by @lunny
Fix#13489
In the original implementation, only `All` will display your owned and
collaborated repositories. For other filters like `Source`, `Mirrors`
and etc. will only display your owned repositories.
This PR removed the limitations. Now except `collbrations`, other
filters will always display your owned and collaborated repositories.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #32066 by @lunny
The previous implementation requires admin permission to rename branches
which should be write permission.
Fix#31993
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #32099 by @maantje
This PR addresses the missing `bin` field in Composer metadata, which
currently causes vendor-provided binaries to not be symlinked to
`vendor/bin` during installation.
In the current implementation, running `composer install` does not
publish the binaries, leading to issues where expected binaries are not
available.
By properly declaring the `bin` field, this PR ensures that binaries are
correctly symlinked upon installation, as described in the [Composer
documentation](https://getcomposer.org/doc/articles/vendor-binaries.md).
Co-authored-by: Jamie Schouten <j4mie@hey.com>
Backport #32050 by @KN4CK3R
Fixes#31937
- Add missing comment reply handling
- Use `onGiteaRun` in the test because the fixtures are not present
otherwise (did this behaviour change?)
Compare without whitespaces.
Co-authored-by: KN4CK3R <admin@oldschoolhack.me>
Backport #31841 by @kemzeb
A 500 status code was thrown when passing a non-existent target to the
create release API. This snapshot handles this error and instead throws
a 404 status code.
Discovered while working on #31840.
Co-authored-by: Kemal Zebari <60799661+kemzeb@users.noreply.github.com>
Backport #32025 by @wolfogre
Fix#32024. Follow #27655.
After this PR, all usage of "new dial context" needs to provide a proxy,
so I dropped the old `NewDialContext` and renamed
`NewDialContextWithProxy` to `NewDialContext`.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #32011 by @wolfogre
Replace #32001.
To prevent the context cache from being misused for long-term work
(which would result in using invalid cache without awareness), the
context cache is designed to exist for a maximum of 10 seconds. This
leads to many false reports, especially in the case of slow SQL.
This PR increases it to 5 minutes to reduce false reports.
5 minutes is not a very safe value, as a lot of changes may have
occurred within that time frame. However, as far as I know, there has
not been a case of misuse of context cache discovered so far, so I think
5 minutes should be OK.
Please note that after this PR, if warning logs are found again, it
should get attention, at that time it can be almost 100% certain that it
is a misuse.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31931 by @lunny
All refs under `refs/pull` should only be changed from Gitea inside but
not by pushing from outside of Gitea.
This PR will prevent the pull refs update but allow other refs to be
updated on the same pushing with `--mirror` operations.
The main changes are to add checks on `update` hook but not
`pre-receive` because `update` will be invoked by every ref but
`pre-receive` will revert all changes once one ref update fails.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31754 by @lunny
When opening a repository, it will call `ensureValidRepository` and also
`CatFileBatch`. But sometimes these will not be used until repository
closed. So it's a waste of CPU to invoke 3 times git command for every
open repository.
This PR removed all of these from `OpenRepository` but only kept
checking whether the folder exists. When a batch is necessary, the
necessary functions will be invoked.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31874 by @bohde
In the OpenID flows, the "CfTurnstileSitekey" wasn't populated, which
caused those flows to fail if using Turnstile as the Captcha
implementation.
This adds the missing context variables, allowing Turnstile to be used
in the OpenID flows.
Co-authored-by: Rowan Bohde <rowan.bohde@gmail.com>
Backport #31825 by @Zettat123
Fix#31395
This regression is introduced by #30273. To find out how GitHub handles
this case, I did [some
tests](https://github.com/go-gitea/gitea/issues/31395#issuecomment-2278929115).
I use redirect in this PR instead of checking if the corresponding `.md`
file exists when rendering the link because GitHub also uses redirect.
With this PR, there is no need to resolve the raw wiki link when
rendering a wiki page. If a wiki link points to a raw file, access will
be redirected to the raw link.
---------
Co-authored-by: Zettat123 <zettat123@gmail.com>
Co-authored-by: yp05327 <576951401@qq.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31843 by @wolfogre
Fix#31625.
If `pull_service.NewPullRequest` return an error which misses each `if`
check, `CompareAndPullRequestPost` will return immediately, since it
doesn't write the HTTP response, a 200 response with empty body will be
sent to clients.
```go
if err := pull_service.NewPullRequest(ctx, repo, pullIssue, labelIDs, attachments, pullRequest, assigneeIDs); err != nil {
if repo_model.IsErrUserDoesNotHaveAccessToRepo(err) {
ctx.Error(http.StatusBadRequest, "UserDoesNotHaveAccessToRepo", err.Error())
} else if git.IsErrPushRejected(err) {
// ...
ctx.JSONError(flashError)
} else if errors.Is(err, user_model.ErrBlockedUser) {
// ...
ctx.JSONError(flashError)
} else if errors.Is(err, issues_model.ErrMustCollaborator) {
// ...
ctx.JSONError(flashError)
}
return
}
```
Not sure what kind of error can cause it to happen, so this PR just
expose it. And we can fix it when users report that creating PRs failed
with error responses.
It's all my guess since I cannot reproduce the problem, but even if it's
not related, the code here needs to be improved.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31832 by @yp05327
Fix#31807
ps: the newly added params's value will be changed.
When the first time you selected the filter, the values of params will
be `0` or `1`
But in pager it will be `true` or `false`.
So do we have `boolToInt` function?
Co-authored-by: yp05327 <576951401@qq.com>
Backport #31829 by @lunny
Fix#31730
This PR rewrote the function `PublicKeysAreExternallyManaged` with a
simple test. The new function removed the loop to make it more readable.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31770 by @emrebdr
When transferring repositories that have issues linked to a project
board to another organization, the issues remain associated with the
original project board. This causes the columns in the project board to
become bugged, making it difficult to move other issues in or out of the
affected columns. As a solution, I removed the issue relations since the
other organization does not have this project table.
Fix for #31538
Co-authored-by: Edip Emre Bodur <emrebdr29@gmail.com>
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31683 by @SimonPistache
As discussed in #31667 & #26561, when a card on a Project contains
images, they can overflow the card on its containing column. This aims
to fix this issue via snapping scrollbars.
---
Issue #31667 is open to discussion as there should be room for
improvement.
Co-authored-by: Simon Priet <105607989+SimonPistache@users.noreply.github.com>
Backport #31790 by @wolfogre
Fix#31271.
When gogit is enabled, `IsObjectExist` calls
`repo.gogitRepo.ResolveRevision`, which is not correct. It's for
checking references not objects, it could work with commit hash since
it's both a valid reference and a commit object, but it doesn't work
with blob objects.
So it causes #31271 because it reports that all blob objects do not
exist.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31778 by @lunny
Fix#31738
When pushing a new branch, the old commit is zero. Most git commands
cannot recognize the zero commit id. To get the changed files in the
push, we need to get the first diverge commit of this branch. In most
situations, we could check commits one by one until one commit is
contained by another branch. Then we will think that commit is the
diverge point.
And in a pre-receive hook, this will be more difficult because all
commits haven't been merged and they actually stored in a temporary
place by git. So we need to bring some envs to let git know the commit
exist.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31702 by @wolfogre
Fix#31137.
Replace #31623#31697.
When migrating LFS objects, if there's any object that failed (like some
objects are losted, which is not really critical), Gitea will stop
migrating LFS immediately but treat the migration as successful.
This PR checks the error according to the [LFS api
doc](https://github.com/git-lfs/git-lfs/blob/main/docs/api/batch.md#successful-responses).
> LFS object error codes should match HTTP status codes where possible:
>
> - 404 - The object does not exist on the server.
> - 409 - The specified hash algorithm disagrees with the server's
acceptable options.
> - 410 - The object was removed by the owner.
> - 422 - Validation error.
If the error is `404`, it's safe to ignore it and continue migration.
Otherwise, stop the migration and mark it as failed to ensure data
integrity of LFS objects.
And maybe we should also ignore others errors (maybe `410`? I'm not sure
what's the difference between "does not exist" and "removed by the
owner".), we can add it later when some users report that they have
failed to migrate LFS because of an error which should be ignored.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31725 by @wolfogre
Fix#31707.
It's split from #31724.
Although #31724 could also fix#31707, it has change a lot so it's not a
good idea to backport it.
Co-authored-by: Jason Song <i@wolfogre.com>
Backport #31672 by @techknowlogick
This lets developers who have direnv enabled to load our nix flake
automatically when entering it
Co-authored-by: techknowlogick <techknowlogick@gitea.com>
Backport #31659 by @Zettat123
Fix#31599Fix#31472
A branch divergence is counted based on the default branch. If the
default branch is updated, all divergence caches of the repo need to be
deleted.
Co-authored-by: Zettat123 <zettat123@gmail.com>
Backport #31548 by @brechtvl
Running git update-index for every individual file is slow, so add and
remove everything with a single git command.
When such a big commit lands in the default branch, it could cause PR
creation and patch checking for all open PRs to be slow, or time out
entirely. For example, a commit that removes 1383 files was measured to
take more than 60 seconds and timed out. With this change checking took
about a second.
This is related to #27967, though this will not help with commits that
change many lines in few files.
Co-authored-by: Brecht Van Lommel <brecht@blender.org>
Backport #31450 by @silverwind
See
https://docs.docker.com/reference/build-checks/legacy-key-value-format/.
Fixes these warnings seen during the docker build:
```
4 warnings found (use --debug to expand):
- LegacyKeyValueFormat: "ENV key=value" should be used instead of legacy "ENV key value" format (line 5)
- LegacyKeyValueFormat: "ENV key=value" should be used instead of legacy "ENV key value" format (line 9)
- LegacyKeyValueFormat: "ENV key=value" should be used instead of legacy "ENV key value" format (line 75)
- LegacyKeyValueFormat: "ENV key=value" should be used instead of legacy "ENV key value" format (line 76)
```
Introduced in: https://github.com/moby/buildkit/pull/4923
Co-authored-by: silverwind <me@silverwind.io>
Backport #31061 by @sergeyvfx
This change fixes cases when a Wiki page refers to a video stored in the
Wiki repository using relative path. It follows the similar case which
has been already implemented for images.
Test plan:
- Create repository and Wiki page
- Clone the Wiki repository
- Add video to it, say `video.mp4`
- Modify the markdown file to refer to the video using `<video
src="video.mp4">`
- Commit the Wiki page
- Observe that the video is properly displayed
Co-authored-by: Sergey Sharybin <sergey.vfx@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Backport #31410 by tobiasbp
This PR modifies the structs for editing and creating org teams to allow
team names to be up to 255 characters. The previous maximum length was
30 characters.
Co-authored-by: Tobias Balle-Petersen <tobias.petersen@unity3d.com>
Backport #31319 by @lunny
Fix a hash render problem like `<hash>: xxxxx` which is usually used in
release notes.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31333 by @lunny
Fix#31330Fix#31311
A workaround to fix the old database is to update object_format_name to
`sha1` if it's empty or null.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Backport #31325 by @bohde
When using the MinIO storage driver for Actions Artifacts, we found that
the chunked artifact required significantly more memory usage to both
upload and merge than the local storage driver. This seems to be related
to hardcoding a value of `-1` for the size to the MinIO client [which
has a warning about memory usage in the respective
docs](https://pkg.go.dev/github.com/minio/minio-go/v7#Client.PutObject).
Specifying the size in both the upload and merge case reduces memory
usage of the MinIO client.
Co-authored-by: Rowan Bohde <rowan.bohde@gmail.com>
Co-authored-by: Kyle D <kdumontnu@gmail.com>
