Issue: After switching the default branch, other branches are still
compared against the old default branch due to outdated divergence
cache.
Change: Clear the divergence cache in SetRepoDefaultBranch to ensure
correct comparisons against the new default branch.
Fixes#34369
Files that should be stored in LFS and are uploaded/edited from the API
or web UI aren't stored in LFS. This may be a regression from #34154.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
In the Gitea GUI, the user can see the time that _AccessTokens_ and
_PublicKeys_ were last used. This information is not returned by the
_/users/{username}/tokens_ and _/user/keys_ endpoints in the API. This
PR adds the missing data.
The time of last usage for for _tokens_ & _keys_ seem to be stored in
the _Updated_ field of the structs internally. For consistency, I have
used the name _updated_at_ for the new field returned by the _API_.
However, for the _API_ user, I don't think that name reflects the data
returned, as I believe it is the time of last usage. I propose that we
use the name _last_used_at_ instead. Let's hear reviewers opinion on
that.
* PublicKey
1. _last_used_at_: string($date-time)
* AccessToken
1. _created_at_: string($date-time) (for parity with public keys)
2. _last_used_at_: string($date-time)
Fix#34313
If archive downloads are are disabled using
_DISABLE_DOWNLOAD_SOURCE_ARCHIVES_, archive links are still returned by
the API.
This PR changes the data returned, so the fields _zipball_url_ and
_tarball_url_ are omitted if archive downloads have been disabled.
Resolve#32159
Fix#34224
The previous implementation in #33744 will get the pushed commits
changed files. But it's not always right when push a merged commit. This
PR reverted the logic in #33744 and will always get the PR's changed
files and get code owners.
Fix#880
Design:
1. A global setting `security.TWO_FACTOR_AUTH`.
* To support org-level config, we need to introduce a better "owner
setting" system first (in the future)
2. A user without 2FA can login and may explore, but can NOT read or
write to any repositories via API/web.
3. Keep things as simple as possible.
* This option only aggressively suggest users to enable their 2FA at the
moment, it does NOT guarantee that users must have 2FA before all other
operations, it should be good enough for real world use cases.
* Some details and tests could be improved in the future since this
change only adds a check and seems won't affect too much.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
`[repository.pull-request] DELAY_CHECK_FOR_INACTIVE_DAYS` is a new
setting to delay the mergeable check for pull requests that have been
inactive for the specified number of days.
This avoids potentially long delays for big repositories with many pull
requests. and reduces system load overall when there are many
repositories or pull requests.
When viewing the PR, checking will start immediately and the PR merge
box will automatically reload when complete. Accessing the PR through
the API will also start checking immediately.
The default value of `7` provides a balance between system load, and
keeping behavior similar to what it was before both for users and API
access. With `0` all conflict checking will be delayed, while `-1`
always checks immediately to restore the previous behavior.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Adds an API POST endpoint under `/repos/{owner}/{repo}/file-contents`
which receives a list of paths and returns a list of the contents of
these files.
This API endpoint will be helpful for applications like headless CMS
(reference: https://github.com/sveltia/sveltia-cms/issues/198) which
need to retrieve a large number of files by reducing the amount of
needed API calls.
Close#33495
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
This pull request adds a GitHub-compatible API endpoint to lock and
unlock an issue.
The following routes exist now:
- `PUT /api/v1/repos/{owner}/{repo}/issues/{id}/lock` to lock an issue
- `DELETE /api/v1/repos/{owner}/{repo}/issues/{id}/lock` to unlock an issue
Fixes#33677Fixes#20012
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Implements runner apis based on
https://docs.github.com/en/rest/actions/self-hosted-runners?apiVersion=2022-11-28#list-self-hosted-runners-for-an-organization
- Add Post endpoints for registration-token, google/go-github revealed
this as problem
- We should deprecate Get Endpoints, leaving them for compatibility
- Get endpoint of admin has api path /admin/runners/registration-token
that feels wrong, /admin/actions/runners/registration-token seems more
consistent with user/org/repo api
- Get Runner Api
- List Runner Api
- Delete Runner Api
- Tests admin / user / org / repo level endpoints
Related to #33750 (implements point 1 and 2)
Via needs discovered in #32461, this runner api is needed to allow
cleanup of runners that are deallocated without user interaction.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
This PR moved git attributes related code to `modules/git/attribute` sub
package and moved language stats related code to
`modules/git/languagestats` sub package to make it easier to maintain.
And it also introduced a performance improvement which use the `git
check-attr --source` which can be run in a bare git repository so that
we don't need to create a git index file. The new parameter need a git
version >= 2.40 . If git version less than 2.40, it will fall back to
previous implementation.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: yp05327 <576951401@qq.com>
Fix#2616
This PR adds a new sort option for exclusive labels.
For exclusive labels, a new property is exposed called "order", while in
the UI options are populated automatically in the `Sort` column (see
screenshot below) for each exclusive label scope.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
When list commits, some of the commits authors are the same at many
situations. But current logic will always fetch the same GPG keys from
database. This PR will cache the GPG keys, emails and users for the
context so that reducing the database queries.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
This PR uniform all temporary directory usage so that it will be easier
to manage.
Relate to #31792
- [x] Added a new setting to allow users to configure the global
temporary directory.
- [x] Move all temporary files and directories to be placed under
os.Temp()/gitea.
- [x] `setting.Repository.Local.LocalCopyPath` now will be
`setting.TempPath/local-repo` and the customized path is removed.
```diff
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-;[repository.local]
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-;;
-;; Path for local repository copy. Defaults to TEMP_PATH + `local-repo`, this is deprecated and cannot be changed
-;LOCAL_COPY_PATH = local-repo
```
- [x] `setting.Repository.Upload.TempPath` now will be
`settting.TempPath/uploads` and the customized path is removed.
```diff
;[repository.upload]
-;;
-;; Path for uploads. Defaults to TEMP_PATH + `uploads`
-;TEMP_PATH = uploads
```
- [x] `setting.Packages.ChunkedUploadPath` now will be
`settting.TempPath/package-upload` and the customized path is removed.
```diff
;[packages]
-;;
-;; Path for chunked uploads. Defaults it's `package-upload` under `TEMP_PATH` unless it's an absolute path.
-;CHUNKED_UPLOAD_PATH = package-upload
```
- [x] `setting.SSH.KeyTestPath` now will be
`settting.TempPath/ssh_key_test` and the customized path is removed.
```diff
[server]
-;;
-;; Directory to create temporary files in when testing public keys using ssh-keygen,
-;; default is the system temporary directory.
-;SSH_KEY_TEST_PATH =
```
TODO:
- [ ] setting.PprofDataPath haven't been changed because it may need to
be kept until somebody read it but temp path may be clean up any time.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
This PR tries to finally fix the bug mentioned in #30011 and #15504,
where the user repo limit is checked when creating a repo in an
organization.
Fix#30011
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: TheFox0x7 <thefox0x7@gmail.com>
Fix#28144
To make the resources will be cleanup once failed. All repository
operations now follow a consistent pattern:
- 1. Create a database record for the repository with the status
being_migrated.
- 2. Register a deferred cleanup function to delete the repository and
its related data if the operation fails.
- 3. Perform the actual Git and database operations step by step.
- 4. Upon successful completion, update the repository’s status to
ready.
The adopt operation is a special case — if it fails, the repository on
disk should not be deleted.
Extract from #33934
In the same goroutine, we should reuse the exist cat file sub process
which exist in `git.Repository` to avoid creating a unnecessary
temporary subprocess.
This PR reuse the exist cate file writer and reader in
`getCommitFromBatchReader`.
It also move `prepareLatestCommitInfo` before creating dataRc which will
hold the writer so other git operation will create a temporary cat file
subprocess.
Fixes [#34027](https://github.com/go-gitea/gitea/issues/34027)
Discord does not allow for description bigger than 2048 bytes. If the
description is bigger than that it will throw 400 and the event won't
appear in discord. To fix that, in the createPayload method we now slice
the description to ensure it doesn’t exceed the limit.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Fix#32886
Add `last_committer_date` and `last_author_date` in the content API
which is not implemented by Github API v3 at the moment.
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
The pull request list API is slow, for every pull request, it needs to
open a git repository. Assume it has 30 records, there will be 30 sub
processes back because every repository will open a git cat-file --batch
sub process. This PR use base git repository to get the head commit id
rather than read it from head repository to avoid open any head git
repository.
Fix#33966
```
;; User must sign in to view anything.
;; It could be set to "expensive" to block anonymous users accessing some pages which consume a lot of resources,
;; for example: block anonymous AI crawlers from accessing repo code pages.
;; The "expensive" mode is experimental and subject to change.
;REQUIRE_SIGNIN_VIEW = false
```
The doctor storage check reconstructs the lfs oid by producing a string
where the path separator is stripped
ab/dc/efg -> abdcefg. Windows however uses a backslash and thus the
ReplaceAll call doesn't produce the correct oid resulting in all lfs
objects being classed as orphaned.
This PR allows this to be more OS agnostic.
Closes#34039
---------
Co-authored-by: Giteabot <teabot@gitea.io>
Follow #33127Fix#8649, fix#639
This is a complete solution. A repo unit could be set to:
* Anonymous read (non-signed-in user)
* Everyone read (signed-in user)
* Everyone write (wiki-only)
Follow #33127
This PR add backend logic and test for "anonymous access", it shares the
same logic as "everyone access", so not too much change.
By the way, split `SettingsPost` into small functions to make it easier
to make frontend-related changes in the future.
Next PR will add frontend support for "anonymous access"
* reuse recoverable error checks across mirror_pull
* add new cases for 'cannot lock ref/not our ref' (race condition in
fetch) and 'Unable to create/lock"
* move lfs sync right after commit graph write, and before other
maintenance which may fail
* try a prune for 'broken reference' as well as 'not our ref'
* always sync LFS right after commit graph write, and before other
maintenance which may fail
This handles a few cases where our very large and very active
repositories could serve mirrored git refs, but be missing lfs files:
## Case 1 (multiple variants): Race condition in git fetch
There was already a check for 'unable to resolve reference' on a failed
git fetch, after which a git prune and then subsequent fetch are
performed. This is to work around a race condition where the git remote
tells Gitea about a ref for some HEAD of a branch, then fails a few
seconds later because the remote branch was deleted, or the ref was
updated (force push).
There are two more variants to the error message you can get, but for
the same kind of race condition. These *may* be related to the git
binary version Gitea has access to (in my case, it was 2.48.1).
## Case 2: githttp.go can serve updated git refs before it's synced lfs
oids
There is probably a more aggressive refactor we could do here to have
the cat-file loop use FETCH_HEAD instead of relying on the commit graphs
to be committed locally (and thus serveable to clients of Gitea), but a
simple reduction in the occurrences of this for me was to move the lfs
sync block immediately after the commit-graph write and before any other
time-consuming (or potentially erroring/exiting) blocks.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
1. Ignore empty inputs in `UnmarshalHandleDoubleEncode`
2. Ignore non-existing `stateEvent.User` in gitlab migration
3. Enable `release` and `wiki` units when they are selected in migration
4. Sanitize repo name for migration and new repo
The pagination on the user dashboard sounds unnecessary, this will
change it to a prev/next buttons. For instances with around `10 million`
records in the action table, this option affects how the user dashboard
is loaded on first visit.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: Giteabot <teabot@gitea.io>
