When a logged in user with no repositories visits their dashboard, it will
display a search box that lists their own repositories.
This is served by the `repo.SearchRepos` handler, which in turn calls
`commitstatus_service.FindReposLastestCommitStatuses()` with an empty
repo list.
That, in turn, will call `git_model.FindBranchesByRepoAndBranchName()`,
with an empty map. With no map, `FindBranchesByRepoAndBranchName()` ends
up querying the entire `branch` table, because no conditions were set
up.
Armed with a gazillion repo & commit shas, we return to
`FindReposLastestCommitStatuses`, and promptly call
`git_model.GetLatestCommitStatusForPairs`, which constructs a monstrous
query with so many placeholders that the database tells us to go
somewhere else, and flips us off. At least on instances the size of
Codeberg. On smaller instances, it will eventually return, and throw
away all the data, and return an empty set, having performed all this
for naught.
We fix this by short-circuiting `FindBranchesByRepoAndBranchName`, and
returning fast if our inputs are empty.
A test case is included.
Fixes #3521.
Signed-off-by: Gergely Nagy <forgejo@gergo.csillger.hu>
The target_url is necessary for the UI, but missed in
commit_status_summary table. This PR fix it.
---------
Co-authored-by: silverwind <me@silverwind.io>
Co-authored-by: delvh <dev.lh@web.de>
Co-authored-by: Giteabot <teabot@gitea.io>
This PR adds a new table named commit status summary to reduce queries
from the commit status table. After this change, commit status summary
table will be used for the final result, commit status table will be for
details.
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Many places have the following logic:
```go
func (jobs ActionJobList) GetRunIDs() []int64 {
ids := make(container.Set[int64], len(jobs))
for _, j := range jobs {
if j.RunID == 0 {
continue
}
ids.Add(j.RunID)
}
return ids.Values()
}
```
this introduces a `container.FilterMapUnique` function, which reduces
the code above to:
```go
func (jobs ActionJobList) GetRunIDs() []int64 {
return container.FilterMapUnique(jobs, func(j *ActionRunJob) (int64, bool) {
return j.RunID, j.RunID != 0
})
}
```
Conflicts:
models/issues/comment_list.go due to premature refactor in #3116
Fix https://github.com/go-gitea/gitea/issues/30428
---
Conflict resolution: trivial and move test to own subtest run directly
after `Normal`.
(cherrypicked commit 9466fec879f4f2c88c7c1e7a5cffba319282ab66)
- Remove options that currently aren't set
on `GarbageCollectLFSMetaObjectsOptions` and
`IterateLFSMetaObjectsForRepoOptions`.
- Simplify `IterateRepositoryIDsWithLFSMetaObjects` and
`IterateLFSMetaObjectsForRepo`.
- `IterateLFSMetaObjectsForRepo` was previously able to get in a
loop (`gc-lfs` doctor check was able to reproduce this) because the code
expected that the records would be updated to not match the SQL query,
but that wasn't the case. Simply enforce that only records higher than
the latest `id` from the previous iteration are allowed.
- For `gc-lfs` doctor check this was because `UpdatedLessRecentlyThan`
option was not set, which caused that records just marked as active in
the iteration weren't being filtered.
- Add unit tests
- Most likely a regression from 2cc3a6381c.
- The bug with `gc-lfs` was found on Codeberg.
This PR replaces the use of `max( id )`, and instead using ``max(
`index` )`` for determining the latest commit status. Building business
logic over an `auto_increment` primary key like `id` is risky and
there’re already plenty of discussions on the Internet.
There‘s no guarantee for `auto_increment` values to be monotonic,
especially upon failures or with a cluster. In the specific case, we met
the problem of commit statuses being outdated when using TiDB as the
database. As [being
documented](https://docs.pingcap.com/tidb/stable/auto-increment),
`auto_increment` values assigned to an `insert` statement will only be
monotonic on a per server (node) basis.
Closes #30074.
(cherry picked from commit 7443a10fc3d722d3326a0cb7b15b208f907c72d7)
- Currently protected branch rules do not apply to admins, however in
some cases (like in the case of Forgejo project) you might also want to
apply these rules to admins to avoid accidental merges.
- Add new option to configure this on a per-rule basis.
- Adds integration tests.
- Resolves #65
Fix #28157
This PR fix the possible bugs about actions schedule.
- Move `UpdateRepositoryUnit` and `SetRepoDefaultBranch` from models to
service layer
- Remove schedules plan from database and cancel waiting & running
schedules tasks in this repository when actions unit has been disabled
or global disabled.
- Remove schedules plan from database and cancel waiting & running
schedules tasks in this repository when default branch changed.
(cherry picked from commit 97292da960)
Conflicts:
modules/actions/github.go
routers/web/repo/setting/default_branch.go
routers/web/repo/setting/setting.go
services/repository/branch.go
services/repository/setting.go
tests/integration/actions_trigger_test.go
Unlike other async processing in the queue, we should sync branches to
the DB immediately when handling git hook calling. If it fails, users
can see the error message in the output of the git command.
It can avoid potential inconsistency issues, and help #29494.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Clarify when "string" should be used (and be escaped), and when
"template.HTML" should be used (no need to escape)
And help PRs like #29059 , to render the error messages correctly.
(cherry picked from commit f3eb835886031df7a562abc123c3f6011c81eca8)
Conflicts:
modules/web/middleware/binding.go
routers/web/feed/convert.go
tests/integration/branches_test.go
tests/integration/repo_branch_test.go
trivial context conflicts
With this change, the "You pushed on branch xyz" banner will be
displayed when either the viewed repository or its base repo (if the
current one's a fork) has pull requests enabled. Previously it only
displayed if the viewed repo had PRs enabled.
Furthermore, if the viewed repository is an original repository that the
viewing user has a fork of, if the forked repository has recently pushed
branches, then the banner will appear for the original repository too.
In this case, the notification will include branches from the viewing
user's fork, and branches they pushed to the base repo, too.
Refs: https://codeberg.org/forgejo/forgejo/pulls/2195
Signed-off-by: Gergely Nagy <forgejo@gergo.csillger.hu>
(cherry picked from commit a29f10661d59f6c33c5cfbee723f03f981aa6b72)
(cherry picked from commit 70c5e2021d7c385b9285622f0b2d878d3807d33c)
(cherry picked from commit 48b25be67a94f739601fbfea951ade6cdfe30691)
Fixes #27114.
* In Gitea 1.12 (#9532), a "dismiss stale approvals" branch protection
setting was introduced, for ignoring stale reviews when verifying the
approval count of a pull request.
* In Gitea 1.14 (#12674), the "dismiss review" feature was added.
* This caused confusion with users (#25858), as "dismiss" now means 2
different things.
* In Gitea 1.20 (#25882), the behavior of the "dismiss stale approvals"
branch protection was modified to actually dismiss the stale review.
For some users this new behavior of dismissing the stale reviews is not
desirable.
So this PR reintroduces the old behavior as a new "ignore stale
approvals" branch protection setting.
---------
Co-authored-by: delvh <dev.lh@web.de>
Fix #28157
This PR fix the possible bugs about actions schedule.
## The Changes
- Move `UpdateRepositoryUnit` and `SetRepoDefaultBranch` from models to
service layer
- Remove schedules plan from database and cancel waiting & running
schedules tasks in this repository when actions unit has been disabled
or global disabled.
- Remove schedules plan from database and cancel waiting & running
schedules tasks in this repository when default branch changed.
The 4 functions are duplicated, especially as interface methods. I think
we just need to keep `MustID` the only one and remove other 3.
```
MustID(b []byte) ObjectID
MustIDFromString(s string) ObjectID
NewID(b []byte) (ObjectID, error)
NewIDFromString(s string) (ObjectID, error)
```
Introduced the new interfrace method `ComputeHash` which will replace
the interface `HasherInterface`. Now we don't need to keep two
interfaces.
Reintroduced `git.NewIDFromString` and `git.MustIDFromString`. The new
function will detect the hash length to decide which objectformat of it.
If it's 40, then it's SHA1. If it's 64, then it's SHA256. This will be
right if the commitID is a full one. So the parameter should be always a
full commit id.
@AdamMajer Please review.
- Remove `ObjectFormatID`
- Remove function `ObjectFormatFromID`.
- Use `Sha1ObjectFormat` directly but not a pointer because it's an
empty struct.
- Store `ObjectFormatName` in `repository` struct
Refactor Hash interfaces and centralize hash function. This will allow
easier introduction of different hash function later on.
This forms the "no-op" part of the SHA256 enablement patch.
Fix #28056
This PR will check whether the repo has zero branch when pushing a
branch. If that, it means this repository hasn't been synced.
The reason caused that is after user upgrade from v1.20 -> v1.21, he
just push branches without visit the repository user interface. Because
all repositories routers will check whether a branches sync is necessary
but push has not such check.
For every repository, it has two states, synced or not synced. If there
is zero branch for a repository, then it will be assumed as non-sync
state. Otherwise, it's synced state. So if we think it's synced, we just
need to update branch/insert new branch. Otherwise do a full sync. So
that, for every push, there will be almost no extra load added. It's
high performance than yours.
For the implementation, we in fact will try to update the branch first,
if updated success with affect records > 0, then all are done. Because
that means the branch has been in the database. If no record is
affected, that means the branch does not exist in database. So there are
two possibilities. One is this is a new branch, then we just need to
insert the record. Another is the branches haven't been synced, then we
need to sync all the branches into database.
The function `GetByBean` has an obvious defect that when the fields are
empty values, it will be ignored. Then users will get a wrong result
which is possibly used to make a security problem.
To avoid the possibility, this PR removed function `GetByBean` and all
references.
And some new generic functions have been introduced to be used.
The recommand usage like below.
```go
// if query an object according id
obj, err := db.GetByID[Object](ctx, id)
// query with other conditions
obj, err := db.Get[Object](ctx, builder.Eq{"a": a, "b":b})
```
Part of #27065
This PR touches functions used in templates. As templates are not static
typed, errors are harder to find, but I hope I catch it all. I think
some tests from other persons do not hurt.
This PR removed `unittest.MainTest` the second parameter
`TestOptions.GiteaRoot`. Now it detects the root directory by current
working directory.
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Related to #26239
This PR makes some fixes:
- do not show the prompt for mirror repos and repos with pull request
units disabled
- use `commit_time` instead of `updated_unix`, as `commit_time` is the
real time when the branch was pushed
In the original implementation, we can only get the first 30 records of
the commit status (the default paging size), if the commit status is
more than 30, it will lead to the bug #25990. I made the following two
changes.
- On the page, use the ` db.ListOptions{ListAll: true}` parameter
instead of `db.ListOptions{}`
- The `GetLatestCommitStatus` function makes a determination as to
whether or not a pager is being used.
fixed #25990
This PR
- Fix #26093. Replace `time.Time` with `timeutil.TimeStamp`
- Fix #26135. Add missing `xorm:"extends"` to `CountLFSMetaObject` for
LFS meta object query
- Add a unit test for LFS meta object garbage collection
Before:
After:
There's a bug in the recent logic, `CalcCommitStatus` will always return
the first item of `statuses` or error status, because `state` is defined
with default value which should be `CommitStatusSuccess`
Then
``` golang
if status.State.NoBetterThan(state) {
```
this `if` will always return false unless `status.State =
CommitStatusError` which makes no sense.
So `lastStatus` will always be `nil` or error status.
Then we will always return the first item of `statuses` here or only
return error status, and this is why in the first picture the commit
status is `Success` but not `Failure`.
af1ffbcd63/models/git/commit_status.go (L204-L211)
Co-authored-by: Giteabot <teabot@gitea.io>
Fix #25776. Close #25826.
In the discussion of #25776, @wolfogre's suggestion was to remove the
commit status of `running` and `warning` to keep it consistent with
github.
references:
-
https://docs.github.com/en/rest/commits/statuses?apiVersion=2022-11-28#about-commit-statuses
## ⚠️ BREAKING ⚠️
So the commit status of Gitea will be consistent with GitHub, only
`pending`, `success`, `error` and `failure`, while `warning` and
`running` are not supported anymore.
---------
Co-authored-by: Jason Song <i@wolfogre.com>
