To avoid deadlock problem, almost database related functions should be
have ctx as the first parameter.
This PR do a refactor for some of these functions.
Related #14180
Related #25233
Related #22639Close#19786
Related #12763
This PR will change all the branches retrieve method from reading git
data to read database to reduce git read operations.
- [x] Sync git branches information into database when push git data
- [x] Create a new table `Branch`, merge some columns of `DeletedBranch`
into `Branch` table and drop the table `DeletedBranch`.
- [x] Read `Branch` table when visit `code` -> `branch` page
- [x] Read `Branch` table when list branch names in `code` page dropdown
- [x] Read `Branch` table when list git ref compare page
- [x] Provide a button in admin page to manually sync all branches.
- [x] Sync branches if repository is not empty but database branches are
empty when visiting pages with branches list
- [x] Use `commit_time desc` as the default FindBranch order by to keep
consistent as before and deleted branches will be always at the end.
---------
Co-authored-by: Jason Song <i@wolfogre.com>
## ⚠️ Breaking
The `log.<mode>.<logger>` style config has been dropped. If you used it,
please check the new config manual & app.example.ini to make your
instance output logs as expected.
Although many legacy options still work, it's encouraged to upgrade to
the new options.
The SMTP logger is deleted because SMTP is not suitable to collect logs.
If you have manually configured Gitea log options, please confirm the
logger system works as expected after upgrading.
## Description
Close#12082 and maybe more log-related issues, resolve some related
FIXMEs in old code (which seems unfixable before)
Just like rewriting queue #24505 : make code maintainable, clear legacy
bugs, and add the ability to support more writers (eg: JSON, structured
log)
There is a new document (with examples): `logging-config.en-us.md`
This PR is safer than the queue rewriting, because it's just for
logging, it won't break other logic.
## The old problems
The logging system is quite old and difficult to maintain:
* Unclear concepts: Logger, NamedLogger, MultiChannelledLogger,
SubLogger, EventLogger, WriterLogger etc
* Some code is diffuclt to konw whether it is right:
`log.DelNamedLogger("console")` vs `log.DelNamedLogger(log.DEFAULT)` vs
`log.DelLogger("console")`
* The old system heavily depends on ini config system, it's difficult to
create new logger for different purpose, and it's very fragile.
* The "color" trick is difficult to use and read, many colors are
unnecessary, and in the future structured log could help
* It's difficult to add other log formats, eg: JSON format
* The log outputer doesn't have full control of its goroutine, it's
difficult to make outputer have advanced behaviors
* The logs could be lost in some cases: eg: no Fatal error when using
CLI.
* Config options are passed by JSON, which is quite fragile.
* INI package makes the KEY in `[log]` section visible in `[log.sub1]`
and `[log.sub1.subA]`, this behavior is quite fragile and would cause
more unclear problems, and there is no strong requirement to support
`log.<mode>.<logger>` syntax.
## The new design
See `logger.go` for documents.
## Screenshot
<details>
</details>
## TODO
* [x] add some new tests
* [x] fix some tests
* [x] test some sub-commands (manually ....)
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: delvh <dev.lh@web.de>
Co-authored-by: Giteabot <teabot@gitea.io>
This PR
- [x] Move some functions from `issues.go` to `issue_stats.go` and
`issue_label.go`
- [x] Remove duplicated issue options `UserIssueStatsOption` to keep
only one `IssuesOptions`
# ⚠️ Breaking
Many deprecated queue config options are removed (actually, they should
have been removed in 1.18/1.19).
If you see the fatal message when starting Gitea: "Please update your
app.ini to remove deprecated config options", please follow the error
messages to remove these options from your app.ini.
Example:
```
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].ISSUE_INDEXER_QUEUE_TYPE`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].UPDATE_BUFFER_LEN`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [F] Please update your app.ini to remove deprecated config options
```
Many options in `[queue]` are are dropped, including:
`WRAP_IF_NECESSARY`, `MAX_ATTEMPTS`, `TIMEOUT`, `WORKERS`,
`BLOCK_TIMEOUT`, `BOOST_TIMEOUT`, `BOOST_WORKERS`, they can be removed
from app.ini.
# The problem
The old queue package has some legacy problems:
* complexity: I doubt few people could tell how it works.
* maintainability: Too many channels and mutex/cond are mixed together,
too many different structs/interfaces depends each other.
* stability: due to the complexity & maintainability, sometimes there
are strange bugs and difficult to debug, and some code doesn't have test
(indeed some code is difficult to test because a lot of things are mixed
together).
* general applicability: although it is called "queue", its behavior is
not a well-known queue.
* scalability: it doesn't seem easy to make it work with a cluster
without breaking its behaviors.
It came from some very old code to "avoid breaking", however, its
technical debt is too heavy now. It's a good time to introduce a better
"queue" package.
# The new queue package
It keeps using old config and concept as much as possible.
* It only contains two major kinds of concepts:
* The "base queue": channel, levelqueue, redis
* They have the same abstraction, the same interface, and they are
tested by the same testing code.
* The "WokerPoolQueue", it uses the "base queue" to provide "worker
pool" function, calls the "handler" to process the data in the base
queue.
* The new code doesn't do "PushBack"
* Think about a queue with many workers, the "PushBack" can't guarantee
the order for re-queued unhandled items, so in new code it just does
"normal push"
* The new code doesn't do "pause/resume"
* The "pause/resume" was designed to handle some handler's failure: eg:
document indexer (elasticsearch) is down
* If a queue is paused for long time, either the producers blocks or the
new items are dropped.
* The new code doesn't do such "pause/resume" trick, it's not a common
queue's behavior and it doesn't help much.
* If there are unhandled items, the "push" function just blocks for a
few seconds and then re-queue them and retry.
* The new code doesn't do "worker booster"
* Gitea's queue's handlers are light functions, the cost is only the
go-routine, so it doesn't make sense to "boost" them.
* The new code only use "max worker number" to limit the concurrent
workers.
* The new "Push" never blocks forever
* Instead of creating more and more blocking goroutines, return an error
is more friendly to the server and to the end user.
There are more details in code comments: eg: the "Flush" problem, the
strange "code.index" hanging problem, the "immediate" queue problem.
Almost ready for review.
TODO:
* [x] add some necessary comments during review
* [x] add some more tests if necessary
* [x] update documents and config options
* [x] test max worker / active worker
* [x] re-run the CI tasks to see whether any test is flaky
* [x] improve the `handleOldLengthConfiguration` to provide more
friendly messages
* [x] fine tune default config values (eg: length?)
## Code coverage:
Before, there was a `log/buffer.go`, but that design is not general, and
it introduces a lot of irrelevant `Content() (string, error) ` and
`return "", fmt.Errorf("not supported")` .
And the old `log/buffer.go` is difficult to use, developers have to
write a lot of `Contains` and `Sleep` code.
The new `LogChecker` is designed to be a general approach to help to
assert some messages appearing or not appearing in logs.
Since #23493 has conflicts with latest commits, this PR is my proposal
for fixing #23371
Details are in the comments
And refactor the `modules/options` module, to make it always use
"filepath" to access local files.
Benefits:
* No need to do `util.CleanPath(strings.ReplaceAll(p, "\\", "/"))),
"/")` any more (not only one before)
* The function behaviors are clearly defined
Extract from #11669 and enhancement to #22585 to support exclusive
scoped labels in label templates
* Move label template functionality to label module
* Fix handling of color codes
* Add Advanced label template
Add a new "exclusive" option per label. This makes it so that when the
label is named `scope/name`, no other label with the same `scope/`
prefix can be set on an issue.
The scope is determined by the last occurence of `/`, so for example
`scope/alpha/name` and `scope/beta/name` are considered to be in
different scopes and can coexist.
Exclusive scopes are not enforced by any database rules, however they
are enforced when editing labels at the models level, automatically
removing any existing labels in the same scope when either attaching a
new label or replacing all labels.
In menus use a circle instead of checkbox to indicate they function as
radio buttons per scope. Issue filtering by label ensures that only a
single scoped label is selected at a time. Clicking with alt key can be
used to remove a scoped label, both when editing individual issues and
batch editing.
Label rendering refactor for consistency and code simplification:
* Labels now consistently have the same shape, emojis and tooltips
everywhere. This includes the label list and label assignment menus.
* In label list, show description below label same as label menus.
* Don't use exactly black/white text colors to look a bit nicer.
* Simplify text color computation. There is no point computing luminance
in linear color space, as this is a perceptual problem and sRGB is
closer to perceptually linear.
* Increase height of label assignment menus to show more labels. Showing
only 3-4 labels at a time leads to a lot of scrolling.
* Render all labels with a new RenderLabel template helper function.
Label creation and editing in multiline modal menu:
* Change label creation to open a modal menu like label editing.
* Change menu layout to place name, description and colors on separate
lines.
* Don't color cancel button red in label editing modal menu.
* Align text to the left in model menu for better readability and
consistent with settings layout elsewhere.
Custom exclusive scoped label rendering:
* Display scoped label prefix and suffix with slightly darker and
lighter background color respectively, and a slanted edge between them
similar to the `/` symbol.
* In menus exclusive labels are grouped with a divider line.
---------
Co-authored-by: Yarden Shoham <hrsi88@gmail.com>
Co-authored-by: Lauris BH <lauris@nix.lv>
Importing labels (via `gitea restore-repo`) did not split them up into
batches properly. The first "batch" would create all labels, the second
"batch" would create all labels except those in the first "batch", etc.
This meant that when importing more than 99 labels (the batch size)
there would always be duplicate ones.
This is solved by actually passing `labels[:lbBatchSize]` to the
`CreateLabels()` function, instead of the entire list `labels`.
This is a fix for https://github.com/go-gitea/gitea/pull/22510
The code assumed that the `AssigneeID` from the comment YAML was an
`int64`, but it is actually an `int`, causing a panic. It also had no
check on whether the type cast was actually valid, so badly formatted
YAML could also cause a panic.
Both these issues have been fixed.
This commit adds support for specifying comment types when importing
with `gitea restore-repo`. It makes it possible to import issue changes,
such as "title changed" or "assigned user changed".
An earlier version of this pull request was made by Matti Ranta, in
https://future.projects.blender.org/blender-migration/gitea-bf/pulls/3
There are two changes with regard to Matti's original code:
1. The comment type was an `int64` in Matti's code, and is now using a
string. This makes it possible to use `comment_type: title`, which is
more reliable and future-proof than an index into an internal list in
the Gitea Go code.
2. Matti's code also had support for including labels, but in a way that
would require knowing the database ID of the labels before the import
even starts, which is impossible. This can be solved by using label
names instead of IDs; for simplicity I I left that out of this PR.
Reposting pull request for #22465
> Migration from GitBucket does not work due to a access for "Reviews"
API on GitBucket that makes 404 response. This PR has following changes.
> 1. Made to stop access for Reviews API while migrating from GitBucket.
> 2. Added support for custom URL (e.g.
`http://example.com/gitbucket/owner/repository`)
> 3. Made to accept for git checkout URL
(`http://example.com/git/owner/repository.git`)
Co-authored-by: zeripath <art27@cantab.net>
Some dbs require that all tables have primary keys, see
- #16802
- #21086
We can add a test to keep it from being broken again.
Edit:
~Added missing primary key for `ForeignReference`~ Dropped the
`ForeignReference` table to satisfy the check, so it closes#21086.
More context can be found in comments.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
Change all license headers to comply with REUSE specification.
Fix#16132
Co-authored-by: flynnnnnnnnnn <flynnnnnnnnnn@github>
Co-authored-by: John Olheiser <john.olheiser@gmail.com>
This PR adds a context parameter to a bunch of methods. Some helper
`xxxCtx()` methods got replaced with the normal name now.
Co-authored-by: delvh <dev.lh@web.de>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
GitHub migration tests will be skipped if the secret for the GitHub API
token hasn't been set.
This change should make all tests pass (or skip in the case of this one)
for anyone running the pipeline on their own infrastructure without
further action on their part.
Resolves https://github.com/go-gitea/gitea/issues/21739
Signed-off-by: Gary Moon <gary@garymoon.net>
GitHub allows releases with target commitish `refs/heads/BRANCH`, which
then causes issues in Gitea after migration. This fix handles cases that
a branch already has a prefix.
Fixes#20317
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
The `go-licenses` make task introduced in #21034 is being run on make vendor
and occasionally causes an empty go-licenses file if the vendors need to
change. This should be moved to the generate task as it is a generated file.
Now because of this change we also need to split generation into two separate
steps:
1. `generate-backend`
2. `generate-frontend`
In the future it would probably be useful to make `generate-swagger` part of `generate-frontend` but it's not tolerated with our .drone.yml
Ref #21034
Signed-off-by: Andrew Thornton <art27@cantab.net>
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: delvh <dev.lh@web.de>
When migrating add several more important sanity checks:
* SHAs must be SHAs
* Refs must be valid Refs
* URLs must be reasonable
Signed-off-by: Andrew Thornton <art27@cantab.net>
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <matti@mdranta.net>
Some Migration Downloaders provide re-writing of CloneURLs that may point to
unallowed urls. Recheck after the CloneURL is rewritten.
Signed-off-by: Andrew Thornton <art27@cantab.net>
When migrating git repositories we should ensure that the commit-graph is written.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: delvh <dev.lh@web.de>
* Fix cli command restore-repo: "units" should be parsed as StringSlice because after #15790 it's read by c.StringSlice("units"). Before, the "units" were processed by strings.Split
* Add checking for invalid unit names
Co-authored-by: 6543 <6543@obermui.de>
* Move access and repo permission to models/perm/access
* fix test
* fix git test
* Move functions sequence
* Some improvements per @KN4CK3R and @delvh
* Move issues related code to models/issues
* Move some issues related sub package
* Merge
* Fix test
* Fix test
* Fix test
* Fix test
* Rename some files
A pr.Reviewer may be nil when migrating from Gitea if this is a team
request review.
We do not migrate teams therefore we cannot map these requests, but we can
migrate user requests.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Update go tool dependencies
Updated all tool dependencies to latest tags, hoping CI will like it.
* fix new lint errors
* handle more strings.Title cases
* remove lint skip
Follows #19266, #8553, Close#18553, now there are only three `Run..(&RunOpts{})` functions.
* before: `stdout, err := RunInDir(path)`
* now: `stdout, _, err := RunStdString(&git.RunOpts{Dir:path})`
* Remove `db.DefaultContext` usage in routers, use `ctx` directly
* Use `ctx` directly if there is one, remove some `db.DefaultContext` in `services`
* Use ctx instead of db.DefaultContext for `cmd` and some `modules` packages
* fix incorrect context usage
Make SKIP_TLS_VERIFY apply to git data migrations too through adding the `-c http.sslVerify=false` option to the git clone command.
Fix#18998
Signed-off-by: Andrew Thornton <art27@cantab.net>
Storing the foreign identifier of an imported issue in the database is a prerequisite to implement idempotent migrations or mirror for issues. It is a baby step towards mirroring that introduces a new table.
At the moment when an issue is created by the Gitea uploader, it fails if the issue already exists. The Gitea uploader could be modified so that, instead of failing, it looks up the database to find an existing issue. And if it does it would update the issue instead of creating a new one. However this is not currently possible because an information is missing from the database: the foreign identifier that uniquely represents the issue being migrated is not persisted. With this change, the foreign identifier is stored in the database and the Gitea uploader will then be able to run a query to figure out if a given issue being imported already exists.
The implementation of mirroring for issues, pull requests, releases, etc. can be done in three steps:
1. Store an identifier for the element being mirrored (issue, pull request...) in the database (this is the purpose of these changes)
2. Modify the Gitea uploader to be able to update an existing repository with all it contains (issues, pull request...) instead of failing if it exists
3. Optimize the Gitea uploader to speed up the updates, when possible.
The second step creates code that does not yet exist to enable idempotent migrations with the Gitea uploader. When a migration is done for the first time, the behavior is not changed. But when a migration is done for a repository that already exists, this new code is used to update it.
The third step can use the code created in the second step to optimize and speed up migrations. For instance, when a migration is resumed, an issue that has an update time that is not more recent can be skipped and only newly created issues or updated ones will be updated. Another example of optimization could be that a webhook notifies Gitea when an issue is updated. The code triggered by the webhook would download only this issue and call the code created in the second step to update the issue, as if it was in the process of an idempotent migration.
The ForeignReferences table is added to contain local and foreign ID pairs relative to a given repository. It can later be used for pull requests and other artifacts that can be mirrored. Although the foreign id could be added as a single field in issues or pull requests, it would need to be added to all tables that represent something that can be mirrored. Creating a new table makes for a simpler and more generic design. The drawback is that it requires an extra lookup to obtain the information. However, this extra information is only required during migration or mirroring and does not impact the way Gitea currently works.
The foreign identifier of an issue or pull request is similar to the identifier of an external user, which is stored in reactions, issues, etc. as OriginalPosterID and so on. The representation of a user is however different and the ability of users to link their account to an external user at a later time is also a logic that is different from what is involved in mirroring or migrations. For these reasons, despite some commonalities, it is unclear at this time how the two tables (foreign reference and external user) could be merged together.
The ForeignID field is extracted from the issue migration context so that it can be dumped in files with dump-repo and later restored via restore-repo.
The GetAllComments downloader method is introduced to simplify the implementation and not overload the Context for the purpose of pagination. It also clarifies in which context the comments are paginated and in which context they are not.
The Context interface is no longer useful for the purpose of retrieving the LocalID and ForeignID since they are now both available from the PullRequest and Issue struct. The Reviewable and Commentable interfaces replace and serve the same purpose.
The Context data member of PullRequest and Issue becomes a DownloaderContext to clarify that its purpose is not to support in memory operations while the current downloader is acting but is not otherwise persisted. It is, for instance, used by the GitLab downloader to store the IsMergeRequest boolean and sort out issues.
---
[source](https://lab.forgefriends.org/forgefriends/forgefriends/-/merge_requests/36)
Signed-off-by: Loïc Dachary <loic@dachary.org>
Co-authored-by: Loïc Dachary <loic@dachary.org>
* logs: add the buffer logger to inspect logs during testing
Signed-off-by: Loïc Dachary <loic@dachary.org>
* migrations: add test for importing pull requests in gitea uploader
Signed-off-by: Loïc Dachary <loic@dachary.org>
* for each git.OpenRepositoryCtx, call Close
* Content is expected to return the content of the log
* test for errors before defer
Co-authored-by: Loïc Dachary <loic@dachary.org>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
When migrating, g.issues is a map with all issues created during the
migration. If an issue is not found in g.issues when inserting a
comment or a review, it cannot exist in the database and trying to get
it via GetIssueByIndex() will always fail and return an error.
Signed-off-by: singuliere <singuliere@autistici.org>
When calling DumpRepository and RestoreRepository on the same Gitea
instance, the users are preserved: all labels, issues etc. belong to
the external user who is, in this particular case, the local user.
Dead code verifying g.gitServiceType.Name() == "" (i.e. plain git) is
removed. The function is never called because the plain git downloader
does not migrate anything that is associated to a user, by definition.
Errors returned by GetUserIDByExternalUserID are no longer ignored.
The userMap is used when the external user is not kown, which is the
most common case. It was only used when the external user exists
which happens less often and, as a result, every occurence of an
unknown external user required a SQL query.
Signed-off-by: Loïc Dachary <loic@dachary.org>
Co-authored-by: Loïc Dachary <loic@dachary.org>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* Replace `sync.Map` with normal maps
- These maps aren't being used in any kind of concurrent read/write and
thus don't need `sync.Map` and can instead use normal maps.
- Special thanks to dachary.
- Added in: https://github.com/go-gitea/gitea/pull/6290
* Remove unannounced feature
* add test coverage for original author conversion during migrations
And create a function to factorize a code snippet that is repeated
five times and would otherwise be more difficult to test and maintain
consistently.
Signed-off-by: Loïc Dachary <loic@dachary.org>
* fix variable scope and int64 formatting
* add missing calls to remapExternalUser and fix misplaced %d
Co-authored-by: Loïc Dachary <loic@dachary.org>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
* GitLab reviews may not have the updated_at field set
Fallback to created_at if that the case and to time.Now() if it is
also missing.
Fixes: 18434
* use assert.WithinDuration
Co-authored-by: Loïc Dachary <loic@dachary.org>
This PR continues the work in #17125 by progressively ensuring that git
commands run within the request context.
This now means that the if there is a git repo already open in the context it will be used instead of reopening it.
Signed-off-by: Andrew Thornton <art27@cantab.net>
The CheckRepoStats function missed the following counters:
- label num_closed_issues & num_closed_pulls
- milestone num_closed_issues & num_closed_pulls
The update SQL statements for updating the repository
num_closed_issues & num_closed_pulls fields were repeated in three
functions (repo.CheckRepoStats, migrate.insertIssues and
models.Issue.updateClosedNum) and were moved to a single helper.
The UpdateRepoStats is implemented and called in the Finish migration method so that it happens immediately instead of wating for the
CheckRepoStats to run.
Signed-off-by: Loïc Dachary loic@dachary.org
---
[source](https://lab.forgefriends.org/forgefriends/forgefriends/-/merge_requests/34)
Git will and can pack references into packfiles and therefore if you write/read the
files directly you will get false results. Instead you should use update-ref and
show-ref. To that end I have created three new functions in git/repo_commit.go that
will do this correctly.
Related #17191
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Some refactors related repository model
* Move more methods out of repository
* Move repository into models/repo
* Fix test
* Fix test
* some improvements
* Remove unnecessary function
Use hostmacher to replace matchlist.
And we introduce a better DialContext to do a full host/IP check, otherwise the attackers can still bypass the allow/block list by a 302 redirection.
