2022-05-07 19:05:52 +02:00
// Copyright 2021 Gitea. All rights reserved.
2022-11-27 13:20:29 -05:00
// SPDX-License-Identifier: MIT
2022-05-07 19:05:52 +02:00
package automerge
import (
"context"
"errors"
"fmt"
2025-03-27 19:40:14 +00:00
"forgejo.org/models/db"
issues_model "forgejo.org/models/issues"
access_model "forgejo.org/models/perm/access"
pull_model "forgejo.org/models/pull"
repo_model "forgejo.org/models/repo"
user_model "forgejo.org/models/user"
"forgejo.org/modules/git"
"forgejo.org/modules/gitrepo"
"forgejo.org/modules/graceful"
"forgejo.org/modules/log"
"forgejo.org/modules/process"
"forgejo.org/modules/queue"
notify_service "forgejo.org/services/notify"
pull_service "forgejo.org/services/pull"
repo_service "forgejo.org/services/repository"
shared_automerge "forgejo.org/services/shared/automerge"
2022-05-07 19:05:52 +02:00
)
// Init runs the task queue to that handles auto merges
func Init ( ) error {
2024-05-21 23:23:22 +08:00
notify_service . RegisterNotifier ( NewNotifier ( ) )
2024-10-23 21:31:52 +02:00
shared_automerge . PRAutoMergeQueue = queue . CreateUniqueQueue ( graceful . GetManager ( ) . ShutdownContext ( ) , "pr_auto_merge" , handler )
if shared_automerge . PRAutoMergeQueue == nil {
2025-05-29 17:34:29 +02:00
return errors . New ( "unable to create pr_auto_merge queue" )
2022-05-07 19:05:52 +02:00
}
2024-10-23 21:31:52 +02:00
go graceful . GetManager ( ) . RunWithCancel ( shared_automerge . PRAutoMergeQueue )
2022-05-07 19:05:52 +02:00
return nil
}
// handle passed PR IDs and test the PRs
Rewrite queue (#24505)
# ⚠️ 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:

2023-05-08 19:49:59 +08:00
func handler ( items ... string ) [ ] string {
for _ , s := range items {
2022-05-07 19:05:52 +02:00
var id int64
var sha string
Rewrite queue (#24505)
# ⚠️ 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:

2023-05-08 19:49:59 +08:00
if _ , err := fmt . Sscanf ( s , "%d_%s" , & id , & sha ) ; err != nil {
log . Error ( "could not parse data from pr_auto_merge queue (%v): %v" , s , err )
2022-05-07 19:05:52 +02:00
continue
}
2024-05-21 23:23:22 +08:00
handlePullRequestAutoMerge ( id , sha )
2022-05-07 19:05:52 +02:00
}
return nil
}
// ScheduleAutoMerge if schedule is false and no error, pull can be merged directly
2024-10-21 21:21:50 +02:00
func ScheduleAutoMerge ( ctx context . Context , doer * user_model . User , pull * issues_model . PullRequest , style repo_model . MergeStyle , message string , deleteBranch bool ) ( scheduled bool , err error ) {
2022-11-13 04:18:50 +08:00
err = db . WithTx ( ctx , func ( ctx context . Context ) error {
2024-10-21 21:21:50 +02:00
if err := pull_model . ScheduleAutoMerge ( ctx , doer , pull . ID , style , message , deleteBranch ) ; err != nil {
2022-05-08 15:46:34 +02:00
return err
}
scheduled = true
2022-05-07 19:05:52 +02:00
2022-06-13 17:37:59 +08:00
_ , err = issues_model . CreateAutoMergeComment ( ctx , issues_model . CommentTypePRScheduledToAutoMerge , pull , doer )
2022-05-08 15:46:34 +02:00
return err
2022-11-13 04:18:50 +08:00
} )
2022-06-20 12:02:49 +02:00
return scheduled , err
2022-05-08 15:46:34 +02:00
}
// RemoveScheduledAutoMerge cancels a previously scheduled pull request
2022-06-13 17:37:59 +08:00
func RemoveScheduledAutoMerge ( ctx context . Context , doer * user_model . User , pull * issues_model . PullRequest ) error {
2022-11-13 04:18:50 +08:00
return db . WithTx ( ctx , func ( ctx context . Context ) error {
2022-05-08 15:46:34 +02:00
if err := pull_model . DeleteScheduledAutoMerge ( ctx , pull . ID ) ; err != nil {
return err
}
2022-06-13 17:37:59 +08:00
_ , err := issues_model . CreateAutoMergeComment ( ctx , issues_model . CommentTypePRUnScheduledToAutoMerge , pull , doer )
2022-05-08 15:46:34 +02:00
return err
2022-11-13 04:18:50 +08:00
} )
2022-05-07 19:05:52 +02:00
}
2024-05-21 23:23:22 +08:00
// StartPRCheckAndAutoMergeBySHA start an automerge check and auto merge task for all pull requests of repository and SHA
func StartPRCheckAndAutoMergeBySHA ( ctx context . Context , sha string , repo * repo_model . Repository ) error {
2024-10-23 21:31:52 +02:00
return shared_automerge . StartPRCheckAndAutoMergeBySHA ( ctx , sha , repo )
2022-05-07 19:05:52 +02:00
}
2024-05-21 23:23:22 +08:00
// StartPRCheckAndAutoMerge start an automerge check and auto merge task for a pull request
func StartPRCheckAndAutoMerge ( ctx context . Context , pull * issues_model . PullRequest ) {
2024-10-23 21:31:52 +02:00
shared_automerge . StartPRCheckAndAutoMerge ( ctx , pull )
2022-05-07 19:05:52 +02:00
}
2024-05-21 23:23:22 +08:00
// handlePullRequestAutoMerge merge the pull request if all checks are successful
func handlePullRequestAutoMerge ( pullID int64 , sha string ) {
2022-05-07 19:05:52 +02:00
ctx , _ , finished := process . GetManager ( ) . AddContext ( graceful . GetManager ( ) . HammerContext ( ) ,
2023-02-03 23:11:48 +00:00
fmt . Sprintf ( "Handle AutoMerge of PR[%d] with sha[%s]" , pullID , sha ) )
2022-05-07 19:05:52 +02:00
defer finished ( )
2022-06-13 17:37:59 +08:00
pr , err := issues_model . GetPullRequestByID ( ctx , pullID )
2022-05-07 19:05:52 +02:00
if err != nil {
log . Error ( "GetPullRequestByID[%d]: %v" , pullID , err )
return
}
// Check if there is a scheduled pr in the db
exists , scheduledPRM , err := pull_model . GetScheduledMergeByPullID ( ctx , pr . ID )
if err != nil {
2023-02-03 23:11:48 +00:00
log . Error ( "%-v GetScheduledMergeByPullID: %v" , pr , err )
2022-05-07 19:05:52 +02:00
return
}
if ! exists {
return
}
2024-05-21 23:23:22 +08:00
if err = pr . LoadBaseRepo ( ctx ) ; err != nil {
log . Error ( "%-v LoadBaseRepo: %v" , pr , err )
return
}
// check the sha is the same as pull request head commit id
baseGitRepo , err := gitrepo . OpenRepository ( ctx , pr . BaseRepo )
if err != nil {
log . Error ( "OpenRepository: %v" , err )
return
}
defer baseGitRepo . Close ( )
headCommitID , err := baseGitRepo . GetRefCommitID ( pr . GetGitRefName ( ) )
if err != nil {
log . Error ( "GetRefCommitID: %v" , err )
return
}
if headCommitID != sha {
log . Warn ( "Head commit id of auto merge %-v does not match sha [%s], it may means the head branch has been updated. Just ignore this request because a new request expected in the queue" , pr , sha )
return
}
2022-05-07 19:05:52 +02:00
// Get all checks for this pr
// We get the latest sha commit hash again to handle the case where the check of a previous push
// did not succeed or was not finished yet.
2022-11-19 09:12:33 +01:00
if err = pr . LoadHeadRepo ( ctx ) ; err != nil {
2023-02-03 23:11:48 +00:00
log . Error ( "%-v LoadHeadRepo: %v" , pr , err )
2022-05-07 19:05:52 +02:00
return
}
2024-05-21 23:23:22 +08:00
var headGitRepo * git . Repository
if pr . BaseRepoID == pr . HeadRepoID {
headGitRepo = baseGitRepo
} else {
headGitRepo , err = gitrepo . OpenRepository ( ctx , pr . HeadRepo )
if err != nil {
log . Error ( "OpenRepository %-v: %v" , pr . HeadRepo , err )
return
}
defer headGitRepo . Close ( )
2022-05-07 19:05:52 +02:00
}
2024-08-20 14:17:21 +08:00
switch pr . Flow {
case issues_model . PullRequestFlowGithub :
headBranchExist := headGitRepo . IsBranchExist ( pr . HeadBranch )
if pr . HeadRepo == nil || ! headBranchExist {
log . Warn ( "Head branch of auto merge %-v does not exist [HeadRepoID: %d, Branch: %s]" , pr , pr . HeadRepoID , pr . HeadBranch )
return
}
case issues_model . PullRequestFlowAGit :
headBranchExist := git . IsReferenceExist ( ctx , baseGitRepo . Path , pr . GetGitRefName ( ) )
if ! headBranchExist {
log . Warn ( "Head branch of auto merge %-v does not exist [HeadRepoID: %d, Branch(Agit): %s]" , pr , pr . HeadRepoID , pr . HeadBranch )
return
}
default :
log . Error ( "wrong flow type %d" , pr . Flow )
2022-05-07 19:05:52 +02:00
return
}
// Check if all checks succeeded
pass , err := pull_service . IsPullCommitStatusPass ( ctx , pr )
if err != nil {
2023-02-03 23:11:48 +00:00
log . Error ( "%-v IsPullCommitStatusPass: %v" , pr , err )
2022-05-07 19:05:52 +02:00
return
}
if ! pass {
2023-02-03 23:11:48 +00:00
log . Info ( "Scheduled auto merge %-v has unsuccessful status checks" , pr )
2022-05-07 19:05:52 +02:00
return
}
// Merge if all checks succeeded
2022-12-03 10:48:26 +08:00
doer , err := user_model . GetUserByID ( ctx , scheduledPRM . DoerID )
2022-05-07 19:05:52 +02:00
if err != nil {
2023-02-03 23:11:48 +00:00
log . Error ( "Unable to get scheduled User[%d]: %v" , scheduledPRM . DoerID , err )
2022-05-07 19:05:52 +02:00
return
}
2022-05-11 18:09:36 +08:00
perm , err := access_model . GetUserRepoPermission ( ctx , pr . HeadRepo , doer )
2022-05-07 19:05:52 +02:00
if err != nil {
2023-02-03 23:11:48 +00:00
log . Error ( "GetUserRepoPermission %-v: %v" , pr . HeadRepo , err )
2022-05-07 19:05:52 +02:00
return
}
2024-05-09 01:11:43 +09:00
if err := pull_service . CheckPullMergeable ( ctx , doer , & perm , pr , pull_service . MergeCheckTypeGeneral , false ) ; err != nil {
2024-08-14 11:43:42 +02:00
if errors . Is ( err , pull_service . ErrUserNotAllowedToMerge ) {
2023-02-03 23:11:48 +00:00
log . Info ( "%-v was scheduled to automerge by an unauthorized user" , pr )
2022-05-07 19:05:52 +02:00
return
}
2024-05-09 01:11:43 +09:00
log . Error ( "%-v CheckPullMergeable: %v" , pr , err )
2022-05-07 19:05:52 +02:00
return
}
2022-11-03 16:49:00 +01:00
if err := pull_service . Merge ( ctx , pr , doer , baseGitRepo , scheduledPRM . MergeStyle , "" , scheduledPRM . Message , true ) ; err != nil {
2022-05-07 19:05:52 +02:00
log . Error ( "pull_service.Merge: %v" , err )
2024-05-21 23:23:22 +08:00
// FIXME: if merge failed, we should display some error message to the pull request page.
// The resolution is add a new column on automerge table named `error_message` to store the error message and displayed
// on the pull request page. But this should not be finished in a bug fix PR which will be backport to release branch.
2022-05-07 19:05:52 +02:00
return
}
2024-10-21 21:21:50 +02:00
if scheduledPRM . DeleteBranchAfterMerge {
err := repo_service . DeleteBranchAfterMerge ( ctx , doer , pr , headGitRepo )
if err != nil {
log . Error ( "%d repo_service.DeleteBranchIfUnused: %v" , pr . ID , err )
}
}
2022-05-07 19:05:52 +02:00
}