0
0
mirror of https://github.com/go-gitea/gitea.git synced 2024-12-22 23:05:37 +01:00
gitea/services/pull/check.go

397 lines
13 KiB
Go
Raw Normal View History

// Copyright 2019 The Gitea Authors.
// All rights reserved.
// SPDX-License-Identifier: MIT
package pull
import (
"context"
"errors"
"fmt"
"strconv"
"strings"
"code.gitea.io/gitea/models"
"code.gitea.io/gitea/models/db"
git_model "code.gitea.io/gitea/models/git"
issues_model "code.gitea.io/gitea/models/issues"
access_model "code.gitea.io/gitea/models/perm/access"
repo_model "code.gitea.io/gitea/models/repo"
"code.gitea.io/gitea/models/unit"
user_model "code.gitea.io/gitea/models/user"
"code.gitea.io/gitea/modules/git"
"code.gitea.io/gitea/modules/graceful"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/notification"
"code.gitea.io/gitea/modules/process"
"code.gitea.io/gitea/modules/queue"
"code.gitea.io/gitea/modules/timeutil"
asymkey_service "code.gitea.io/gitea/services/asymkey"
)
// prPatchCheckerQueue represents a queue to handle update pull request tests
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: ![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 13:49:59 +02:00
var prPatchCheckerQueue *queue.WorkerPoolQueue[string]
var (
ErrIsClosed = errors.New("pull is closed")
ErrUserNotAllowedToMerge = models.ErrDisallowedToMerge{}
ErrHasMerged = errors.New("has already been merged")
ErrIsWorkInProgress = errors.New("work in progress PRs cannot be merged")
ErrIsChecking = errors.New("cannot merge while conflict checking is in progress")
ErrNotMergableState = errors.New("not in mergeable state")
ErrDependenciesLeft = errors.New("is blocked by an open dependency")
)
// AddToTaskQueue adds itself to pull request test task queue.
func AddToTaskQueue(ctx context.Context, pr *issues_model.PullRequest) {
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: ![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 13:49:59 +02:00
pr.Status = issues_model.PullRequestStatusChecking
err := pr.UpdateColsIfNotMerged(ctx, "status")
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: ![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 13:49:59 +02:00
if err != nil {
log.Error("AddToTaskQueue(%-v).UpdateCols.(add to queue): %v", pr, err)
return
}
log.Trace("Adding %-v to the test pull requests queue", pr)
err = prPatchCheckerQueue.Push(strconv.FormatInt(pr.ID, 10))
if err != nil && err != queue.ErrAlreadyInQueue {
log.Error("Error adding %-v to the test pull requests queue: %v", pr, err)
}
}
type MergeCheckType int
const (
MergeCheckTypeGeneral MergeCheckType = iota // general merge checks for "merge", "rebase", "squash", etc
MergeCheckTypeManually // Manually Merged button (mark a PR as merged manually)
MergeCheckTypeAuto // Auto Merge (Scheduled Merge) After Checks Succeed
)
// CheckPullMergable check if the pull mergable based on all conditions (branch protection, merge options, ...)
func CheckPullMergable(stdCtx context.Context, doer *user_model.User, perm *access_model.Permission, pr *issues_model.PullRequest, mergeCheckType MergeCheckType, adminSkipProtectionCheck bool) error {
return db.WithTx(stdCtx, func(ctx context.Context) error {
if pr.HasMerged {
return ErrHasMerged
}
if err := pr.LoadIssue(ctx); err != nil {
log.Error("Unable to load issue[%d] for %-v: %v", pr.IssueID, pr, err)
return err
} else if pr.Issue.IsClosed {
return ErrIsClosed
}
if allowedMerge, err := IsUserAllowedToMerge(ctx, pr, *perm, doer); err != nil {
log.Error("Error whilst checking if %-v is allowed to merge %-v: %v", doer, pr, err)
return err
} else if !allowedMerge {
return ErrUserNotAllowedToMerge
}
if mergeCheckType == MergeCheckTypeManually {
// if doer is doing "manually merge" (mark as merged manually), do not check anything
return nil
}
if pr.IsWorkInProgress() {
return ErrIsWorkInProgress
}
if !pr.CanAutoMerge() && !pr.IsEmpty() {
return ErrNotMergableState
}
if pr.IsChecking() {
return ErrIsChecking
}
if err := CheckPullBranchProtections(ctx, pr, false); err != nil {
if !models.IsErrDisallowedToMerge(err) {
log.Error("Error whilst checking pull branch protection for %-v: %v", pr, err)
return err
}
// Now the branch protection check failed, check whether the failure could be skipped (skip by setting err = nil)
// * when doing Auto Merge (Scheduled Merge After Checks Succeed), skip the branch protection check
if mergeCheckType == MergeCheckTypeAuto {
err = nil
}
// * if the doer is admin, they could skip the branch protection check
if adminSkipProtectionCheck {
if isRepoAdmin, errCheckAdmin := access_model.IsUserRepoAdmin(ctx, pr.BaseRepo, doer); errCheckAdmin != nil {
log.Error("Unable to check if %-v is a repo admin in %-v: %v", doer, pr.BaseRepo, errCheckAdmin)
return errCheckAdmin
} else if isRepoAdmin {
err = nil // repo admin can skip the check, so clear the error
}
}
// If there is still a branch protection check error, return it
if err != nil {
return err
}
}
if _, err := isSignedIfRequired(ctx, pr, doer); err != nil {
return err
}
if noDeps, err := issues_model.IssueNoDependenciesLeft(ctx, pr.Issue); err != nil {
return err
} else if !noDeps {
return ErrDependenciesLeft
}
return nil
})
}
// isSignedIfRequired check if merge will be signed if required
func isSignedIfRequired(ctx context.Context, pr *issues_model.PullRequest, doer *user_model.User) (bool, error) {
pb, err := git_model.GetFirstMatchProtectedBranchRule(ctx, pr.BaseRepoID, pr.BaseBranch)
if err != nil {
return false, err
}
if pb == nil || !pb.RequireSignedCommits {
return true, nil
}
sign, _, _, err := asymkey_service.SignMerge(ctx, pr, doer, pr.BaseRepo.RepoPath(), pr.BaseBranch, pr.GetGitRefName())
return sign, err
}
// checkAndUpdateStatus checks if pull request is possible to leaving checking status,
// and set to be either conflict or mergeable.
func checkAndUpdateStatus(ctx context.Context, pr *issues_model.PullRequest) {
// If status has not been changed to conflict by testPatch then we are mergeable
if pr.Status == issues_model.PullRequestStatusChecking {
pr.Status = issues_model.PullRequestStatusMergeable
}
// Make sure there is no waiting test to process before leaving the checking status.
has, err := prPatchCheckerQueue.Has(strconv.FormatInt(pr.ID, 10))
if err != nil {
log.Error("Unable to check if the queue is waiting to reprocess %-v. Error: %v", pr, err)
}
if has {
log.Trace("Not updating status for %-v as it is due to be rechecked", pr)
return
}
if err := pr.UpdateColsIfNotMerged(ctx, "merge_base", "status", "conflicted_files", "changed_protected_files"); err != nil {
log.Error("Update[%-v]: %v", pr, err)
}
}
// getMergeCommit checks if a pull request has been merged
// Returns the git.Commit of the pull request if merged
func getMergeCommit(ctx context.Context, pr *issues_model.PullRequest) (*git.Commit, error) {
if err := pr.LoadBaseRepo(ctx); err != nil {
return nil, fmt.Errorf("unable to load base repo for %s: %w", pr, err)
}
prHeadRef := pr.GetGitRefName()
// Check if the pull request is merged into BaseBranch
if _, _, err := git.NewCommand(ctx, "merge-base", "--is-ancestor").
AddDynamicArguments(prHeadRef, pr.BaseBranch).
RunStdString(&git.RunOpts{Dir: pr.BaseRepo.RepoPath()}); err != nil {
if strings.Contains(err.Error(), "exit status 1") {
// prHeadRef is not an ancestor of the base branch
return nil, nil
}
// Errors are signaled by a non-zero status that is not 1
return nil, fmt.Errorf("%-v git merge-base --is-ancestor: %w", pr, err)
}
// If merge-base successfully exits then prHeadRef is an ancestor of pr.BaseBranch
// Find the head commit id
prHeadCommitID, err := git.GetFullCommitID(ctx, pr.BaseRepo.RepoPath(), prHeadRef)
if err != nil {
return nil, fmt.Errorf("GetFullCommitID(%s) in %s: %w", prHeadRef, pr.BaseRepo.FullName(), err)
}
// Get the commit from BaseBranch where the pull request got merged
mergeCommit, _, err := git.NewCommand(ctx, "rev-list", "--ancestry-path", "--merges", "--reverse").
AddDynamicArguments(prHeadCommitID + ".." + pr.BaseBranch).
RunStdString(&git.RunOpts{Dir: pr.BaseRepo.RepoPath()})
if err != nil {
return nil, fmt.Errorf("git rev-list --ancestry-path --merges --reverse: %w", err)
} else if len(mergeCommit) < git.SHAFullLength {
// PR was maybe fast-forwarded, so just use last commit of PR
mergeCommit = prHeadCommitID
}
mergeCommit = strings.TrimSpace(mergeCommit)
gitRepo, err := git.OpenRepository(ctx, pr.BaseRepo.RepoPath())
if err != nil {
return nil, fmt.Errorf("%-v OpenRepository: %w", pr.BaseRepo, err)
}
defer gitRepo.Close()
commit, err := gitRepo.GetCommit(mergeCommit)
if err != nil {
return nil, fmt.Errorf("GetMergeCommit[%s]: %w", mergeCommit, err)
}
return commit, nil
}
// manuallyMerged checks if a pull request got manually merged
// When a pull request got manually merged mark the pull request as merged
func manuallyMerged(ctx context.Context, pr *issues_model.PullRequest) bool {
if err := pr.LoadBaseRepo(ctx); err != nil {
log.Error("%-v LoadBaseRepo: %v", pr, err)
return false
}
if unit, err := pr.BaseRepo.GetUnit(ctx, unit.TypePullRequests); err == nil {
config := unit.PullRequestsConfig()
if !config.AutodetectManualMerge {
return false
}
} else {
log.Error("%-v BaseRepo.GetUnit(unit.TypePullRequests): %v", pr, err)
return false
}
commit, err := getMergeCommit(ctx, pr)
if err != nil {
log.Error("%-v getMergeCommit: %v", pr, err)
return false
}
if commit == nil {
// no merge commit found
return false
}
pr.MergedCommitID = commit.ID.String()
pr.MergedUnix = timeutil.TimeStamp(commit.Author.When.Unix())
pr.Status = issues_model.PullRequestStatusManuallyMerged
Add context cache as a request level cache (#22294) To avoid duplicated load of the same data in an HTTP request, we can set a context cache to do that. i.e. Some pages may load a user from a database with the same id in different areas on the same page. But the code is hidden in two different deep logic. How should we share the user? As a result of this PR, now if both entry functions accept `context.Context` as the first parameter and we just need to refactor `GetUserByID` to reuse the user from the context cache. Then it will not be loaded twice on an HTTP request. But of course, sometimes we would like to reload an object from the database, that's why `RemoveContextData` is also exposed. The core context cache is here. It defines a new context ```go type cacheContext struct { ctx context.Context data map[any]map[any]any lock sync.RWMutex } var cacheContextKey = struct{}{} func WithCacheContext(ctx context.Context) context.Context { return context.WithValue(ctx, cacheContextKey, &cacheContext{ ctx: ctx, data: make(map[any]map[any]any), }) } ``` Then you can use the below 4 methods to read/write/del the data within the same context. ```go func GetContextData(ctx context.Context, tp, key any) any func SetContextData(ctx context.Context, tp, key, value any) func RemoveContextData(ctx context.Context, tp, key any) func GetWithContextCache[T any](ctx context.Context, cacheGroupKey string, cacheTargetID any, f func() (T, error)) (T, error) ``` Then let's take a look at how `system.GetString` implement it. ```go func GetSetting(ctx context.Context, key string) (string, error) { return cache.GetWithContextCache(ctx, contextCacheKey, key, func() (string, error) { return cache.GetString(genSettingCacheKey(key), func() (string, error) { res, err := GetSettingNoCache(ctx, key) if err != nil { return "", err } return res.SettingValue, nil }) }) } ``` First, it will check if context data include the setting object with the key. If not, it will query from the global cache which may be memory or a Redis cache. If not, it will get the object from the database. In the end, if the object gets from the global cache or database, it will be set into the context cache. An object stored in the context cache will only be destroyed after the context disappeared.
2023-02-15 14:37:34 +01:00
merger, _ := user_model.GetUserByEmail(ctx, commit.Author.Email)
// When the commit author is unknown set the BaseRepo owner as merger
if merger == nil {
if pr.BaseRepo.Owner == nil {
if err = pr.BaseRepo.LoadOwner(ctx); err != nil {
log.Error("%-v BaseRepo.LoadOwner: %v", pr, err)
return false
}
}
merger = pr.BaseRepo.Owner
}
pr.Merger = merger
pr.MergerID = merger.ID
if merged, err := pr.SetMerged(ctx); err != nil {
log.Error("%-v setMerged : %v", pr, err)
return false
} else if !merged {
return false
}
notification.NotifyMergePullRequest(ctx, merger, pr)
log.Info("manuallyMerged[%-v]: Marked as manually merged into %s/%s by commit id: %s", pr, pr.BaseRepo.Name, pr.BaseBranch, commit.ID.String())
return true
}
// InitializePullRequests checks and tests untested patches of pull requests.
func InitializePullRequests(ctx context.Context) {
prs, err := issues_model.GetPullRequestIDsByCheckStatus(issues_model.PullRequestStatusChecking)
if err != nil {
log.Error("Find Checking PRs: %v", err)
return
}
for _, prID := range prs {
select {
case <-ctx.Done():
return
default:
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: ![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 13:49:59 +02:00
log.Trace("Adding PR[%d] to the pull requests patch checking queue", prID)
if err := prPatchCheckerQueue.Push(strconv.FormatInt(prID, 10)); err != nil {
log.Error("Error adding PR[%d] to the pull requests patch checking queue %v", prID, err)
}
}
}
}
// 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: ![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 13:49:59 +02:00
func handler(items ...string) []string {
for _, s := range items {
id, _ := strconv.ParseInt(s, 10, 64)
testPR(id)
}
return nil
}
func testPR(id int64) {
pullWorkingPool.CheckIn(fmt.Sprint(id))
defer pullWorkingPool.CheckOut(fmt.Sprint(id))
ctx, _, finished := process.GetManager().AddContext(graceful.GetManager().HammerContext(), fmt.Sprintf("Test PR[%d] from patch checking queue", id))
defer finished()
pr, err := issues_model.GetPullRequestByID(ctx, id)
if err != nil {
log.Error("Unable to GetPullRequestByID[%d] for testPR: %v", id, err)
return
}
log.Trace("Testing %-v", pr)
defer func() {
log.Trace("Done testing %-v (status: %s)", pr, pr.Status)
}()
if pr.HasMerged {
log.Trace("%-v is already merged (status: %s, merge commit: %s)", pr, pr.Status, pr.MergedCommitID)
return
}
if manuallyMerged(ctx, pr) {
log.Trace("%-v is manually merged (status: %s, merge commit: %s)", pr, pr.Status, pr.MergedCommitID)
return
}
if err := TestPatch(pr); err != nil {
log.Error("testPatch[%-v]: %v", pr, err)
pr.Status = issues_model.PullRequestStatusError
if err := pr.UpdateCols("status"); err != nil {
log.Error("update pr [%-v] status to PullRequestStatusError failed: %v", pr, err)
}
return
}
checkAndUpdateStatus(ctx, pr)
}
// CheckPRsForBaseBranch check all pulls with baseBrannch
func CheckPRsForBaseBranch(ctx context.Context, baseRepo *repo_model.Repository, baseBranchName string) error {
prs, err := issues_model.GetUnmergedPullRequestsByBaseInfo(ctx, baseRepo.ID, baseBranchName)
[Enhancement] Allow admin to merge pr with protected file changes (#12078) * [Enhancement] Allow admin to merge pr with protected file changes As tilte, show protected message in diff page and merge box. Signed-off-by: a1012112796 <1012112796@qq.com> * remove unused ver * Update options/locale/locale_en-US.ini Co-authored-by: Cirno the Strongest <1447794+CirnoT@users.noreply.github.com> * Add TrN * Apply suggestions from code review * fix lint * Update options/locale/locale_en-US.ini Co-authored-by: zeripath <art27@cantab.net> * Apply suggestions from code review * move pr proteced files check to TestPatch * Call TestPatch when protected branches settings changed * Apply review suggestion @CirnoT * move to service @lunny * slightly restructure routers/private/hook.go Adds a lot of comments and simplifies the logic Signed-off-by: Andrew Thornton <art27@cantab.net> * placate lint Signed-off-by: Andrew Thornton <art27@cantab.net> * skip duplicate protected files check * fix check logic * slight refactor of TestPatch Signed-off-by: Andrew Thornton <art27@cantab.net> * When checking for protected files changes in TestPatch use the temporary repository Signed-off-by: Andrew Thornton <art27@cantab.net> * fix introduced issue with hook Signed-off-by: Andrew Thornton <art27@cantab.net> * Remove the check on PR index being greater than 0 as it unnecessary Signed-off-by: Andrew Thornton <art27@cantab.net> Co-authored-by: techknowlogick <matti@mdranta.net> Co-authored-by: Cirno the Strongest <1447794+CirnoT@users.noreply.github.com> Co-authored-by: zeripath <art27@cantab.net> Co-authored-by: techknowlogick <techknowlogick@gitea.io>
2020-10-13 20:50:57 +02:00
if err != nil {
return err
}
for _, pr := range prs {
AddToTaskQueue(ctx, pr)
[Enhancement] Allow admin to merge pr with protected file changes (#12078) * [Enhancement] Allow admin to merge pr with protected file changes As tilte, show protected message in diff page and merge box. Signed-off-by: a1012112796 <1012112796@qq.com> * remove unused ver * Update options/locale/locale_en-US.ini Co-authored-by: Cirno the Strongest <1447794+CirnoT@users.noreply.github.com> * Add TrN * Apply suggestions from code review * fix lint * Update options/locale/locale_en-US.ini Co-authored-by: zeripath <art27@cantab.net> * Apply suggestions from code review * move pr proteced files check to TestPatch * Call TestPatch when protected branches settings changed * Apply review suggestion @CirnoT * move to service @lunny * slightly restructure routers/private/hook.go Adds a lot of comments and simplifies the logic Signed-off-by: Andrew Thornton <art27@cantab.net> * placate lint Signed-off-by: Andrew Thornton <art27@cantab.net> * skip duplicate protected files check * fix check logic * slight refactor of TestPatch Signed-off-by: Andrew Thornton <art27@cantab.net> * When checking for protected files changes in TestPatch use the temporary repository Signed-off-by: Andrew Thornton <art27@cantab.net> * fix introduced issue with hook Signed-off-by: Andrew Thornton <art27@cantab.net> * Remove the check on PR index being greater than 0 as it unnecessary Signed-off-by: Andrew Thornton <art27@cantab.net> Co-authored-by: techknowlogick <matti@mdranta.net> Co-authored-by: Cirno the Strongest <1447794+CirnoT@users.noreply.github.com> Co-authored-by: zeripath <art27@cantab.net> Co-authored-by: techknowlogick <techknowlogick@gitea.io>
2020-10-13 20:50:57 +02:00
}
return nil
}
// Init runs the task queue to test all the checking status pull requests
func Init() error {
prPatchCheckerQueue = queue.CreateUniqueQueue(graceful.GetManager().ShutdownContext(), "pr_patch_checker", handler)
if prPatchCheckerQueue == nil {
return fmt.Errorf("unable to create pr_patch_checker queue")
}
go graceful.GetManager().RunWithCancel(prPatchCheckerQueue)
go graceful.GetManager().RunWithShutdownContext(InitializePullRequests)
return nil
}