2019-11-02 01:51:22 +03:00
// Copyright 2019 The Gitea Authors. All rights reserved.
2022-11-27 21:20:29 +03:00
// SPDX-License-Identifier: MIT
2019-11-02 01:51:22 +03:00
package webhook
import (
2022-05-20 17:08:52 +03:00
"context"
2023-07-28 20:46:48 +03:00
"errors"
2019-11-02 01:51:22 +03:00
"fmt"
"strings"
2021-12-10 04:27:50 +03:00
repo_model "code.gitea.io/gitea/models/repo"
2022-10-21 19:21:56 +03:00
user_model "code.gitea.io/gitea/models/user"
2021-11-10 08:13:16 +03:00
webhook_model "code.gitea.io/gitea/models/webhook"
2019-11-02 01:51:22 +03:00
"code.gitea.io/gitea/modules/git"
2022-04-25 21:03:01 +03:00
"code.gitea.io/gitea/modules/graceful"
2019-11-02 01:51:22 +03:00
"code.gitea.io/gitea/modules/log"
2022-04-25 21:03:01 +03:00
"code.gitea.io/gitea/modules/queue"
2019-11-02 01:51:22 +03:00
"code.gitea.io/gitea/modules/setting"
api "code.gitea.io/gitea/modules/structs"
2021-08-12 15:43:08 +03:00
"code.gitea.io/gitea/modules/util"
2023-01-01 18:23:15 +03:00
webhook_module "code.gitea.io/gitea/modules/webhook"
2021-08-12 15:43:08 +03:00
2019-11-02 01:51:22 +03:00
"github.com/gobwas/glob"
)
2020-12-08 13:41:14 +03:00
type webhook struct {
2023-01-01 18:23:15 +03:00
name webhook_module . HookType
payloadCreator func ( p api . Payloader , event webhook_module . HookEventType , meta string ) ( api . Payloader , error )
2020-12-08 13:41:14 +03:00
}
2023-01-01 18:23:15 +03:00
var webhooks = map [ webhook_module . HookType ] * webhook {
webhook_module . SLACK : {
name : webhook_module . SLACK ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetSlackPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . DISCORD : {
name : webhook_module . DISCORD ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetDiscordPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . DINGTALK : {
name : webhook_module . DINGTALK ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetDingtalkPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . TELEGRAM : {
name : webhook_module . TELEGRAM ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetTelegramPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . MSTEAMS : {
name : webhook_module . MSTEAMS ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetMSTeamsPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . FEISHU : {
name : webhook_module . FEISHU ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetFeishuPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . MATRIX : {
name : webhook_module . MATRIX ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetMatrixPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . WECHATWORK : {
name : webhook_module . WECHATWORK ,
2022-01-20 20:46:10 +03:00
payloadCreator : GetWechatworkPayload ,
} ,
2023-01-01 18:23:15 +03:00
webhook_module . PACKAGIST : {
name : webhook_module . PACKAGIST ,
2022-01-23 16:46:30 +03:00
payloadCreator : GetPackagistPayload ,
} ,
2022-01-20 20:46:10 +03:00
}
2020-12-08 13:41:14 +03:00
// IsValidHookTaskType returns true if a webhook registered
func IsValidHookTaskType ( name string ) bool {
2023-01-01 18:23:15 +03:00
if name == webhook_module . GITEA || name == webhook_module . GOGS {
2020-12-12 18:33:19 +03:00
return true
}
2022-06-20 13:02:49 +03:00
_ , ok := webhooks [ name ]
2020-12-08 13:41:14 +03:00
return ok
}
2019-11-02 05:35:12 +03:00
// hookQueue is a global queue of web hooks
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 14:49:59 +03:00
var hookQueue * queue . WorkerPoolQueue [ int64 ]
2019-11-02 01:51:22 +03:00
// getPayloadBranch returns branch for hook event, if applicable.
func getPayloadBranch ( p api . Payloader ) string {
switch pp := p . ( type ) {
case * api . CreatePayload :
if pp . RefType == "branch" {
return pp . Ref
}
case * api . DeletePayload :
if pp . RefType == "branch" {
return pp . Ref
}
case * api . PushPayload :
if strings . HasPrefix ( pp . Ref , git . BranchPrefix ) {
return pp . Ref [ len ( git . BranchPrefix ) : ]
}
}
return ""
}
2022-10-21 19:21:56 +03:00
// EventSource represents the source of a webhook action. Repository and/or Owner must be set.
type EventSource struct {
Repository * repo_model . Repository
Owner * user_model . User
}
2022-04-25 21:03:01 +03:00
2022-10-21 19:21:56 +03:00
// handle delivers hook tasks
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 14:49:59 +03:00
func handler ( items ... int64 ) [ ] int64 {
2022-10-21 19:21:56 +03:00
ctx := graceful . GetManager ( ) . HammerContext ( )
2022-04-25 21:03:01 +03:00
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 14:49:59 +03:00
for _ , taskID := range items {
task , err := webhook_model . GetHookTaskByID ( ctx , taskID )
2022-04-25 21:03:01 +03:00
if err != nil {
2023-07-28 20:46:48 +03:00
if errors . Is ( err , util . ErrNotExist ) {
log . Warn ( "GetHookTaskByID[%d] warn: %v" , taskID , err )
} else {
log . Error ( "GetHookTaskByID[%d] failed: %v" , taskID , err )
}
2022-11-23 17:10:04 +03:00
continue
}
if task . IsDelivered {
// Already delivered in the meantime
log . Trace ( "Task[%d] has already been delivered" , task . ID )
continue
}
if err := Deliver ( ctx , task ) ; err != nil {
log . Error ( "Unable to deliver webhook task[%d]: %v" , task . ID , err )
2022-04-25 21:03:01 +03:00
}
}
2022-10-21 19:21:56 +03:00
2022-04-25 21:03:01 +03:00
return nil
}
2022-11-23 17:10:04 +03:00
func enqueueHookTask ( taskID int64 ) error {
err := hookQueue . Push ( taskID )
2022-04-25 21:03:01 +03:00
if err != nil && err != queue . ErrAlreadyInQueue {
return err
}
return nil
}
2021-11-10 08:13:16 +03:00
func checkBranch ( w * webhook_model . Webhook , branch string ) bool {
2019-11-02 01:51:22 +03:00
if w . BranchFilter == "" || w . BranchFilter == "*" {
return true
}
g , err := glob . Compile ( w . BranchFilter )
if err != nil {
// should not really happen as BranchFilter is validated
log . Error ( "CheckBranch failed: %s" , err )
return false
}
return g . Match ( branch )
}
2022-10-21 19:21:56 +03:00
// PrepareWebhook creates a hook task and enqueues it for processing
2023-01-01 18:23:15 +03:00
func PrepareWebhook ( ctx context . Context , w * webhook_model . Webhook , event webhook_module . HookEventType , p api . Payloader ) error {
2021-02-11 20:34:34 +03:00
// Skip sending if webhooks are disabled.
if setting . DisableWebhooks {
return nil
}
2019-11-02 01:51:22 +03:00
for _ , e := range w . EventCheckers ( ) {
if event == e . Type {
if ! e . Has ( ) {
return nil
}
2021-02-22 21:54:01 +03:00
break
2019-11-02 01:51:22 +03:00
}
}
2020-09-03 05:46:02 +03:00
// Avoid sending "0 new commits" to non-integration relevant webhooks (e.g. slack, discord, etc.).
// Integration webhooks (e.g. drone) still receive the required data.
if pushEvent , ok := p . ( * api . PushPayload ) ; ok &&
2023-01-01 18:23:15 +03:00
w . Type != webhook_module . GITEA && w . Type != webhook_module . GOGS &&
2020-09-03 05:46:02 +03:00
len ( pushEvent . Commits ) == 0 {
return nil
}
2019-11-02 01:51:22 +03:00
// If payload has no associated branch (e.g. it's a new tag, issue, etc.),
// branch filter has no effect.
if branch := getPayloadBranch ( p ) ; branch != "" {
if ! checkBranch ( w , branch ) {
log . Info ( "Branch %q doesn't match branch filter %q, skipping" , branch , w . BranchFilter )
return nil
}
}
var payloader api . Payloader
var err error
2021-01-06 18:11:23 +03:00
webhook , ok := webhooks [ w . Type ]
2020-12-08 13:41:14 +03:00
if ok {
payloader , err = webhook . payloadCreator ( p , event , w . Meta )
2020-03-28 16:09:55 +03:00
if err != nil {
2022-10-24 22:29:17 +03:00
return fmt . Errorf ( "create payload for %s[%s]: %w" , w . Type , event , err )
2020-03-28 16:09:55 +03:00
}
2020-12-08 13:41:14 +03:00
} else {
2019-11-02 01:51:22 +03:00
payloader = p
}
2022-10-21 19:21:56 +03:00
task , err := webhook_model . CreateHookTask ( ctx , & webhook_model . HookTask {
2021-06-27 22:21:09 +03:00
HookID : w . ID ,
Payloader : payloader ,
EventType : event ,
2022-10-21 19:21:56 +03:00
} )
if err != nil {
2022-10-24 22:29:17 +03:00
return fmt . Errorf ( "CreateHookTask: %w" , err )
2019-11-02 01:51:22 +03:00
}
2022-10-21 19:21:56 +03:00
2022-11-23 17:10:04 +03:00
return enqueueHookTask ( task . ID )
2019-11-02 01:51:22 +03:00
}
// PrepareWebhooks adds new webhooks to task queue for given payload.
2023-01-01 18:23:15 +03:00
func PrepareWebhooks ( ctx context . Context , source EventSource , event webhook_module . HookEventType , p api . Payloader ) error {
2022-10-21 19:21:56 +03:00
owner := source . Owner
2019-11-02 05:35:12 +03:00
2022-10-21 19:21:56 +03:00
var ws [ ] * webhook_model . Webhook
2019-11-02 01:51:22 +03:00
2022-10-21 19:21:56 +03:00
if source . Repository != nil {
repoHooks , err := webhook_model . ListWebhooksByOpts ( ctx , & webhook_model . ListWebhookOptions {
RepoID : source . Repository . ID ,
IsActive : util . OptionalBoolTrue ,
} )
if err != nil {
2022-10-24 22:29:17 +03:00
return fmt . Errorf ( "ListWebhooksByOpts: %w" , err )
2022-10-21 19:21:56 +03:00
}
ws = append ( ws , repoHooks ... )
2022-11-19 11:12:33 +03:00
owner = source . Repository . MustOwner ( ctx )
2019-11-02 01:51:22 +03:00
}
2023-03-10 17:28:32 +03:00
// append additional webhooks of a user or organization
if owner != nil {
ownerHooks , err := webhook_model . ListWebhooksByOpts ( ctx , & webhook_model . ListWebhookOptions {
OwnerID : owner . ID ,
2021-08-12 15:43:08 +03:00
IsActive : util . OptionalBoolTrue ,
} )
2019-11-02 01:51:22 +03:00
if err != nil {
2022-10-24 22:29:17 +03:00
return fmt . Errorf ( "ListWebhooksByOpts: %w" , err )
2019-11-02 01:51:22 +03:00
}
2023-03-10 17:28:32 +03:00
ws = append ( ws , ownerHooks ... )
2019-11-02 01:51:22 +03:00
}
2020-03-09 01:08:05 +03:00
// Add any admin-defined system webhooks
2022-05-20 17:08:52 +03:00
systemHooks , err := webhook_model . GetSystemWebhooks ( ctx , util . OptionalBoolTrue )
2020-03-09 01:08:05 +03:00
if err != nil {
2022-10-24 22:29:17 +03:00
return fmt . Errorf ( "GetSystemWebhooks: %w" , err )
2020-03-09 01:08:05 +03:00
}
ws = append ( ws , systemHooks ... )
2019-11-02 01:51:22 +03:00
if len ( ws ) == 0 {
return nil
}
for _ , w := range ws {
2022-10-21 19:21:56 +03:00
if err := PrepareWebhook ( ctx , w , event , p ) ; err != nil {
2019-11-02 01:51:22 +03:00
return err
}
}
return nil
}
2022-01-06 00:00:20 +03:00
// ReplayHookTask replays a webhook task
2022-10-21 19:21:56 +03:00
func ReplayHookTask ( ctx context . Context , w * webhook_model . Webhook , uuid string ) error {
task , err := webhook_model . ReplayHookTask ( ctx , w . ID , uuid )
2022-01-06 00:00:20 +03:00
if err != nil {
return err
}
2022-11-23 17:10:04 +03:00
return enqueueHookTask ( task . ID )
2022-01-06 00:00:20 +03:00
}