* Move pluginCommands into Channels
We move pluginCommands, pluginCommandsLock
into Channels.
We also move the plugin related route handlers
under Channels and move the init code under
NewChannels. To achieve this, the router initialization
is bumped up.
Along with it, we clean up some App methods
which were just wrappers over Channel methods.
Instead, we call the Channel method directly
to make things more readable and easy to understand.
```release-note
NONE
```
* fix tests
```release-note
NONE
```
* Move plugins under Channels
We move all plugin related fields under *Channels.
This essentially migrates several methods from being under
*Server to under *Channels.
We also move the plugin startup and shutdown code
to be under Channels.Start and Channels.Shutdown.
While here, we remove the getPluginPublicKeyFiles
method which was a one-line method which uselessly
returned an error.
Lastly, we fix the product initialization order which
was incorrect previously. Products are dependent on
the main server.
So startup should be products -> server.
And shutdown should be server -> products.
```release-note
NONE
```
* Added app layer
```release-note
NONE
```
* Incorporate suggestions
```release-note
NONE
```
* Move Channels into App
In this PR, we make Channels as part of App
instead of Server. This is part of the transition period
of moving fields from Server to Channels.
For now, Channels contains Server. So the hierarchy is
App -> Channels -> Server.
And as a first step, we also move httpService to Channels.
```release-note
NONE
```
* Fixing another test
```release-note
NONE
```
* new method
```release-note
NONE
```
* add request context
* move initialialization to server
* use app interface instead of global app functions
* remove app context from webconn
* cleanup
* remove duplicated services
* move context to separate package
* remove finalize init method and move content to NewServer function
* restart workers and schedulers after adding license for tests
* reflect review comments
Co-authored-by: Mattermod <mattermod@users.noreply.github.com>
* Replacing require.nil in app layer
* Fixing some tests
* Fixing tests
* Reverting the error correction, defering it for another PR
* Fixing golangci-lint errors
Co-authored-by: Mattermod <mattermod@users.noreply.github.com>
* Reducing the filestore dependencies from the rest of the source code
* Making more generic config conversion to FileBackendSettings
* Fixing usage of the NewFileBackend function
* Fixing more usages of the NewFileBackend function
* Fix some linter errors
* Fix more linter errors
* Fixing some unit tests
* Fixing linter problem
* Addressing PR review comments
* Simplifing the CopyFile for tests
Co-authored-by: Mattermod <mattermod@users.noreply.github.com>
We shutdown all plugins prior to checking the log buffer
so that there are no active plugins writing to the buffer
when we read the contents.
https://mattermost.atlassian.net/browse/MM-31324
```release-note
NONE
```
* Start plugins concurrently
* Determine which plugins need to be activated/deactivated before changing their state
* Test activation/deactivation of plugins in TestSyncPluginsActiveState
* Change test comments
* remove unneeded temp var
th.InitBasic initializes users, channels, and teams. But a lot of
tests don't use it at all, causing unncessary delays. A single InitBasic
call takes around 400ms. This PR removes 71 of them, approximately shaving
off 28 seconds. Arguably, this would be more because some tests run in loops.
* Implemented AmazonS3PathPrefix
* Remove unecessary method
* fix for test
* fix for test which are failing
* fix for test which are failing
* fix for test
Co-authored-by: Dusan Panic <dusan@salestrekker.com>
* MM-22706: pass along set_online flag in websocket response
To let the client know whether a user has created a post without
being online or not, we get the set_online query param and
pass it down to the websocket event being passed down to the client.
With this PR, the "data" field of the `posted` event will contain
a `set_online` boolean field set to true/false depending on the
query_param set_online value set in the createPost call.
* Setting to false for auto responder
Co-authored-by: mattermod <mattermod@users.noreply.github.com>
* explicitly assert panic as error log
* Revert "[MM-18150] plugin panic trace should not be lost (#13559)"
This reverts commit 5d928b4f94, while leaving the unit tests intact
and now asserting debug logs instead.
* missing license header
* Transit panic from debug to error
* Parse plugin's StdErr and output panic to the mlog.Error
* Add unit tests
* Change log test
* Remove buffer from logger
* Remove 'panic' string filter
* Change *Buffer to io.Writer
Co-authored-by: mattermod <mattermod@users.noreply.github.com>
Setting a `nil` value should actually delete the row instead of preserve it with a `nil` value.
The existing tests passed because they only checked the return value of `nil`. When running the tests on Postgres, this value is returned as an empty `[]byte{}` instead of `nil`, so update the test to reflect this semantically equivalent behaviour.
* MM-21103: change plugin signature path
Save as `<plugin_id>.tar.gz.sig` instead of `<plugin_id>.sig`. The latter was a relic of the previous design to support multiple plugin signatures, but now creates an inconsistency with how the original source files were supplied as `<some_name>.tar.gz` and `<some_name>.tar.gz.sig`.
Fixes: https://mattermost.atlassian.net/browse/MM-21103
* relax signature matches to avoid assuming signatures always exist
* Consistent license message for all the go files
* Fixing the last set of unconsistencies with the license headers
* Addressing PR review comments
* Fixing busy.go and busy_test.go license header
* MM-17549: use StatusNotFound when deleting plugin (#12983)
Trying to delete a plugin that does not exist should fail with a 404, not a 400.
Relates-to: https://mattermost.atlassian.net/browse/MM-17549
* MM-19630: marketplace: model ReleaseNotesURL (#13083)
Automatic Merge
* MM-20065: allow prepackaged plugin upgrade (#13076)
* MM-20065: allow prepackaged plugin upgrade
When locally installing prepackaged plugins, skip if a plugin exists with the same id and is the same or a newer version.
This is effectively a "poor man's" rework of prepackaged plugins to allow upgrade of prepackaged plugins via the marketplace. The larger plan to rework prepackaged plugins was deferred from v5.18.
Fixes: https://mattermost.atlassian.net/browse/MM-20065
* eliminate unnecessary installPlugin
* fix TestPluginSync defaults to match minio
* cleanExistingBundles
* close prepackaged filereader
* simplify
* Add SetWithOptions
* Avoid passing two structs to the functions
* Rename ExpiryInSeconds -> ExpireInSeconds
* Use t.Run for the tests
* Fix build
* Address feedback
* Update log message
* Update docs and use KVSetWithOptions in KVCompareAndSetJSON
* Improve code style
* Use struct instead of pointer to struct
* Fix minimum server versions
* Update documentation
* Address feedback
* Revert new implemention of kv helpers
* Adress feedback
* MM-16272 - Synchronize plugins across cluster (#11611)
* MM-16272 - Synchronize plugins across cluster
* Adding a test
* MM-16272 - Fixed tests
* MM-16272 - PR feedback
* MM-16270 - Plugin Sync (#11615)
* Initial implementation for plugin synch with file store. WIP
* Removed ListAll implementation. Used ListDirectory and change localstore to be consistent and return all items (files and folders) from directory
* Refactored plugin filestore operations out of main install/remove plugin
* Fixing error handling details
* Changes to use structured logging
* More logging fixes
* Wording and comments improvements
* Error handling and control flow improvements
* Changed managed flag check to use os.stat
* Added file store plugin dir and filename consts
* Replaced FileRead to use a the FileReader in PluginSync
* Minor styling and PR feedback changes
* Minor error handling improvements
* Added unit test for SyncPlugins. Changed SyncPlugins to use plugins environment to list available plugins
* PR Feedback improvements
* Minor err handling fix
* Removing FileStorePath from PluginEventData (#11644)
* Fix plugin path (#11654)
* tweak path, logging
Fix an issue not finding the plugins folder in S3. Tweak logging messages to add additional clarity.
* Removing FileExists check when Syncing plugins. Updated localstore to not return an error when directory does not exist
* PR Feedback
* Install prepackaged plugins locally only (#11656)
* s/uninstall/remove
* Updated ClusterMessage comment
* Updated PluginSync to test against s3 + local storage
* MM-14246 - Plugin framework: support transactional semantics with KV Store
Rename old, new variable names
Moving New function to the bottom
* Made CompareAndUpdate sync, updated tests
* Removed going through channel in CompareAndSetPluginKey
* Inserting new key when oldValue is nil to KVCompareAndSet
* Updated error text to include CompareAndSet
* Rerouted the unit test logs through t.Log
* resolving merge confilvts
* Update testing.go
* Update helper_test.go
* Added godocs for NewTestingLogger
* Added go docs for NewTestingLogger
* Resolving conflicts
* Stop inserting hashed plugin keys
* Address comments
* Remove else statement
* Address comments
* Add comment
* Update as per comments
* Refactor as per comments
* Update plugin_key_value_store.go
minor comment tweaks
* fail on non StatusNotFound for original key query
* idiomatic error handling, and do not clean up if insert fails
* guard app plugins with mutex
Shutting down the app could race with a goroutine that uses the plugins environment, since we shut down the plugins first before cleaning up goroutines.
* fix go vet issues
* Adding blank request context to plugin hooks for future use.
* Rename RequestContext to Context
* Adding context to ServeHTTP and ExecuteCommand
* Fixing import cycle in test.
* allow `Wait()`ing on the supervisor
In the event the plugin supervisor shuts down a plugin for crashing too
many times, the new `Wait()` interface allows the `ActivatePlugin` to
accept a callback function to trigger when `supervisor.Wait()` returns.
If the supervisor shuts down normally, this callback is invoked with
a nil error, otherwise any error reported by the supervisor is passed
along.
* improve plugin activation/deactivation logic
Avoid triggering activation of previously failed-to-start plugins just
becase something in the configuration changed. Now, intelligently
compare the global enable bit as well as the each individual plugin's
enabled bit.
* expose store to manipulate PluginStatuses
* expose API to fetch plugin statuses
* keep track of whether or not plugin sandboxing is supported
* transition plugin statuses
* restore error on plugin activation if already active
* don't initialize test plugins until successfully loaded
* emit websocket events when plugin statuses change
* skip pruning if already initialized
* MM-8622: maintain plugin statuses in memory
Switch away from persisting plugin statuses to the database, and
maintain in memory instead. This will be followed by a cluster interface
to query the in-memory status of plugin statuses from all cluster nodes.
At the same time, rename `cluster_discovery_id` on the `PluginStatus`
model object to `cluster_id`.
* MM-8622: aggregate plugin statuses across cluster
* fetch cluster plugin statuses when emitting websocket notification
* address unit test fixes after rebasing
* relax (poor) racey unit test re: supervisor.Wait()
* make store-mocks
