With the schema version available, a job can query for
the cluster info to confirm whether or not all nodes
in a cluster are upgraded to the same version or not.
This will help it in determining whether to start
the job or not.
https://mattermost.atlassian.net/browse/MM-53228
```release-note
NONE
```
* remove feature flag managed plugins
* remove unneeded plugin blocklist
* remove unnecessary wrappers
* documentation and logging improvements
* avoid use of global logger
* leverage wrapped loggers (e.g. consistently log `plugin_id`)
* promote some logs from `Debug` to `Info` for better visibility.
* extract installPluginToFilestore
* rename some variables for consistency / clarity
* make generated
If the store fails to initialize (e.g. run a migration), it would `log.Fatal` and then `os.Exit`. Unfortunately, this trips up `TestMain`, which happily keeps running tests, now guaranteed to fail.
Avoid this by instead returning an error from the store initialization, handling appropriately at the layer above.
After some analysis of vacuuming statistics of our large
Postgres customers, it was observed that the default vacuuming
thresholds for some tables are not optimal enough for our workload.
We need more frequent vacuuming for some tables. The analysis
showed that the top 4 tables that needed rectification
were posts, preferences, threadmemberships and fileinfo.
So we reduce the threshold by half for those tables, triggering
autovacuum and autoanalyze doubly often.
```release-note
NONE
```
We missed this out last time. It's possible in an HA
scenario for a second pod to start later while the other
job is in-progress. In that case, it would schedule
two jobs.
https://mattermost.atlassian.net/browse/MM-53747
```release-note
NONE
```
Bifrost now encodes all image paths. Due to this
one-way translation, we need to encode all the older
image paths as well.
After this is done, we can remove the double-lookup.
https://mattermost.atlassian.net/browse/MM-53747
```release-note
NONE
```
Co-authored-by: Mattermost Build <build@mattermost.com>
* Revert "POC: Cross-team recent search (#20027)"
This reverts commit aa59c28b04, preserving
a few code tidyings unrelated to the original PR.
* Revert "Add feature flag for command palette (#20011)"
This reverts commit c78c5ce3f3.
Now that we have identified the true fix
from Bifrost side, we will slowly revert
the changes that have happened.
Starting with this first fix, we will revert
to writing out non-encoded paths while still
doing the double read. This will break uploading
files with + for a short while, and then
the new Bifrost fix should fix that as well.
https://mattermost.atlassian.net/browse/MM-53747
```release-note
NONE
```
* drop public submodule replace directive
During development of the public/ submodule, a replace directive remained that was both unnecessary and harmful as discussed in https://community.mattermost.com/private-core/pl/w77sh7igwpfb9ecj5o4jjjbbyo.
Remove that, and bump the explicit dependency (even though we use go.work) to v0.0.6 so the import paths match (e.g. `mattermost` vs `mattermost-server`).
* make modules-tidy
* consistently setup-go-work
* fix build-api-spec
* fix mmctl test template
* fix workflow copy/paste typo
We do a check to see if a non-encoded path is present
or not, and in that case, choose not to encode it.
This accrues an additional StatFile call. But after
we have encoded all paths to the new style, we will
get rid of this.
https://mattermost.atlassian.net/browse/MM-53709
```release-note
NONE
```
* add base for calling the endpoint
* add endpoint and handler
* update store and layers
* call the endpoint
* align types
* update app layers
* generate mocks
* complete handler
* finish store query
* add todos
* add ui for member count
* add selector
* add a todo
* add cache layer
* optimize calls in FE
* handle invalidation of the cache
* fix go style
* fix test
* use existing channel layer count
* fix import error
* delete unnecessary code
* write tests for channel cache layer
* fix testname
* fix mocks
* fix cache layer test
* fix a test
* really fix the test
* write more tests for server
* address PR comments
* remove comment
* rename more_channels to browse_channels
* fix style
* update snapshot
* add translations
* Revert "add translations"
This reverts commit 56476a5dabe357703ef02be9a38b3e88f5c8b1e7.
* add only related translations
* address PR review points
* add test
* fix test
---------
Co-authored-by: Mattermost Build <build@mattermost.com>
There was already a cache present for emoji names. But we weren't using
it for the GetMultipleByName method. Now we implement that method
to look up the cache for every emoji name passed.
Secondly, a bigger problem was that we were making the DB call for system
emojis as well. Since system emojis aren't stored in the DB, it would
fall through the cache layer and always make a redundant DB call. In the
profiles, this should up as taking 16% of the total time to serve
a getPostsForChannel API endpoint.
We fix this by filtering the emojis to only custom emojis
before making the call.
https://mattermost.atlassian.net/browse/MM-53669
```release-note
NONE
```
* [MM-53192] Patch full name leak in Insights team_members API
* Update server/channels/app/team.go
Co-authored-by: Ibrahim Serdar Acikgoz <serdaracikgoz86@gmail.com>
---------
Co-authored-by: Ibrahim Serdar Acikgoz <serdaracikgoz86@gmail.com>
Sometimes a broad distinction of just a channelID or a userID
is not enough to efficiently send a websocket event to users.
In several cases, depending on the user and channel, we might
need to modify the message. Therefore, we introduce the
concept of a channel hook that will get executed if the scope
is set to a channel. This hook can be populated at the app layer
to perform any application specific logic to the event.
Care must be taken to avoid race conditions as the passed event
is not deep copied. It is left to the user to treat it carefully.
For this issue, the main problem was that since we don't know
which users have permissions to which channels, we had to go through
_all_ members of a channel to figure that out. This was redundant
since a large portion of those users might not even be connected
at that time.
We solve this with the channel hook where we push this check
to be performed later while actually sending the event. This
reduces the computation to be done only for _connected_ users
rather than _all_ users of a channel.
The next iteration of this should be to use websocket subscriptions
to monitor exactly which users are on that channel to even
trim down that list. That is a larger initiative to be taken later.
Tested locally with a channel of 50 users. Here are rough results:
```
With PR:
patchPost 97ms
createPost 90ms
Master:
patchPost 306ms
createPost - 298ms
```
https://mattermost.atlassian.net/browse/MM-42810
```release-note
Improve performance while sending messages
with permalinks to channels with large number of users.
```
* add teams to allowed frame-ancestors
* fix unit tests
* set SameSite attribute for session cookie
* further restrict ancestors
* skip landing page if in iframe
* Only set cookie SameSite=None if embedded in iframe
* don't set MMEMBED cookie on landing page (check only)
* fully parse MMEMBED cookie
* add comment
* more comments
---------
Co-authored-by: Mattermost Build <build@mattermost.com>
