* revert 4e5f6fcfbc
This reverts the configuration setting to disable plugin health checks, preferring instead to retain this functionality for advanced cases. In part, this was driven by the discovery that the health checks were failing on Windows, though that will be addressed separately.
Fixes: MM-16378
* simplify interval handling/logging
Also shutdown the job when plugins are altogether disabled.
* MM-17071 Extract and add tests for handling of out of channel mentions
* Rename checkForOutOfChannelMentions back to sendOutOfChannelMentions
* Log errors returned by sendOutOfChannelMentions
* add docker-compose for ci
* docker-compose logs does not have the option to set the dc file
* leverage docker-compose extends
This moves the developer `docker-compose.yml` to the root, extending the `build/docker-compose.yml` and overriding the `container_name` to preserve compatibility with the older, non-docker-compose setup.
Note that this required downgrading to docker-compose 2.4's file format (still supported by the newer tooling) due to the long and frustrating converstaion at https://github.com/moby/moby/issues/31101.
* remove -f docker-compose-ci.yml references
* replace dockerhost with localhost
* remove uneeded setup-max build step (no more dockerhost)
* changes as recommended by @cpanato
* make clean-docker with docker-compose
* added ports to docker-compose.yml (needed for osx). ignore error for ldapadd (when already exists)
* add clean-old-docker to legacy.mk
* docker-compose stop instead of down for `make stop-docker`
* Time Layer for measure the Store methods calls time
* Fixing build
* Fixing a formating problem
* Fixing tests
* generating store mocks
* Fixing build
* Updating generated timer layer
* Updating timer layer to the last store interface
* Updating timer layer
* Generating time layer
* MM-16479: panic on io.Copy error in test
I'm unable to reproduce the original report on MM-16479, but in a bid to help track down a potential failure in the future, I've added a panic on err returned from `io.Copy` within the plugin, as this seems like the only path on which the hook could return successfully but fail to show the updated contents.
* assert number of copied bytes
* log n on io.Copy error too
* changes for channel and post import for corp slack
* add fix for missing dms in import
* add suggestions from PR review
* Apply suggestions from code review
Co-Authored-By: George Goldberg <george@gberg.me>
* fix gofmt
* MM-17383: Update query to include users who are not members of any groups.
* MM-17383: Fixes govet complaint.
* MM-17383: Sorts by username.
* MM-17383: Removes accidental staging.
* Filtered incoming webhooks for users wihtout PERMISSION_MANAGE_OTHERS_INCOMING_WEBHOOKS
* Filtered outgoing webhooks for users without PERMISSION_MANAGE_OTHERS_OUTGOING_WEBHOOKS
* Refactored GetOutgoingByTeamByUser to its own method in app and store
* Fixed paging condition for outgoing webhooks in store
* Separated test cases into separate t.run in WebhookStore
* Improved unit test. PR Feedback
* Filtered outgoing webhooks by channel for users without PERMISSION_MANAGE_OTHERS
* Filtered getting full list of outgoing webhooks for users without PERMISSION_MANAGE_OTHERS
* Added missing signature for GetOutgoingWebhooksPage in app
* Expanded permissions in test to SYSTEM_USER_ROLE
* Filtered getting full list of incoming webhooks for users without PERMISSION_MANAGE_OTHERS
* Removed unnecessary sq.and operator
Revert https://mattermost.atlassian.net/browse/MM-14559. The tests were refactored, but all the hooks were be run as:
_, ret := hooks.MessageWillBePosted(nil, nil)
and the plugins were panicking, but then nothing was actually failing because this hook doesn't support returning a "proper" error. Good news is that the old tests are still passing, so nothing has regressed in the toolkit, but thinking we should revisit this holistically instead of trying to fix in place.
