* FIX: revert chat-integration move to discourse-automation
this reverses the changes made in 57b460737c
without deleting the migration.
* Revert "DEV: removing create post for category and tag changes setting (#210)"
This reverts commit 35925c4dac.
* DEV: merge both chat-integration rules and automation rules
* FEATURE: add topic tags changed trigger to chat integration
* FEATURE: add placeholder for reply to topic trigger
add description on how to use the placeholder
* DEV: Move slack message creation to provider
Add tests to new method
* FEATURE: add ${URL} to placeholder replacements and added tags link
If triggered when a topic tag is changed, message behavior will follow what user defined in message.
* DEV: Update tests with tags
* DEV: add post to topic for testing
* DEV: update test strings
* DEV: add early return for topic tags changed trigger
* DEV: move early return to use try/catch
* DEV: update `create_slack_message` to not send a tuple of values
* DEV: refactor method to be more readable
* FEATURE: add `${ADDED_AND_REMOVED}` for default texts
* DEV: Update typo in test
* DEV: Add tests to check when if `create_slack_message` raises an error
* DEV: Remove the `tag_added` from chat-integration filter
Added migration to handle the migration of the `tag_added` filter from the chat-integration plugin.
Only removed the logic from the plugin, data removal will happen in a future PR
* DEV: lint migration file
* DEV: update chat-integration to not show "tag_added" rules
* DEV: update added and missing tags logic
* DEV: update context variable name
* DEV: update migration to include `begin/rescue` block and added a list with available filters
We're promoting problem checks to first class citizens in core. This migrates the problem check to the new API.
In the process of adding tests for this check, I discovered what seems like a mistake that likely means this check never worked until now. (See inline comment.)
We're updating core to change TL based access settings to be group based. This requires some updates of tests to work correctly. (The existing test setup gives false positives.)
Previously we were using the `ts` of the previous message we sent to the thread. While this did work under some situations, it's not recommended in the Slack API docs, and can lead to some unexpected behavior (e.g. when one of the threaded messages is deleted).
This commit updates our logic to always use Slack's returned `thread_ts`, which represents the thread's parent.
This patch concerns the Telegram integration. Currently, we always try
to process commands when we receive a hook from Telegram. To do so we
rely on the `text` parameters from a Telegram message but the
API documentation tells us this parameters is actually optional. It
means sometimes it’s not present in the payload we receive but we still
try to access it resulting in a crash.
This patch addresses the issue by simply returning early from the
`#process_command` method when `text` is missing from the payload since
we don’t have anything to process then.
This commit centralizes 'display name' generation into a helper, and updates it to respect core preferences for `enable_names` and `prioritize_username_in_ux`.
We were using the internal UID of the channel in the hint, which would then fail to match any 'channel' configs in Discourse (which use the human-friendly `#channel` format). This was causing automatic threading for transcripts to fail.
When notifications about a topic are posted to multiple slack channels, and the Discourse channels are configured to "thread" the notifications, each channel will have a different thread_id. Previously we were only storing a single slack thread id per Discourse topic. This commit fixes that logic, so that threads in different channels are tracked separately.
Once configured, this adds a new item to the context menu of slack messages. When clicked, the menu item will generate a transcript and present the user with a custom "Post to Discourse" modal. This provides the same functionality as the existing slash-command interface, but is much more user friendly.
- Fix multi-line code blocks
- Add strikethrough support
- Fix HTML entities inside code blocks
- Do not process formatting inside code blocks
- Ensure links are never created with no URL
- Replace - with _ in emoji names
- Stop using `break` in a block - it doesn't work. The failure here was hidden because it was within a `defer` block, so would not cause a server error
- Refactor the error handling so that the error is passed back to Slack and displayed to the user
- Return specific error messages for user / message-history / message errors
- Tidy up the SlackCommandController to make all non-requestable methods private
- Add a test to ensure error messages are passed correctly to Slack
The `site_setting_changed` event passes the setting name as a symbol, not a string. This commit also removes the backwards-compatible `site_setting_saved` event, which was deprecated in 2.3.0.
When creating a new Discourse post from slack with the `post` feature, record the
slack `ts` thread ID for the resulting topic post using an HTML comment to pass
the `ts` through.
When notifying slack of new Discourse posts, record the slack `ts` thread ID in
the post's topic if it has not yet been recorded. (Normally, this will be done
for the topic post, except where notifications are being posted for old topics
before this feature was created.)
Add a new rule filter `thread` which posts threaded responses to slack if there
is a `ts` recorded for the post topic.
Modify the `trigger_notifications` interface to enable other integrations to
implement similar functionality.
Present the `thread` rule in the help text and admin UI only for the slack
providers.
https://meta.discourse.org/t/optionally-threading-posts-to-parent-topic-in-slack-integration/150759
* adds groupme configuration and relative
* first pass at groupme provider
* add group id site setting w/ english translation
* rework to use bots groupme api intstead of developer, no access tokens reqd
also should catch 404 response codes
* add strings to locale yml files for i18n
* better error handling for multi-bot case
* add channel param for separate Groupme instances, include name in errorbody
* bugfix for multi bot msg forwarding
this gives us the ability to treat diff groupme instances like slack channels, lots of people use them this way for better or worse. use case is certain category posts only go to a particular GM instance
* add spec for groupme provider
* fix channel param constraint
* specify channels by groupme name, not bot api token
* fix some linting issues w/ spacing
* newline and trailing space lint fixes
* Use Slack display names instead of the "name" field
The "name" field is the left-hand side of the email address in many
cases. This information is not otherwise available so we shouldn't
expose it in forum posts.
* Fall back to real_name, add comments
* Store users as a hash
This should avoid a lot of O(n) lookups
* Format user mentions with the correct name also
* Fix the tests (first try)
* Fix the tests (second try); add a test for user mentions
* Fix the tests (third try)
* Empty commit to trigger Travis
* Fix the tests (fourth try)
* Fix the tests (fifth try)
* Change spaces to underscores
* Updates per feedback
Gabriel Grubba
Benoit
Ted Johansson
Jarek Radosz
Mark VanLandingham
David Taylor
Martin Lim
Loïc Guitaut
Leonardo Mosquera
Mark VanLandingham
Mathieu Trudel-Lapierre
jjaffeux
Ahmed Gagan
Robin Ward
Michael K Johnson
Tosin Sonuyi
James Nylen
Bianca Nenciu
Guo Xiang Tan