All integrations currently use their name as their client name as
well. Removed the redundancy of declaring the same name twice, while
retaining the support for client names.
This commit allows users to be assigned to custom groups when
inviting them to join Zulip, similar to how channels are handled.
The implementation follows a similar pattern for adding pills,
ensuring consistency, as user groups and channels are parallel
in nature.
Fixes#24365.
After #28478 we no longer change the name when deactivating a channel,
so this is particularly needed - now there's always a "collision" on
new_name.
Also fixes a second detail - the name should be matched
case-insensitively.
This timeout needs to be short enough that we don't drop the RabbitMQ
connection. Also drop the offending message (by returning with no
further exception) so we don't hit a head-of-queue failure situation.
Ideally, the parser would just be lightning-fast, so this would never
happen.
We do not allow realms on a limited plan to create groups. This commit
adds a banner to upgrade to the can_create_groups org setting and
disables editing it on the backend.
This commit also adds a new method called
`disable_group_permission_setting` to easily disable similar settings.
Since "Nobody" and other user-defined empty groups has no
members or subgroups, we need a UserGroup query.
Since we are doing UserGroup query, updated the code to
check whether the setting is set to anonymous group or
not just after doing UserGroup query and not in membership
queries.
Also, added tests to check query count when setting is set
to anonymous group.
We give the new users some messages in their feed.
Earlier, we were including upto 1000 messages which were sent
within the last 12 weeks.
For realms with low-traffic it results in very few messages
being included in the new user's feed.
This commit removes the 12 week limit.
Now, we simply include upto 1000 recent messages.
This commit extends the test for the `/api/v1/register` endpoint
for spectator users by adding an assertion to check the number
of database queries.
Fixes: #32409
This commit adds a custom rule to check python files
and raise lint error if they have transaction.atomic used
without any argument or savepoint=True is used.
It helps to avoid creating unintended savepoints in the future.
We create savepoint at a couple of places in backend tests
to avoid rollback due to error raised within test's transaction.
This commit explicitly adds 'savepoint=True' at those places.
This commit adds `durable=True` to the short independent db transaction
in `UserGroupRaceConditionTestCase.tearDown` method.
This is as a part of our plan to explicitly mark all the
transaction.atomic calls with either 'savepoint=False' or
'durable=True' as required.
It helps to avoid creating unintended savepoints in the future.
This is as a part of our plan to explicitly mark all the
transaction.atomic calls with either 'savepoint=False' or
'durable=True' as required.
This commit removes the 'email_token' field from the
'Stream' model, introduces a new model 'ChannelEmailAddress',
and backfills it.
This is a prep work towards our plan to generate unique
channel emails for different users, which can be used
to check post permissions at message send time.
- Removed the extra screenshot.
- Updated the fixture used to generate the screenshot to use a more
suitable topic name and commit message description.
Removed `invite_to_realm_policy` property, as the permission
to invite other users to the organization is now controlled by
`can_invite_users_group` setting.
This commit fixes the code to pass AnonymousSettingGroupDict
value instead of group ID when the setting is set to anonymous
groups in never subscribred streams subscription dicts.
Updates text of generate-webhook-url-basic.md to replace the use of
the "Create the channel" step.
Removes {!create-channel.md!} from all the integrations docs that
also include {!generate-webhook-url-basic.md!}.
Fixes#32269.
This change improves error handling in `do_import_realm` by replacing
the use of a generic Exception with CommandError. The updated approach
provides clearer, user-friendly error messages when there is a version
mismatch between the exported data and the Zulip server.
Fixes#32292.
This commit updates code to optimize code for computing stream
objects to be sent with stream creation event and in response
for 'GET /streams/{stream_id}' endpoint by optimizing number
of queries while computing values for group permission settings.
This change does not benefit much currently as we only have one
stream group permission setting, but is important before we add
more stream permission settings.
There are a couple of places left where we can still optimize
the code and that would be done in further commits.
This commit adds a new function to compute setting group
values for a list of streams, so we can avoid having duplicate
code for computing setting group IDs from streams.
This commit updates code to compute the values for group permission
settings in send_stream_creation_events_for_previously_inaccessible_streams
only when we need to send the events. This helps us in avoiding
unnecessary DB queries.
There is no need to call "locals()" inside the loop for
stream permission settings. It should just be called once
to get values of all the settings passed to the endpoint.
It was fine as we only had one group permission setting
for stream, but is a good fix before we add more settings.
This commit refactors code to create subscription dicts so
that we can efficiently compute can_remove_subscribers_group
setting when we add support to set the setting to anonymous
groups.
The initial implementation for this feature used the issue UUID
to create the topics, as some of the webhook events only provide
this and it is the only way to group sub-issues with parents.
However, if we sacrifice this and allow sub-issues to be
their own topic then we can avoid using the UUID as the topic name.
I think the quality of life improvement to topics here warrants this change,
as UUID topic names can be come unruly very quickly.
Part of #23118.
The in:home narrow filter is used to filter messages that appear
in the home view, i.e., messages that are not muted. Conversely,
`-in:home` should filter messages that are not in the home view,
i.e., muted messages. However, `-in:home` did not work as expected
because this filter lacked negation support, unlike similar code
paths.
This commit adds negation support for the in:home filter.
For more information, see:
<https://chat.zulip.org/#narrow/channel/378-api-design/topic/mark.20muted-topic.20messages.20as.20read/near/1980534>.
orjson’s use of Unicode is more consistent with what we get from
Transifex. (We could alternatively use json’s ensure_ascii=False
flag.)
Signed-off-by: Anders Kaseorg <anders@zulip.com>
