Additionally, migrate existing EditMessageTest to use this helper
method, with the side effect of migrating the tested flow from a
/json/messages URL to a /api/v1/messages URL.
This code path isn't currently reached for the logged out "spectator"
view, but it will be in upcoming commits. This makes sure that case
is handled properly.
This change ensures that every call to find the stream name or topic
in the composebox is calling compose_state functions instead of directly
taking a value from an HTML element specified by a classname.
This has better code readability and abstracts away the classname,
and also makes it easier to change which classname we use (which
will be happening in an upcoming change as part of switching the
stream name field to a dropdown).
Note that this change means that the stream name will always be
trimmed, whereas before it was whatever the user had written even
if it had trailing whitespace. This shouldn't be an issue as far
as we (me + Tim) can tell, and also it will become irrelevant as
soon as the dropdown changes land.
This code is equivalent, because the keep_leading_whitespace parameter
of get_or_set was never used for the stream name.
This addresses an open TODO and makes the code more readable.
Updates `compute_narrow_title` to cover the two cases for views
with undefined filters: "All messages" and "Recent conversations".
Also, moves the call to `notifications.redraw_title` from
`set_narrow_title` and to `update_narrow_title`.
Creates a new function, `compute_narrow_title`, that returns a
string for the narrow title based on the current filter, so that
`update_narrow_title` can call that function, and then use the
returned string to set the narrow title.
Adds a node test for the new `compute_narrow_title` function for
cases that differ from the title generated by `filter.get_title`.
The authenticate_by_username limit of 5 attempts per 30 minutes can get
annoying in some cases where the user really forgot their password and
should be allowed to keep trying with admin approvial - so we should
document the command that allows unblocking them.
This is implemented by reusing the stream_privacy handlebars helper.
We can't use it in the public stream case, because the design of the
invite users modal uses no space between the `#` and the stream name.
Fixes part of #23496.
Instead of removing all compose banners, we now remove only warnings
and errors -- this makes no difference right now, but this change
will allow us to add other compose banners that are cleared at
different times.
This fixes a visual regression in newer Electron builds (without
regressing Firefox) for which I still don't entirely know the root
cause, where extra "borders" were being applied to messages in both
streams and PMs. Applying a negative "spread radius" to the box-shadow
properties of these elements, and moving that pixel to the "horizontal
shadow" aspect of the property (which is used to create the left-side
"ruler" effect), restores the expected look and feel.
Tested in qutebrowser (Chromium 87-based), Electron v18+v19, and Firefox
107.
Refs (and should unblock) zulip/zulip-desktop#1251
If new stream is created as one of the two private options, the view stream button was not visible
immediately, but does appear after a page refresh. This bug turns out to be caused by the wrong
selector being used in the code intended to make the button visible.
Fixes: #22556.
Refactored (moved) the code for preserving the cursor's initial logical
position from `quote_and_reply()` in `compose_actions.js` which calls
`replace_syntax()` directly into `replace_syntax()` in `compose_ui.js`.
This ensures that anytime text in a textarea is replaced, the original
cursor position is always restored.
Earlier, this was needed to be done separately, and missing that would
lead to bugs with the cursor unexpectedly jumping on replacement.
Fixes: #23863.
This margin doesn't make sense in a wider view, and is actually
a hinderance to the layout on narrower screens, and removing it still
keeps more than enough separation from the buttons on the right.
This commit removes get_editability function and we can instead
use is_content_editable and is_topic_editable/can_move_message
functions to check the permissions. We can remove get_editability
since there is no longer a sort order to the possibilities and
it is not necessary that if a user has permission to edit content
then they can edit the topics also.
So, this commit adds is_message_editable_ignoring_permissions
check to is_content_editable and also modifies the tests accordingly.
We separate the message editing and message moving settings
into different sections.
We do not change the help-links for the section headings and keep
both of them same linked to "Configure message eiditng and deletion"
as we would anyway be changing or spliting this page and can update
the links at that time.
This commit renames "can_edit_topic_of_any_message" function
in models.py to "can_move_messages_to_another_topic" and
"user_can_edit_topic_of_any_message" function in settings_data.js
to "user_can_move_messages_to_another_topic".
This change is done since topic editing permission does not
depend on message sender now and messages are considered same
irrespective of whether the user who is editing the topic had sent
the message or not. This also makes the naming consistent with
what we use for the label of this setting in webapp and how we
describe this action in help documentation.
This commit updates the labels for "edit_topic_policy" and
"move_messages_between_streams_policy" settings.
The label for topic editing setting is changed to
"Who can move messages to another topic" since the topic
edit permissions do not depend on message sender now and
messages are considered same irrespective of whether the
message was sent by the user editing the topic or by
someone else.
The label for stream editing setting is changed to
"Who can move messages to another stream".
These changes also makes the labels consistent with how these
actions are described in help center documentation.
This commit changes the topic edit permssions in webapp to not depend
whether the user editing the message had sent the message or it was sent
by someone else. Backend changes were done already in previous commits.
Previously, we always allowed topic edits when the user themseleves had
sent the message not considering the edit_topic_policy and the 3-day time
limit. But now we consider all messages as same and editing is allowed only
according to edit_topic_policy setting and the time limit of 3 days in
addition for users who are not admins or moderators.
This commit changes the topic edit permssions to not depend whether the user
editing the message had sent the message or it was sent by someone else.
We only do backend changes in this commit and frontend changes will be done
in further commits.
Previously, we always allowed topic edits when the user themseleves had
sent the message not considering the edit_topic_policy and the 3-day time
limit. But now we consider all messages as same and editing is allowed only
according to edit_topic_policy setting and the time limit of 3 days in
addition for users who are not admins or moderators.
We now allow editing stream and topic of message even if
allow_message_editing if False using webapp given that is
allowed as per other topic and stream edit specific settings.
Fixes a part of #21739.
We change the topic and stream edit permssions to not depend on
allow_message_editing setting in the API and are allowed even
if allow_message_editing is set to False based on other settings
like edit_topic_policy and can_move_message_between_streams.
Fixes a part of #21739.
All instances of `choice` have been replaced with `option` in the UI
for editing a poll. In code, `option` was already being used. This
ensures the same terminology is used across the code, the UI and the
related help center article.
"is_pm_recipient" is not supposed to be called with an arbitrary
recipient object which might have "to_user_ids" being undefined.
Since this helper is only used with focused_recipient in
compose_fade_helper, we move it there.
Note that the helper is no longer separately tested. It is now covered
by the test case of "compose_fade_helper.would_receive_message".
(See
5e74a8d0cc/static/js/compose.js (L156-L162))
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
This avoids "@typescript-eslint/restrict-template-expressions" when we
convert util.js to util.ts.
Note that prettier would otherwise split the first template literal
expression in half if we don't start a newline.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
This uses the linkifier index among the list of linkifiers in the
replacement as the priority to order the replacement order for
patterns in the topic. This avoids having multiple overlapping matches
that each produce a link.
The linkifier with the lowest id will be prioritized when its pattern
overlaps with another. Linkifiers are prioritized over raw URLs.
Note that the same algorithm is used for local echoing and the
backend markdown processor.
Fixes#23715.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
This explicitly enforces ordering on the linkifiers. This is useful when
there are overlapping linkifier patterns that matches the same text. In
our current linkifier implementation, this order affects how the
patterns are handled in the markdown processor, with the earlier ones
being prioritized.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
The same pattern being matched multiple times in a topic cannot be
properly ordered using topic_name.find(match_text) and etc. when there
are multiple matches of the same pattern in the topic.
Signed-off-by: Zixuan James Li <p359101898@gmail.com>
Josh Klar
evykassirer
Tim Abbott
Lauryn Menard
Mateusz Mandera
Anders Kaseorg
Alya Abbott
SameepAher
Josh Klar
Rishant Rokaha
N-Shar-ma
Rohitt Vashishtha
Sahil Batra
Ujjawal Modi
Yuri Sulyma
Zixuan James Li