Fixes#30851.
Since there can be multiple messages being edited at once, with a
possibility for them to be out of the DOM, we use
compose_state.get_last_focused_compose_type_input to determine which
message edit to toggle the preview mode on.
We keep the compose box preview toggle behaviour more or less the same,
where we do not check whether the compose box was last focused or not if
there are no eligible message edit boxes and the compose box is open.
This behaviour works well for compose box since there's always one
possible compose box at a time compared to multiple message edit boxes.
One thing to note behaviour wise is that if you have 1 message edit
box open and you shift the focus to compose box, and then close the
compose box, `Alt + P` will not work until the focus is brought back again to
the message edit box.
We want to able to use the clear_preview_area in other places than
click_handlers and that's why it has been moved to message_edit.js.
The name was chosen to be consistent with a similar function in
compose.js.
We want to use render_and_show_preview in message_edit, but that will
cause a dependency cycle because of message_events. In order to avoid
the dependency cycle, the function has been moved to compose_ui and
relevant types have been added.
This commit does not audit the existing function for improvements, just
moves it around and adds types.
The `msg` and `result` are same as the other files (e.g.
activity.test.js) with success responses.
This adds the `LONG_HOVER_DELAY` delay to the user card tooltips, which
prevent the tooltips from becoming distracting when the user is just
navigating the popover.
When the user card is opened via the keyboard shortcut `U`, the initial
focus should be on the first popover menu option, rather than the copy
buttons or the custom profile field links which can be distracting due
to the presence of tooltips on them.
Fixes part of #31027.
Changing the button text of one-time modals to introduce the recent and
inbox view to say "Got it" rather than "Continue" helps reinforce the
vocabulary we're using "Got it" = dismiss one-time content forever.
This is a followup commit to #30824.
Fixes
https://chat.zulip.org/#narrow/stream/9-issues/topic/.F0.9F.8E.AF.20help.20not.20updated.20for.20new.20preferences.20sections/near/1899330.
git grep in the help directory will still show up results for `Advanced`
but those are mostly used in either plain english language instructions
or refer to some other part of the app. At the point of writing this
commit, there were 6 results for `Advanced`, 4 of which were plain
english, 1 refers to the Advanced settings in chrome for desktop
notifications, and another one refers to `Advanced configurations` for
channel settings.
The previous logic did not properly account for the fact that in muted
channels, unmuted/followed topics are sorted above normal topics.
Rather than trying to reproduce that logic, we just call the function
for computing the topics in the left sidebar, and pull off the top
topic.
If the email subject is something like `Fwd:`, it gets stripped to an
empty string, activating the "(no topic)" override. This however leads
to failure if the organization enables the setting forcing every message
to have a topic. Such emails should still go through, so we should just
change the topic value used.
Earlier the focus was not being set to the modal when it was opened.
This was because `is_in_focus` checks if `modals.any_active` which
checks `$(".micromodal").hasClass("modal--open")`.
But when `is_in_focus` is called
`$(".micromodal").hasClass("modal--opening")`.
This commit fixes the issue by creating a function
`is_active_or_animating` which checks if the modal has the class
`modal--open` or `modal--opening`.
Earlier, when the user deactivation modal was open, the focus was not
set on the modal. This commit fixes that by setting
`focus_submit_on_open` to `true` in the modal options.
This allows clients to potentially lay out the thumbnails more
intelligently, or to provide a better "progressive-load" experience
when enlarging the thumbnail.
- Make page title more general.
- Document new option for animated image previews.
- Split instructions for showing different preview types.
- Update intro.
Previously animated images were automatically played in the
message feed of the web app.
Now that we have still thumbnails available for them, we can add a new
personal setting, "web_animate_image_previews", which controls how the
animated images would be played in the web app message feed -- always
played, on hover, or only in the image viewer.
Fixes#31016.
This commit increases right padding for the th elements in recent
views table to make sure that the sorting arrow is visible correctly
for various screen width and font sizes. This change is similar to
what we did for tables in settings UI.
There was a bug when stream_id field was being passed as NaN when user
tried to edit topic of a message if the user did not had permission
to edit the channel.
This commit fixes it to not pass stream_id field in such cases because
channel cannot be changed.
Fixes#30992.
When a user changed an email, we updated the compose and narrow state
before updating that person's email in `people_dict`. This raised a
warning in `get_direct_message_permission_hints` when using
`emails_strings_to_user_ids_string` since that person's email wasn't
updated in `people_dict` by then.
We just changed the order of update so that the email is updated in
`people_dict` before we update compose and narrow state.
In 'fetch_initial_state_data' we were doing one database query
per announcement stream.
This commit updates the logic to prefetch those streams using
select_related hence avoiding the extra db queries.
Fixes#28909.
The libvips cache is 100MB, 100 operations, or 100 files, whichever is
less. A single Django process or worker is extremely unlikely to ever
see the same image twice, much less within those timeframes.
Disable the cache, since it is mostly useless memory usage for our use
case.
After navigating to a `near` narrow, user is likely to press back
to go back to previous selected message. To support this, we have
to select `then_select_id` without rerender, if we can.
Tested by using breakpoints with `near` links to check if the
correct code is being executed.
There's no need for sharding, but this allows one to spend a bit of
extra memory to reduce image-processing latency when bursts of images
are uploaded at once.
This moves the cache key for parse_media_data to being the "canonical"
image URL, not the displayed `img src`. We may not even have the `img
src`, if the thumbnail has not yet been generated.