We had the `3px 0 1px` padding before migration to use grid,
then I switched it to use `4px 0 1px` since we were planning to
use blue box border which seemed to have helped that case.
Since we switched to using outline for blue box, it makes sense
to just use equal padding.
Backend incorrectly renders @everyone and @stream as `user-mention`
while they are clearly `group_mention`. To fix this, we use
`data-user-id` property of @everyone and @stream which are
set to `*` for them.
The use of named areas with `grid-area` make it unnecessary to
declare `grid-row` or `grid-column` values. (Note also that
grid areas must not be presented in quotation marks.)
Additionally, because `.unread_marker` is no longer placed on the
grid by itself (i.e., it always accompanies `.date_unread_marker`
or `.message_unread_marker`), it does not need any manual grid
placement, `grid-area` or otherwise.
This change enables the unread marker to participate as a grid item,
rather than the product of various absolute/relative positioning
hacks. The intention is to therefore prevent the blue active-message
box from disappearing on browsers that have zoomed out (~80% zoom).
With grid in place, this also makes for a more robust presentation
of each message row, and named grid areas should make it possible to
modify and extend the grid into the future.
Finally, this change removes styles that are no longer necessary in
the context of CSS Grid.
Fixes#25413.
The old code was disabling the save button wrong by using
`.addClass("disabled")` instead of `prop()`.
Added tooltip for the disabled save button as per issue #25413 and changed
"Times up!" color to red.
The textbox readonly logic was changed to no longer becoming readonly. Reason
being there are edge cases involving the compose buttons such that simply
marking the textbox as readonly is not sufficient.
E.g. using the compose buttons after readonly still modifies the content.
One solution might be to just hide the compose buttons visually. However, there
are edge cases for that too. If preview mode was previously active, then
perhaps that state needs to be reverted. If any modal is open, such as the emoji
picker, then that needs to be closed. Solving these edge cases doesn't
improve the user experience. Keeping the textbox editable allows an easier way
for user to copy the text and don't have weird cases.
Zulip's select widgets have a 30px height; this comes from Bootstrap
but is also generally nice for visual consistency.
In modals, we use a 15px font-size, instead of the 14px used in the
rest of the app, and in that context, the 4px vertical padding plus
30px fixed height resulted in the text not being vertically aligned.
Fix this by removing that vertical padding; all of our select elements
with these classes appear to position the text in the center of the
dropdown through other CSS mechanisms.
In #22524, we have updated the compose banner to the new, updated design
while the unread banners have remained the same. When comparing them side
by side, they look rather old and outdated. We should apply the updated
design to the unread banners as well.
Fixes: #25551.
I didn't want to use JS for it since calculating height based on
item height before rendering the dropdown is expensive and after
rendering the dropdown will cause it jump.
`Version` and `Forked from` copy buttons in the "About Zulip" UI
will now change color when hovered over.
This allows the copy button to pop out more when hovering over it.
Fixes part of #23210.
Now that the navbar is a different color, we want it to also
take up the full width of the screen. Because the navbar has
margin, this commit adds a new div. To not have to add a new
div, the scrollbar would need to be moved to the `html`
element, which is a bigger project.
This commit also moves the zulip logo to be left-justified instead
of centered, since it looks strange centered with the full
width navbar.
This looks visually nicer, and I think is necessary to merge it for
now, but I'm doing this as a separate commit so it's easy to revert if
we change our mind.
In move_topic modal, Renamed inline_topic_edit to
move_messages_edit_topic to be more specific selector as
inline_topic_edit is also used in message header.
Additionally, Removed id inline_topic_edit as it was only used
in zulip.css instead used `.inline_topic_edit` selector as that
element also have class with same name inline_topic_edit.
Fixes#24805
Abbrevated DM recipient names in message feeds, Drafts and Scheduled
messages overlays by wrapping recipient names in span with class
private_message_header_name and applying `overflow: hidden` and
`text-overflow: ellipsis` CSS to it.
Fixes#25353
Earlier, the sender info would overlap with the edit compose
area because of another class applying "margin-top" to keep
the sender info properly aligned when the message is not being
edited. However, that "margin-top" had to be changed when the edit
compose is opened.
This commits adds that "margin-top" to sender info to make sure
that it doesn't overlap with the edit compose area.
When there was an unread message below date row, the date
row overlaps with the message header.
This was a result of #23538 not adjusting the message header
z-index along with the z-index of date row.
This is a more natural place for this extra padding than below the
navbar. Padding below navbar was primarily needed to hide the
message content above the sticky header while scrolling the message feed.
We are still able to do with this change without partially hiding the
loading message indicator.
Fixes#25159
Updated mute_or_unmute_topic to take visibility policy as parameter
and set topic visibility_policy to the passed visibility_policy.
In zulip.css, updated CSS to set opacity as per visibility_policy.
Updated click handlers for recipient_row mute/unmute icons to
work as per stream.
Fixes#25124
Previously the typeahead container was being created at the bottom
of `body`, and its width (and `top` and `left`) were being set to
move it to the right position.
Now it sits in the search box container, which gives it the correct
position and width by default. This is better for DOM readability,
and is also better for the new 100% width (which is part of the
search bar redesign) because it can change width more smoothly
with the search bar when the page changes width.
This commit adds custom functionality to the bootstrap typeahead
to allow the typehead to be placed in the search box container
(whereas previously, it could only be appended to `body`).
The logic to avoid first unread message with a date row to not
have an unread marker works purely based on `z-index` and seems
to be working nicely.
This avoids a bug where the background color of the date row
of the first unread message is different from others.
Following important changes are being made here:
* color_class lib is removed since it not used anywhere now.
* We don't need the `dark_background` class since the background
color is already adjusted based on color scheme. So, all
instances of it being used is removed.
At times, it might get confusing for users who are on
slow connections if their messages has not been sent
even after 5s. Including a spinner that will only show
up after 5 seconds has elapsed will keep user informed
about their slow connection.
5s is set as minimum time because showing up a spinner
before than might be distracting for users on normal
connections.
Fixes: #19328.
This commit fixes the alignment of the `search_close` icon on devices
with a width less than $sm_min (576px) by removing a top padding
of 5px and adjusting the position to top 5px from the previous 4px.
The dropdown icon for direct messages is partially visible
on small screens. This was due to removing padding when the
screen size is below a certain size by media query.
The intent of the original code appears to have been to remove top
padding, so make the CSS do just that.
This commit adds width property to CSS for "#inline_topic_edit"
element in zulip.css to set it to 206px, as we will be removing
the bootstrap CSS for it in further commits.
For some reason, browser is treating clicking on the button as
submitting the form, which results in the page getting redirected
to the same page with an additional empty query `?` in the URL.
postcss-preset-env transpiles this back as necessary. (It does a
better job than we did, in fact: we had several four-argument hsl()
calls that should have been hsla().)
Signed-off-by: Anders Kaseorg <anders@zulip.com>
This adds a new search icon which we prefer over the one made
available from bootstrap, and replaces search icons in navbar
search with the Ionic icon.
The previous hide-date CSS class had a semantic meaning of "this
recipient row has the same date as the previous one"; since we're now
having logic read that value, it's worth giving it a semantic name
that makes that code easier to understand.
Previously, when a user view the message source of a message at the
very top with the blue box around, the tooltip for the button will
be partially hidden by the recipient bar. Ths cause is some legacy
CSS from, for example, 3cd33c0fea,
which increased the z-index for the bodies of selected messages.
The intent of that code appeared to be something around handling overlaps
between unread indicators and the blue selected message box. It's logically
incorrect, and testing demonstrates that the blue box works fine next to unread
messages without this change, so we can safely remove these z-index values.
This reverts commit e16e7630e6 (#19542).
This was compiling to
a.messagebox-content .message_time {
z-index: 1;
}
which did nothing, because .messagebox-content is a <div>, not an <a>.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Ever since we started bundling the app with webpack, there’s been less
and less overlap between our ‘static’ directory (files belonging to
the frontend app) and Django’s interpretation of the ‘static’
directory (files served directly to the web).
Split the app out to its own ‘web’ directory outside of ‘static’, and
remove all the custom collectstatic --ignore rules. This makes it
much clearer what’s actually being served to the web, and what’s being
bundled by webpack. It also shrinks the release tarball by 3%.
Signed-off-by: Anders Kaseorg <anders@zulip.com>