Previously [{operator: "topic", operand: "one\xa0two"}] would be
unparsed to "topic:one\xa0two" which parses as [{operator: "topic",
operand: "one"}, {operator: "search", operand: "two"}], leading to
exceptions in the search pill system.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Previously, the "stream_topic_history" used to store unacked float
message ids as well in its "max_message_id" of stream and
"message_id" of topic histories.
This commit updates it to rather store only the acked message ids
here, and rather use the "echo_state" module so as to look up
for unacked messages in case of looking for recent topics, or
max message id in functions.
Previously, when `insert_local_message` was called, the
data structures in the `echo_state` are updated with the new
local messages after calling `insert_new_message`. This would
update the stream sidebar before even updating the `echo_state`
with the new local messages.
This commit introduces `track_local_message`, which basically
updates the `echo_state` data structures with the local
message before actually updating the stream sidebar.
This is a preparatory commit to update stream_topic_history
to only contain acked message ids.
The narrow_state module has a function that returns what this needs.
This does remove a log statement, but I don't think it was useful; we
don't need to or have a useful way to colorize a channel that doesn't
exist.
This commit refactors the `ensure_channel_topic_terms` of filter.ts.
Previously, this method used to add channel and topic terms, with
operands as placeholders in case the `with` narrow doesn't have
channel-topic terms.
This commit updates it to rather correct the narrow with the right
terms in case the channel-topic terms are missing in the `with`
narrow, but leave it as it is in case the channel-topic terms are
present, so that it can later be corrected if the channel-topic
terms are not pointing to the right conversation.
Previously, no custom styling was being applied to the enter send
choice options, which led to uneven styling from the other popover
options, as well as the outline ring being cut-off from the edges of
the popover. This commit fixes these issues by adding custom styling
for the outline ring when the enter send choice options are focused.
This was causing a bug where the participants weren't necessarily
all getting rendered, specifically when there were many subscribers
to a channel, because we render users in batches as buddy list
scrolls, and those users would only show up after some scrolling.
This fix makes sure we always load the participants first.
We create an unnamed user group with just the group creator as it's
member when trying to set the default. The pattern I've followed across
most of the acting_user additions is to just put the user declared
somewhere before the check_add_user_group and see if the test passes.
If it does not, then I'll look at what kind of user it needs to be set
to `acting_user`.
We also add the exception for the group creator to be able to edit their
group in this commit. This exception was added in the backend in earlier
commits.
This commit does not add the logic of using this setting to actually
check the permission on the backend. That will be done in a later
commit.
Only owners can modify this setting, but we will add that logic in a
later commit in order to keep changes in this commit minimal.
Adding the setting breaks the frontend, since the frontend tries to find
a dropdown widget for the setting automatically. To avoid this, we've
added a small temporary if statement to `settings_org.js`.
Although, most lists where we insert this setting follow an unofficial
alphabetical order, `can_manage_all_groups` has been bunched together
with `can_create_groups` since keeping those similar settings together
would be nicer when checking any code related to creating/managing a
user group.
This commit does not add the logic of using this setting to actually
check the permission on the backend. That will be done in a later
commit.
Adding the setting breaks the frontend, since the frontend tries to find
a dropdown widget for the setting automatically. To avoid this, we've
added a small temporary if statement to `settings_org.js`.
When adding or removing recipients for a direct message, the
placeholder text may cause a change in the size of the compose box,
so this adds a call to `compose_ui.autosize_textarea` when setting
that placeholder attribute.
If the user group being deactivated is used as a setting for a stream
which cannot be accessed by the user trying to deactivate the group,
we show "Unknown stream" in the banner mentioning where the group
is used.
This commit adds code to list the groups and streams
where the group is being used as settings in a banner.
The banner also mentions if the group is being used for
realm settings.
This isn't a bug fix, but it does clean up the code. `reload_setup`
only uses the `stream_id` if it can parse it as an integer and
otherwise falls back to `undefined`, so it makes more sense to not
include undefined stream ids.
Fixes#31750.
The control buttons of the edit box were flickering b/w default cursor
and pointer cursor after c1d155d923.
The addition of `position: relative` to `.edit-content-container` was
causing that. We've now added a div to apply position:relative to, this
div does not have any edit controls in it's children.
This is helpful for when a user wants to delete multiple pills
in a row, now that a pill needs to be highlighted before it can
be deleted.
It makes more sense intuitively for the previous pill to be selected
after backspace, since the direction of deletion is backwards. We
still potentially focus next() if there is no previous pill, which
is most helpful when the input element has no more pills left, since
this focuses the text field.
Select the pill on the first backspace and delete the whole pill on
the second backspace. If the pill is already highlighted from
left-pressing, then backspace would delete it right away.
We're making this change because it can be quite annoying to
re-type out a pill that's accidentally deleted, and users might
think pills are editable and accidentally delete the whole thing
with a backspace stroke.
The HTML id attribute is supposed to be globally unique; it’s not an
appropriate place to store a user-controlled string, or to identify
part of a component that’s rendered more than once.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Anders Kaseorg
Aman Agrawal
roanster007
Tim Abbott
Sayam Samal
Sahil Batra
evykassirer
opmkumar
Shubham Padia
Vector73
Lauryn Menard
sanchi-t
Karl Stolley