Rather than use a bulk insert via Django, use the faster
`bulk_insert_all_ums` that we already have. This also adds a `ON
CONFLICT` clause, to make the insert resilient to race conditions.
There are currently two callsites, with different desired `ON
CONFLICT` behaviours:
- For `notify_reaction_update`, if the `UserMessage` had already been
created, we would have done nothing to change it.
- For `do_update_message_flags`, we would have ensured a specific bit
was (un)set.
Extend `create_historical_user_messages` and `bulk_insert_all_ums` to
support `ON CONFLICT (...) UPDATE SET flags = ...`.
This commit updates the code for parsing strings to
numbers to use correct types -
- Number.parseInt and Number.parseFloat take strings
as inputs but we were passing Number objects.
- We were using string as an operator for "*" arithmetic
operation, which is incorrect.
This commit fixes the above mentioned things and would
be helpful when we convert settings_components file to
typescript.
THis commit refactors get_realm_time_limits_in_minutes
code to not convert the setting value to integer and just
keep it string. This change will help us in not defining
the type of the variable as "number | string" when we can
easily use the string values.
Fix overlapping UI layout for stream subscribers search
on screens smaller than 320px. Break the search field
to a new row in the layout. Also fix the same issue in
the group settings > members panel.
Fixes#29368
These default to off, because in situations with thousands of queues,
consumers, and producers, they cause unreasonable overhead. Our use
case has few enough queues that we do want to be able to inspect them
individually.
Enable per-object Prometheus metrics, per [1].
[1]: 78851828ec/deps/rabbitmq_prometheus (configuration)
There seem to be two different ways of making the unread
marker visible:
(1) `user-with-count` classname on the `user_sidebar_entry`
element, which is added on render and not modified after
render at all.
(2) `hide` classname, which is toggled on the `unread_count`
element in `update_unread_count_in_dom`.
This creates a bug where sometimes the buddy list doesn't
update when messages are marked unread. This actually
doesn't always happen, since if there was an unread marker
that was there on the original render (and then removed
and put back) it will display again.
This commit keeps the `hide` class and removes the
`user-with-count` strategy.
Instead of only checking for visible `a` tags, we should also check
for visible elements with the `tabindex=0` attribute defined. This
allows us to support navigation for other components which we
explicitly set to be tabbable.
This commit introduces a theme switcher feature within the personal
menu popover. The implementation begins with the development of a
tab picker, which has the following features:
* Utilization of radio buttons to emulate the tab picker.
Radio input buttons provides the native way in HTML to select one
value out of a set.
* Support for both horizontal (default) and vertical orientations.
Vertical orientation can be achieved by appending the
`.tab-picker-vertical` class.
* Respects the `prefers-reduced-motion` option set by the user.
Disables the sliding tab animation to improve accessibility.
Additionally, the theme switcher component incorporates error handling
mechanisms. In the event of a server/network error, the tab slider
reverts to the previous theme option after a delay of 500ms. This
behavior visually communicates the occurrence of an error to the user,
improving the UX.
Fixes: #22803.
To increase the discoverability of the possibility of specifying the
language for a code block, we trigger the language typeahead when code
syntax is added using the code formatting button. A blank option is
shown preselected in the typeahead, so that pressing enter will not
mistakenly add a language to the code block.
We only trigger the typeahead on empty opening fences when added by the
button, by setting a state variable to true when adding the syntax using
the button, checking for this state when sourcing languages for the code
typeahead, and then resetting the state variable to false.
Fixes: #29150.
Now we don't hide the typeahead if the focus is back in the input within
150 ms. This is common when using the compose formatting buttons, which
only momentarily take the focus away from the input.
This is a prep commit for the next, to show typeahead on adding code
syntax with the code formatting button.
The inline `format` function defined in `format_text` function` now
returns true if formatting was added rather than removed, else false.
This is a prep commit for the upcoming change to show typeahead on using
the code formatting button to add code formatting to any selected text.
The bots do not exist in the user table to look up their active
status, and attempting to import them into the analytics table will
result in duplicate rows.
We only expect 'data-tooltip-template-id' to be present if 'tippy_content'
is not valid.Hence, only when 'tippy_content' is undefined, should we create
'$template'.
Uptil now, the `query_matches_string_in_order` function would respect
the order of words in the query string when matching a source string.
This meant that for query "two one", the source string "one two three"
would not be matched.
For more flexibility, a new function, `query_matches_string_in_any_order`,
has been added., which returns true if each word in the query string matches
the prefix of a distinct word in the source string, else it returns false.
The algorithm for computing this is quadratic in terms of the source word
count so can be a little expensive, but it is only currently used for
searching topics in Recent Conversations, where the strings' length is
limited by the max stream / topic name length allowed, so this should be
performant enough for this use case.
Replace the long string for organisations that have notification
body/content disabled (settings.PUSH_NOTIFICATION_REDACT_CONTENT
set to true) with "New message".
This allows more of the limited space on the mobile device screen to
be used for additional messages rather than this verbose content.
Fixes#29152
Using --rotate-key without write access to the secrets file is currently
quite painful, since you end up rotating your registration's secret with
no local record of it; so effectively you lose your registration and
need help from support. We should just prevent this failure mode.
Previously, #26419 addressed the majority of these calls, but did not
prevent more from creeping in. Remove the one remaining
callsite (after the cleanup from the previous commits), and ban any
future use of the pattern.
For the common case of not needing to reference the UserMessage row
later, and for being a stream without private history, the UserMessage
row is irrelevant. Convert `has_user_message` to a thunk, and defer
loading it unless necessary.
Calling `.select_related()` with no arguments joins through every
possible table, recursively. In this case, this currently produces a
query which joins through forty-three tables.
This is rather inefficient, particularly for what is a very common
call which should be very fast.
No callsite depends on having prefetched any joined table on the
object; drop all of the joins.