For unknown reasons, deleting 10,000s of ArchiveTransaction objects
results in rapidly growing memory in the job making the request in the
Django process, eventually leading to an OOM kill.
I don't understand why Django behaves that way; I would have expected
the failure mode to instead be a serious load problem on the database
server, but perhaps the way Django's internal deletion logic handles
cascading the deletes to many millions of ArchiveMessages and other
ForeignKey objects requires tracking a lot of data in memory.
The solution is the same in any case, which is to batch the deletions
to execute a reasonable number of them at once. Doing a single
ArchiveTransaction at a time would likely result in huge numbers of
database queries in a loop, which performs very poorly. So we balance
by batching deletions in groups of 100 ArchiveTransactions; testing
this in production, I saw no spike of memory usage materially beyond
that of a normal Django process, and each bulk-deletion transaction
takes several seconds to process (meaning per-transaction overhead is
negligible).
memcached 1.5.22 in Ubuntu 20.04 has a bug where it looks for its SASL
configuration at /etc/sasl2/memcached.conf/memcached.conf instead of
/etc/sasl2/memcached.conf.
https://bugs.launchpad.net/ubuntu/+source/memcached/+bug/1878721
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Before we used a selector to check whether the "Subscribed" tab
is active or not. The problem is that if a new sibling of the tab
switcher is inserted in the DOM and uses the same classes as the
"Subscribed" and "All Streams" button do, the selector can get
the sibling element instead the tab button.
To avoid that, we are using a variable that tracks the current
active tab instead of using the selector.
I'm not sure exactly what series of history got us here, but we were
fetching the mobile_user_ids data for all users in the organization,
regardless of whether they were recently active (and thus relevant for
the main presence data set). And doing so in a sloppy fashion
(sending every user ID over the wire, rather than just having the
database join on Realm).
Fixing this saves a factor of 4-5 on the total runtime of a presence
request on organizations with 10Ks of users like chat.zulip.org; more
like 25% in an organization with 150. Since large organizations are
very heavily weighted in the overall cost of presence, this is a huge
win.
Fixes part of #13734.
* Stream bar color logic is borrwoed from compose stream bar.
* Use flex containers to align elements and automatically set their
height to be same, them automatically filling the stream color bar
height to be the height of the select box.
* Use flex-wrap to wrap the propagate selector when out of space.
* To make sure stream select box and stream color box are closest possible,
select box has been moved under stream color box.
A separate outer span with the title text needs to be added because:
(1) The default bootstrap popover behavior takes the title as the
popover title, if provided.
(2) We need to avoid having the title area be too big in the
me_message template.
Fixes#12769.
Co-Author-By: Vaibhav <vrongmeal@gmail.com>
The query to fetch the latest user activity was missing an
`.order_by('last_visit')`. This meant that the results were being
ordered by the `id`, which resulted in us getting `update_message_flags`
action performed on the client that the user installed last, instead of
being client agnostic and fetching the "global" last
`update_message_flags` action performed by the user.
"pyr_" events are like refund events some source called it a pseudo
refund event (https://stackoverflow.com/questions/46296374/how-can-
i-get-the-original-charge-and-refund-ids-of-an-automatic-payout).
Though due to the lack of any documentation on this event I'm not
confident if this is the right fix.
Signed-off-by: Hemanth V. Alluri <hdrive1999@gmail.com>
Fix a bug where the compose box didn't collapse when sending a message
from the preview area by hitting the send button. The bug ocurred because
the preview area wasn't being properly cleared when this flow was executed.
This was fixed by moving the clear_preview_area function call for a place
that will be reached by both the enter and button flow.
Fixes: #14889
Previously, we handled these updates in server_events_dispatch
and could accidentally call widget.render() before initializing
the widget.
Original report: https://chat.zulip.org/#narrow/near/875608.
The sync_realm_settings function ensures that if the settings are
not open, any updates are a noop.
Zulip's openapi specification in zulip.yaml has various examples
for various schemas. Validate the example with their respective
schemas to ensure that all the examples are schematically correct.
Part of #14100.
Some examples mentioned in zulip.yaml did not match their schema.
Change either the schema or the example so that all examples are
valid with respect to their schemas.
The `email` field for identifying the user being modified in these
events was not used by either the webapp or other official Zulip
clients. Instead, it was legacy data from before we switched years
ago to sending user_id fields as the correct way to uniquely identify
a user.
Use get_release_event_message from webhooks/git.py to format release
events using the newly implemented release message template.
Tweaked by tabbott to handle name=None.
Builds on #14746. Proposed in #14934.
If the #random channel in Slack is deactivated, we should follow
Zulip's data model of not allowing deactivated, default streams.
This had apparently happened in zulipchat.com for a few organizations,
resulting in weird exceptions trying to invite new users.
Aman Agrawal
Tim Abbott
Vishnu KS
Anders Kaseorg
clarammdantas
Aman Agrawal
Gillian Pantaleon
Puneeth Chaganti
Hemanth V. Alluri
arpit551
Mateusz Mandera
João Maurício Carvalho
Kartik Srivastava
Rohitt Vashishtha
majordwarf
orientor
cestrell
Rohan Prashant