Previously, navigating from any stream to the recent topics view would
cause a forced reflow every time we checked `is_visible()` because it
would call `$("#recent_topics_view").is(":visible")`.
The reason for this is related to how browsers ship frames, the
process follows these steps:
JavaScript > style calculations > layout > paint > composite.
(The layout step is called Reflow in firefox.)
Typically, the browser will handle these steps in the most optimal
manner possible, delaying expensive operations until they're needed.
However, it is possible to cause the browser to perform a layout
earlier than necessary. An example of this is what we previously did:
When we call `top_left_corner.narrow_to_recent_topics()`, we ask to
add a class via `.addClass()`, this schedules a Style Recalculation,
then, when we call `message_view_header.make_message_view_header()` it
calls `recent_topics_util.is_visible()` which calls
`$("#recent_topics_view").is(":visible")`.
Before the browser can get this value, it realizes that our dom was
invalidated by `.addClass()` and so it must execute the scheduled
Style Recalculation and cause a layout.
This is called a forced synchronous layout.
This commit adds a JavaScript variable representing the visible state,
in order to prevent the above behavior.
This commit reduces the main thread run time of
`build_message_view_header` from 131.81 ms to 5.20 ms.
Unfortunately we still have the case where
`recent_topics_ui.revive_current_focus()` calls
`recent_topics_ui.set_table_focus()` which causes a reflow.
However, by eliminating this reflow we still save ~100ms.
(It's important to note that we only save this sometimes, as other
things can still cost us a reflow.)
Further reading: https://developers.google.com/web/fundamentals/
performance/rendering/avoid-large-complex-layouts-and-layout-thrashing
Previously, the tests relied on `recent_topics_util.is_visible`
returning false unless we used override to return true, this is
because `recent_topics_util.is_viible` would return
`$("#recent_topics_view").is(":visible");` which would always be false
since we don't create a stub for `$("#recent_topics_view")`.
This would cause the tests to fail if we changed how `is_visible`
found it's return value, which we want to do for performance reasons,
and as such, this commit adds explicit overrides where needed.
Further more, even when we do override `is_visible` as override(rt,
"is_in_focus", ...) we do not override the `is_visible` used in
`recent_topics_uttil.is_in_focus`, and so our tests stil rely on that
`is_visible` always returning false. This commit address that by also
adding explicit overrides for `is_in_focus` where needed.
The `analyze_new_cluster.sh` script output by `pg_upgrade` just runs
`vacuumdb --all --analyze-in-stages`, which runs three passes over the
database, getting better stats each time. Each of these passes is
independent; the third pass does not require the first two.
`--analyze-in-stages` is only provided to get "something" into the
database, on the theory that it could then be started and used. Since
we wait for all three passes to complete before starting the database,
the first two passes add no value.
Additionally, PosttgreSQL 14 and up stop writing the
`analyze_new_cluster.sh` script as part of `pg_upgrade`, suggesting
the equivalent `vacuumdb --all --analyze-in-stages` call instead.
Switch to explicitly call `vacuumdb --all --analyze-only`, since we do
not gain any benefit from `--analyze-in-stages`. We also enable
parallelism, with `--jobs 10`, in order to analyze up to 10 tables in
parallel. This may increase load, but will accelerate the upgrade
process.
This increases the possible maximum wait time to send exceptions
during shutdown. The larger value makes it possible to send larger
exceptions, and weather larger network hiccups, during shutdown. In
instances where a service is crash-looping, it is already not serving
requests reliably, and better ensuring those exceptions are captured
is of significant value.
Previously, our codebase contained links to various versions of the
Django docs, eg https://docs.djangoproject.com/en/1.8/ref/
request-response/#django.http.HttpRequest and https://
docs.djangoproject.com/en/2.2/ref/settings/#std:setting-SERVER_EMAIL
opening a link to a doc with an outdated Django version would show a
warning "This document is for an insecure version of Django that is no
longer supported. Please upgrade to a newer release!".
Most of these links are inside comments.
Following the replacement of these links in our docs, this commit uses
a search with the regex "docs.djangoproject.com/en/([0-9].[0-9]*)/"
and replaces all matches with "docs.djangoproject.com/en/3.2/".
All the new links in this commit have been generated by the above
replace and each link has then been manually checked to ensure that
(1) the page still exists and has not been moved to a new location
(and it has been found that no page has been moved like this), (2)
that the anchor that we're linking to has not been changed (and it has
been found that no anchor has been changed like this).
One comment where we mentioned a Django version in text before linking
to a page for that version has also been changed, the comment
mentioned the specific version when a change happened, and the history
is no longer relevant to us.
Previously, our docs had links to various versions of the Django docs,
eg https://docs.djangoproject.com/en/1.10/topics/migrations/ and
https://docs.djangoproject.com/en/2.0/ref/signals/#post-save, opening
a link to a doc with an outdated Django version would show a warning
"This document is for an insecure version of Django that is no longer
supported. Please upgrade to a newer release!".
This commit uses a search with the regex
"docs.djangoproject.com/en/([0-9].[0-9]*)/" and replaces all matches
inside the /docs/ folder with "docs.djangoproject.com/en/3.2/".
All the new links in this commit have been generated by the above
replace and each link has then been manually checked to ensure that
(1) the page still exists and has not been moved to a new location
(and it has been found that no page has been moved like this), (2)
that the anchor that we're linking to has not been changed (and it has
been found that this happened once, for https://docs.djangoproject.com
/en/1.8/ref/django-admin/#runserver-port-or-address-port, where
/#runserver-port-or-address-port was changed to /#runserver).
In commit f6c78a35a4 we accidentally
deleted these link definitions, probably thinking that the end of the
markdown file would be the same as the end of the rendered doc. This
broke the links `[cloning your fork of the Zulip
repository][zulip-rtd-git-cloning]` and `[connecting the Zulip
upstream repository][zulip-rtd-git-connect]`.
This commit fixes things by adding back the definitions.
This resolves the issues reported in #20108, major chunk of which were
due to the incomplete support for importing the livechat streams/messages
in the tool. So, it's best not to import any livechat streams/messages for
now until a complete support for importing the same is developed.
In small screens, the quote used as a standin for image used
to overflow from screen as it didn't had responsive size set.
This image has additional bound of `max-width: 100%` which
stops them from overflowing which the quote did not.
Unless this was working around a specific bug that has somehow
persisted for eight years, this was just pointlessly defeating
Python’s bytecode cache every time you started the dev server.
This reverts commits deffda072f,
59228f7458, and
ae45217671.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
After the Stripe migration to the hosted checkout page, we were
missing one fixture for the test for switching from Standard to
Plus. This commit updates the fixtures for that particular test.
This more closely matches email_change_by_user and
password_reset_form_by_email limits; legitimate users are unlikely to
need to send more than 5 emails to themselves during a day.
PR #19576 added a settings option for selecting a notification batching
period. We want to extend that UI option with the ability to select
a custom period.
Tweaked by tabbott to have the natural model that picking a value
present in the dropdown live-updates to remove the custom input,
rather than only having the custom input disappear on reload.
Fixes#19713.
The decorator form is clearer by being more explicit; additionally,
the api_by_user rate-limit only currently used in one place, and makes
it difficult to test per-user rate-limits that are more specific.
Both `create_realm_by_ip` and `find_account_by_ip` send emails to
arbitrary email addresses, and as such can be used to spam users.
Lump their IP rate limits into the same bucket; most legitimate users
will likely not be using both of these endpoints at similar times.
The rate is set at 5 in 30 minutes, the more quickly-restrictive of
the two previous rates.
YashRE42
Alex Vandiver
Sahil Batra
Ganesh Pawar
Alya Abbott
Priyansh Garg
Eeshan Garg
Aman Agrawal
Anders Kaseorg
Vishnu KS
Andrew McAfee
Tim Abbott