Apparently, our nginx configuration's use of "localhost", combined
with the default in modern Linux of having localhost resolve to both
the IPv4 and IPv6 addresses on a given machine, resulted in `nginx`
load-balancing requests to a given Zulip server between the IPv4 and
IPv6 addresses. This, in turn, resulted in irrelevant 502 errors
problems every few minutes on the /events endpoints for some clients.
Disabling IPv6 on the server resolved the problem, as does simply
spelling localhost as 127.0.0.1 for the `nginx` upstreams that we
declare for proxying to non-Django services on localhost.
This improves test coverage for a lot of our webhooks that relied
on ad-hoc methods to handle unexpected event types.
Note that I have deliberately skipped github_legacy, it isn't
advertised and is officially deprecated.
Also, I have refrained from making further changes to Trello, I
believe further improvements to test coverage should be covered
in separate per-webhook commits/PRs.
UnexpectedWebhookEventType is a generic exception that we may
now raise when we encounter a webhook event that is new or one
that we simply aren't aware of.
Fixes#9492.
`is` operator uses predefined categories. This commit
displays an invalid operand message if the operand does not fall into
any of these categories and the `is` operator is not at the last.
e.g. `is:abc sender:abc@zulipchat.com` will have `invalid abc operand
for has operator, sent by abc@zulipchat.com` as a prefix for all its
suggestions.
Fixes#9492.
Default suggestion e.g `abc messages` as a suggestion for `is:abc`
is not shown in a new suggestion. But if the is operator is already
present before any other operator, the default message text will be
used. e.g `is:abc sender:abc@zulipchat.com` will have all the suggestions
with the prefix `abc messages, sent by abc@zulipchat.com`.
The mypy daemon only supports error and skip. Choosing error
immediately surfaces when a file excluded from the build is
being imported, so that something can be done about it.
`get_containing_suggestions` was used to get the operand suggestions
for the `has` operator. `get_special_filter_suggestions` is now used
to get both the operand and operator suggestions for `has`.
Partially fixes#9461.
Negated suggestion for both operand and operators are handle in
get_special_filter_suggestions. A bug is get_operator_suggestions
causing the removal of `-` symbol from the operand was also fixed.
Now that we've moved it into a bulleted set of options inside a modal,
there's no good reason to have separate variables for the corner cases
around who can manage a stream.
Commit fa18913b8 preserved the fix to the actual import, but added
back the `ignore_missing_imports` line that had shielded the
erroneous import. Remove that line again.
Our logic for stream_has_topics never accounted for
us creating essentially "empty" stream buckets that
don't have topics, but we recently added some code
related to unread counts that violated the original
assumptions of the code.
Now we check deeper into the stream bucket to find
actual topics.
This bug manifested in the left sidebar where users
were seeing streams as recently active just because
some muted topics had unread counts, when in fact
the stream was inactive for practical purposes.
Having submessages will become common enough that the
info message here is too spammy, and there are other
ways to observe incoming submessages if you're doing dev
debugging, which this was originally written for.
We could get submessage events for messages that weren't
in our message store if somebody played with a widget
that was on an "old" message for somebody else.
We've had this sort of logic for GCM for a long time; it's worth
adding for APNS as well.
Writing this is a bit of a reminder that I'm not a fan of how our unit
tests for push notifications work.
Dropdown element for outgoing interface type was not showing correct
value, cause the way default value was set to dropdown was incorrect
(it should have been setting the selected parameter on the selected
option if it were going to be selected via the template code).
Fixes#9419.
We add conditional infinite sleep to this delivery job as a means to
handle case of multiple servers in service to a realm running this
job. In such a scenerio race conditions might arise leading to
multiple deliveries for same message. This way we try to match the
behaviour of what other jobs do in such a case.
Note: We should eventually do something to make such jobs work
while being running on multiple servers.
This revised GitHub auth backend test is inspired by the end-to-end
flow model of the Google auth backend test. My hope is that we will
be able to migrate the rest of the important cases in the GitHub auth
backend tests to this model and then delete what is now
GitHubAuthBackendLegacyTest.
The next step after that will be to merge the GitHub and Google auth
tests (since actually, the actual test functions are basically
identical between the two).
Apparently, the bug here was that we were aliasing the user_profile
variable, so that the results depended on what the last iteration in
the loop landed on.
Now that we have tsearch_extras packages uploaded, this mostly works.
There's a few issues being debugged in #9460; they should be fixed
soon, and regardless, merging this will make it easier to develop.
Since this is a logged-out view, need to actually write code for the
case of deactivated realms.
The change to get_active_user is more for clarity; the Django password
reset form already checks for whether the user is active earlier.
Tim Abbott
Eeshan Garg
Umair Khan
Shubham Dhama
Shubham Padia
Michael J. Sullivan
Greg Price
Steve Howell
Yashashvi Dave
Aditya Bansal