This is a pretty pure code move, where we moved stuff from
message_store to pm_conversations:
insert_recent_private_message() -> recent.insert()
recent_private_messages -> recent.get()
The object message_store.recent_private_messages was not
encapsulated in a function before this change. Now it is
hidden in the scope of pm_conversations.recent.
Both of the modules touched here maintain 100% line coverage.
The error message, in relevant part:
```
zerver/templatetags/app_filters.py: error: INTERNAL ERROR --
please report a bug at https://github.com/python/mypy/issues version: 0.560
File ".../site-packages/mypy/nodes.py", line 497, in accept
return visitor.visit_func_def(self)
File ".../site-packages/mypy/report.py", line 317, in visit_func_def
assert start_indent is not None and start_indent > old_indent
AssertionError:
zerver/templatetags/app_filters.py: : note: use --pdb to drop into pdb
```
Seems mypy gets confused by the comment following the decorator.
I got distracted, came back later to a successful test run in my
terminal, and thought I remembered finishing the change and just
kicking off a final test run to check.
In fact, there was an `assert False` right in the normal case for
production, and I just hadn't finished a test for that path. (m.-)
Definitely the most grateful I've been for our coverage checks,
which highlighted this for me.
Remove the `assert False`, and also finish writing the test it was
there to help me write. Those lines are covered now.
This is a wrapper over lru_cache function. It adds following features on
top of lru_cache:
* It will not cache result of functions with unhashable arguments.
* It will clear cache whenever zerver.lib.cache.KEY_PREFIX changes.
This will essentially run the code paths to go from whatever you were
at before to /all and back in the case of /new, which will call the
render function three times (!!), so remove this call because it isn’t
really necessary anyways.
Currently the new streams user list will populate twice when you click
the new stream button (or “+”), because it is triggered once directly
by the button click and then once by the hash change to /new, so we
want to ignore the changes by the hash change.
This should make it possible to use the zulip_ops base rules
successfully on chat.zulip.org. Many of the changes in this commit
are hacks and probably can be cleaned up later, but given that we plan
to drop trusty support soon, it's likely that most of them will simply
be deleted then.
All the event handler did was resetting some entries in the edit
bot form. This is unnecessary, because the whole form gets
destroyed anyway when closed.
This is done by rewriting JS manipulations of the DOM tree
in the bot-settings.handlebars template. Dead code involving
the affected JS variables is removed.
This is the first step in cleaning up the bot edit code.
Since the bot edit form appears dynamically, we remove
it from the static HTML scaffold, of which settings_sidebar
is a part of.
The fresh imported data shows that the users emails are not included
in the data. However, the data received from the older method of slack
(which is using legacy tokens) contains the email data of the users.
When removing the description from a stream (i.e. setting it to ""),
the UI was not correctly updating the description. This is because we
were checking incorrectly for a falsey value, rather than the specific
value undefined (which means the description wasn't changed).
It's too easy to go over the rate limits when using the webapp.
The correct fix for this probably involves some changes to which
routes get covered by what sort of rate limit, but for now, just
increase the limits.
If some bug in Bugdown results in a rendered message content that is
bigger than twice the message size, we now just throw an exception
from Bugdown. This is considerably better than the old behavior,
which might result in an enormous message being placed in the database
(potentially, bigger than the 1MB limit to store in memcached), which
would in turn result in tragic consequences.
This fixes#8322, in that it prevents the super bad outcome seen there
(where basically Zulip became unusable for everyone on the stream
where the message is posted). Now, the failure mode is just the
message failing to send. Still not ideal (and requires further work
on the URL embed feature), but not a minor problem, not a major one.
Steve Howell
Aman Jain
Rhea Parekh
Greg Price
Umair Khan
Priscilla
Brock Whittaker
Tim Abbott
Rishi Gupta
Priyank Patel
Robert Hönig
Sumana Harihareswara
YJDave