In a decorator annotated with generic type (ViewFuncT) -> ViewFuncT,
the type variable ViewFuncT = TypeVar(…) must be instantiated to
the *same* type in both places. This amounts to a claim that the
decorator preserves the signature of the view function, which is not
the case for decorators that add a user_profile parameter.
The corrected annotations enforce no particular relationship between
the input and output signatures, which is not the ideal type we might
get if mypy supported variadic generics, but is better than enforcing
a relationship that is guaranteed to be wrong.
This removes a bunch of ‘# type: ignore[call-arg] # mypy doesn't seem
to apply the decorator’ annotations. Mypy does apply the decorator,
but the decorator’s incorrect annotation as signature-preserving made
it appear as if it didn’t.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Also enable warn_unused_ignores. I think the fact that there are so
few of these is good evidence that it’s not a significant burden for
people fixing type errors.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Fixes these compilation errors from webpack and PostCSS, exposed by
commit b10f156250 (#14997) which tries
to @extend these directives:
Unexpected '}' at app.d5da4b9d46e79634b8fb.css:9103:4.
Unexpected '}' at app.d5da4b9d46e79634b8fb.css:9104:0.
Invalid property name '@-moz-document url-prefix() {
@nest & #settings_page select {
background-color' at night_mode.scss:788:0. Ignoring.
Invalid selector '}
}
.user_status_overlay .overlay-content' at night_mode.scss:797:4. Ignoring.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
Migrations that add items to a table in the same migration htat
craetes the table require atomic=False.
This fixes this exception:
Running migrations:
Applying zerver.0287_clear_duplicate_reactions... OK
Applying zerver.0288_reaction_unique_on_emoji_code... OK
Applying zerver.0289_tighten_attachment_size... OK
Applying zerver.0290_remove_night_mode_add_color_scheme...Traceback (most recent call last):
File "/home/zulip/deployments/2020-06-22-23-20-36/zulip-py3-venv/lib/python3.6/site-packages/django/db/backends/utils.py", line 84, in _execute
return self.cursor.execute(sql, params)
File "/home/zulip/deployments/2020-06-22-23-20-36/zerver/lib/db.py", line 33, in execute
return wrapper_execute(self, super().execute, query, vars)
File "/home/zulip/deployments/2020-06-22-23-20-36/zerver/lib/db.py", line 20, in wrapper_execute
return action(sql, params)
psycopg2.errors.ObjectInUse: cannot ALTER TABLE "zerver_userprofile" because it has pending trig
This adds a convenient way to review the upgrade notes for all Zulip
releases that one is upgrading across.
I thought about moving all the upgrade notes to a common section, but
in some cases the language is clearly explaining changes in the
release that are not duplicated elsewhere, and I think it reads better
having them inline alongisde related changes.
This is important because the Hashicorp boxes are better designed, and
in particular, https://www.vagrantup.com/docs/boxes.html recommends
using the hashicorp boxes over the "ubuntu" namespace.
This commit changes stream_data.create_sub_from_server_data to use
stream id, instead of stream name, for checking whether subscription
already exists or not. We are using stream ids so that we can avoid
bugs related to live update after stream rename.
This commit changes would_receive_message and want_normal_display
functions in compose_fade.js to use stream id, instead of stream
name, to get sub object. We are using stream ids so that we can
avoid bugs related to live update after stream rename.
This commit changes populate_group_for_message_container function to
use stream id to get sub object, instead of stream name.
We already have stream id from message object, so we can easily
use it to fetch sub. We are using stream ids so that we can
avoid bugs related to live update after stream rename.
This commit changes stream_data.remove_subscriber to use stream id
instead of stream name. We are using stream ids so that we can
avoid bugs related to live update after stream rename.
Thsi commit changes stream_data.add_subscriber to use stream_id
instead of stream name. We are using stream ids so that we can
avoid bugs related to live update after stream rename.
This commit removes stream_edit.rerender function. We directly
call subs.rerender_subscriptions_settings directly from
server_events_dispatch.js, which was the only caller of rerender
function, as we already have sub object.
We are using stream ids so that we can avoid bugs related to live
update after stream rename.
We can use get_sub_by_id instead of get_sub to get the stream info,
as we already have stream id from the message object. We are using
stream ids so that we can avoid bugs related to live update after
stream rename.
This commit changes receives_notifications function to use
stream_ids instead of stream names. We are using stream ids so
that we can avoid bugs related to live update after stream rename.
We only use this in a few places, but they're really important places
for understanding the types in the codebase, and so it's worth having
a bit of expository documentation explaining how we use it.
(And I expect we'll add more with time).
Prior to this commit has:link, has:attachment, has:image
filter couldn't be applied locally and deferred filtering to
web server. This commits make sure client filters all messages
it can instead of completely deferring to the server and hence
improve speed.
A tradeoff is also made to turn off local echo for has: narrows
as messages with link sent to has:link narrow were locally echoing
to another narrow and not appearing in the active has:link narrow.
Fixes: #6186.
Previously, the message controls had a bug where they would trigger on
mobile with a single tap over the area they occupy when visible. This
is wrong because a user would expect to first see the controls and
only trigger them once they are visible (with a second tap).
The above bug is caused by the fact that we were using "opacity: 0" on
".message_controls > div" to hide the controls and "opacity: 1" on
".messagebox &:hover .message_controls > div" to show the controls on
hover, however, this would not effect the click action because
"opacity". So we used "pointer-events: none;" and "pointer-events:
all;" with the hopes that it would prevented the above bug, but in
practice, it didn't.
(the most probable explanation being that tapping the message_control
area would cause the "&:hover" rule to trigger and change the
"pointer-event" to "all" before it could prevent the click trigger,
But that explanation is just conjecture.)
This commit replaces both "pointer-events" attributes with
"visibility: hidden" and "visibility: visible" respectively. The
result being that the message_controls behave identically to before,
except without the above bug.
The addition to the ".has_actions_popover .info" selector is important
because without it, we would regress on issue #3172.
Trivia:
An alternate approach to using "opacity" is to set the
"display" attribute to "none", however, using "display" prevents the
transition from animating (which is probably why we were using opacity
here in the first place). "visibility" does not prevent the transition
from animating.
History: The "pointer-events" attribute was introduced in
4d5aa3ddc9 and it replaced prior code
which relied on the "visibility" attribute... But it seems PR #3792
was mostly focused on improving the positioning through removal of
`display: none`, but introduced opacity to make the animations work
rather than visibility as the replacement solution, which requires the
pointer-events hack and resulted in the bug described here.
Fixes the second bug described in #13642.
This helps user understand "Include muted" filter better and
saves keypress to focus on more used filters like "participated"
and "unread".
See https://github.com/zulip/zulip/issues/15482.
Anders Kaseorg
Tim Abbott
Felix
kadrianne
sahil839
Aman Agrawal
Priyank Patel
Gittenburg
Mohit Gupta
YashRE42
dangraw