Might not be necessary, but if it is then it prevents a sitation where a
websocket is still open but we've forgotten about it, so we will never
broadcast any new messages to it and the client will be practically frozen in
time until they disconnect and open a new websocket.
Also update the user's last_seen when the websocket is closed. This prevents a
user with js enabled who's actually idle being considered absent and being
rotated when their websocket accidentally closes for a few seconds.
This works around a bug in mobile Firefox where under certain cirucmstances two
elements inside an iframe both become the iframe's target elment at the same
time, which breaks the CSS logic so instead of exactly one form being displayed,
nothing is displayed.
Hacky workaround of weird behaviour in Firefox where on a page whose url has a
fragment/hash/anchor in it, sometimes a urlless meta refresh tag will jump to
the element instead of refreshing the page. Same thing happens if the meta
refresh tag's url component is the same as the page's url.
The nojs button appears when the stream is online and the user is not watching.
The js button appears when the stream is online and the media element either
(1) is not using the network or (2) fires an error event.
Matches the behaviour of the js chat. Makes it so if you submit an empty
message but with a correct captcha, you won't be deverified and given another
captcha until you successfully send a message (and exceed the flood threshold).
Previously you could fill in the captcha with no message and be given back a
new captcha, which doesn't make that much sense.
Made the 'Users in chat' header above the overflow area, so it always stays on
top. Now using `visibility: hidden;` instead of `display: none;` to show/hide
messages/users so that nojs css animations don't reset.
This adds a field 'watching' in `user_for_websocket` that's True iff WATCHING,
False iff NOTWATCHING, and None otherwise (since clients don't need to know if
a user is tentative or absent). When the value of this field changes for any
user, they get added to the update buffer (like with any other change).
Removed race condition in `t_sunset_users`: `broadcast_users_update` was being
called *after* a user was removed from memory (and for each user being removed,
which was redundant). In that scenario if there's a user in the update buffer
and `t_sunset_users` wins the race between it and `t_broadcast_users_update`,
then when `t_sunset_users` calls `broadcast_users_update` a KeyError would be
raised since the user's already been removed.
Fixed unintended behaviour of `t_sunset_users`: it was removing users based on
the result of `is_visible`, so users who were actually tenative (as opposed to
absent) were being removed.