We now default to passive event listeners on all services,
as well as 100ms minimum granularity on setTimeout.
This is because many of the web-wrapped services constantly fire
totally unnecessary repaint events as fast as they can,
as well as actively listening for mouse events which also cause
unnecessary reflow.
On my machine, this cuts the CPU usage by ~70% while not affecting
the usability of any services.
There's a user-toggle available there. There's not a per-service
option for the ServiceList because it shouldn't affect anything
except in truly abnormal cases. If there needs to be one, we can add
it.
We've completely overhauled badge updates by introducing
a callable function which should cover the standard case.
We also allow separate custom JS and JS just for badge updates.
We also make the service editing box taller by default.
We fix the logging for injected stuff to actually remain accurate.
And replace each js_unread in each service with modernized version.
We allow for separate custom badge updates and custom JS to
avoid doing unnecessary title event hooking allow overrides in a
sensible order.
Also remove Yahoo messenger service, as it was shutdown earlier in
July.
Squashed commit of the following:
commit d661565c25
Author: Ryan Zumstein <[email protected]>
Date: Thu May 31 00:10:26 2018 -0500
Fix Gmail important count
commit 30abb3a3c0
Author: rzumstein <[email protected]>
Date: Mon May 21 01:25:45 2018 -0500
Show notifications in new Gmail interface
Fix#819
I think this issue was closed due to lack of feedback. This problem is still a thing and this small change should fix it.
In my custom service I use only "https://app.mysms.com/" instead of "https://app.mysms.com/#login" which every service restart go to login page instead of main app page.
- Slack got rid of the .unread_msgs and .unread_highlights selectors
- This uses the selectors currently available to determine if there
are any unreads, as well as the selector on the highlight badge
to find a count of unread highlights.