Timeout bugs

[LoneFenris]

Sorry for the delay, I had to test the most recent commit quite a bit to figure out what was going on. There's a few things happening here:

~~1. When the screen is sleeping due to the timeout, an incoming notification will wake it, and it will not sleep again until the notification is addressed. Likewise, it may not sleep at all if there is an active notification (my testing on that is incomplete).~~

~~2. When the screen is sleeping due to the timeout, and a user wakes it with input, but doesn't unlock it, it will eventually timeout and sleep again. However, on this second sleep, user input (ie. keyboard/mouse activity) will not wake the screen. A notification will still wake the screen in this state, though.~~

~~3. When the screen blanks after timing out when the user manually locked the screen (e.g. with Super+L), user input (keyboard/mouse activity) will not wake the screen. Similar to the situation in 2 above, an incoming notification will still wake the screen in this state. The difference here is that even the first monitor sleep off a manual lock won't wake up to user input.~~

4. If the the user locks the screen manually, the screen will wait for both the GNOME blanking timeout and the extension's timeout before blanking. For example, if the user's power settings have the monitor blanking/locking after 10 minutes, and has the extension set to timeout the screen after 30 minutes, then the screen will blank 40 minutes after the user manually locked the desktop. This is a relatively minor problem, though.

[mvnural]

I have the first issue happening after updating to the latest version as well.

[sunwxg]

I updated some code, could you do a test?

[LoneFenris]

2 and 3 above are fixed.

The behavior for 1 has improved somewhat, but is still not quite right. An incoming notification wakes the screen and successfully can go back to sleep, however, it will wait for the full timeout set in the extension before going back to sleep. This behavior changes if a user has previously woke the screen with input and didn't unlock it. An incoming notification after that will only wake the screen for roughly 30 seconds before sleeping it again. I think the latter behavior is probably the intended one, it just doesn't work that way without user action.

I think an incoming notification while the screen is locked in the unblanked state will also reset the timer for the timeout at the moment, but I need to test that again to be sure. If that is the case, it seems like a relatively minor problem like 4. EDIT: Tested again. This is indeed what's happening.

4 above has not changed, but it remains of little concern.

[sunwxg]

The code has been updated, could you test the case 1? For the case 4, it's acceptable for me, I don't want to fix it.

[LoneFenris]

All seems fixed now except 4

[sunwxg]

Thank you.