Add ignore_dbusclose settings

[matclab]

closes #731

However I was not able to test it because of X BadAtom error:

./dunst -print -verbosity debug
{
	appname: 'dunst'
	summary: 'startup'
	body: 'dunst is up and running'
	icon: 'dialog-information'
	raw_icon set: false
	icon_id: 'dialog-information'
	desktop_entry: ''
	category: 
	timeout: 10000
	urgency: LOW
	transient: 0
	formatted: '<b>startup</b>
dunst is up and running'
	fg: #888888
	bg: #222222
	frame: #aaaaaa
	fullscreen: pushback
	progress: -1
	stack_tag: 
	id: 2
	actions: {}
	script: (null)
}
DEBUG: RUN
DEBUG: XEvent: Ignoring '65'
DEBUG: XEvent: processing 'CreateNotify'
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Ignoring '65'
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Checking for active screen changes
DEBUG: XEvent: Ignoring '19'
DEBUG: XEvent: Ignoring '19'
DEBUG: XEvent: Ignoring '15'
DEBUG: XEvent: processing 'Expose'
X Error of failed request:  BadAtom (invalid Atom parameter)
  Major opcode of failed request:  23 (X_GetSelectionOwner)
  Atom id in failed request:  0x0
  Serial number of failed request:  218
  Current serial number in output stream:  218

[tsipinakis]

Some points: The notification spec has no concept of disallowing closing notifications, so to make this 'compliant' I'd choose to lie to the clients that the notification is actually closed, but keep it open.
To do this, we need to:

• Send a NotificationClosed signal back (see signal_notification_closed)
• Invalidate all client-based actions as they most likely won't work anymore (see notification_invalidate_action)

Additionally please update the example dunstrc and the man page with some documentation about the new option.

[tsipinakis]

Oh, I almost forgot the elephant in the room:

However I was not able to test it because of X BadAtom error:

Can you please expand more into this? Does it happen on master without this PR? If so, that's an important bug, please file an issue about it!

[matclab]

Oh, I almost forgot the elephant in the room:

However I was not able to test it because of X BadAtom error:

Can you please expand more into this? Does it happen on master without this PR? If so, that's an important bug, please file an issue about it!

I just tried with master and cannot reproduce the problem, and switching back on this branch I cannot reproduce it anymore. And then I remember that I had an external screen connected when the problem occurred. Checking it again with an external screen and the problem happens also with master. I'll open a bug report.

[matclab]

Note that I didn't call notification_invalidate_action which is already call by signal_notification_closed.

[matclab]

The CI fails on codecov timeout. Nothing I can do on my side.

[tsipinakis]

The changes look good. One thing left is that when a notification is closed, a warning is printed WARNING: Closing notification 'Test' not supported. Notification already closed. from signal_notification_closed. I guess the easiest way to solve this is to remove that line entirely.

As for the docs, Useful to enforce _the_ timeout set by dunst configuration, just missing a the but otherwise it looks good to me.

[matclab]

commits updated to edit the english according to review and remove the warning.

[tsipinakis]

LGTM

[tsipinakis]

What does received mean in this context?
I'd say queues_get_by_id would be a shorter and more descriptive name.

[matclab]

Hmmm I was trying to say that id was not looked for in all the queues…
queues_get_by_id would be a bit misleading as history queues is not used.

The best would probably to name the function queues_get_by_id and search in all queues. In our use case it won't change the looking time as we expect the id to be in one of the first two queues.
Would this be ok ?

[tsipinakis]

That's fine by me. Default history max is ~20 items so performance impact is negligible.

[tsipinakis]

Thanks for the PR! And sorry for the delay in merging this.

[matclab]

No problem. Thanks for your help.