Behavior of chrome-Gmail-Ubuntu notification

[eliabntt]

Hey, I'm in Ubuntu and I have a custom rule for chrome that keeps notifications up for 30 seconds + override the dbus timeout. That's working fine. However, gmail notification are immediately displayed as (A)(U) and after 2-3 seconds (A) disappears and if I click in the notification the action if I click on the notification is opening the URL.
Also, the URL is only kept on the first level, so since I have two users (i.e. mail.google.com/mail/u/[1,0]) it just open the default (0) one.
Is it something we can control through configuration? A possible bug?

The dunst -print shows the correct actions

	actions: {
		"default": "Activate"
		"settings": "Settings"
	}

I am running dunst on xwayland on

Linux 6.11.0-26-generic 
Ubuntu 24.04.2 LTS

[bynect]

what is the value of urls printed by dunst -print?

[eliabntt]

{
	appname: 'Google Chrome'
	summary: 'Elia'
	body: '<a href="https://mail.google.com/">mail.google.com</a>

adsadas'
	icon: '/tmp/.com.google.Chrome.7Zndp2/icon.png'
	raw_icon set: false
	icon_id: '(null)'
	desktop_entry: 'google-chrome'
	category: 
	timeout: 20000
	start: 0
	timestamp: 694233787
	urgency: NORMAL
	transient: 0
	formatted: '<b>Elia</b>
mail.google.com

adsadas'
	fg: #ffffffff
	bg: #285577ff
	frame: #aaaaaaff
	highlight: #1745d1ff
	fullscreen: show
	format: <b>%s</b>\n%b
	progress: -1
	stack_tag: 
	id: 3
	urls:
	{
		[mail.google.com] https://mail.google.com/
	}
	actions: {
		"default": "Activate"
		"settings": "Settings"
	}
	script_count: 0
}

this is what dunst -print shows

[bynect]

From the notification body it seems like the link of Gmail did not include the user /u/N

[eliabntt]

That I totally Agree. What it's not clear is why if I click on it during the first few seconds the email message that I just received pops out while if I wait 5+ seconds it just opens the webpage. Is it possible that you're not getting the full/you are "cleaning" the notification?

…

On Wed, Jul 9, 2025, 19:06 nect ***@***.***> wrote: *bynect* left a comment (dunst-project/dunst#1488) <#1488 (comment)> From the notification body it seems like the link of Gmail did not include the user /u/N — Reply to this email directly, view it on GitHub <#1488 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEXDUJQ2QJDVT3UK2SELTOD3HWN7FAVCNFSM6AAAAAB7UBNHFCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANJUGM3TOOBVGI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bynect]

That I totally Agree. What it's not clear is why if I click on it during
the first few seconds the email message that I just received pops out while
if I wait 5+ seconds it just opens the webpage. Is it possible that you're
not getting the full/you are "cleaning" the notification?

On Wed, Jul 9, 2025, 19:06 nect @.***> wrote:

bynect left a comment (#1488)
#1488 (comment)

From the notification body it seems like the link of Gmail did not include
the user /u/N

—
Reply to this email directly, view it on GitHub
#1488 (comment),
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AEXDUJQ2QJDVT3UK2SELTOD3HWN7FAVCNFSM6AAAAAB7UBNHFCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANJUGM3TOOBVGI
.
You are receiving this because you authored the thread.Message ID:
@.***>

I am not sure about how gmail sends notification, but I suspect that the program waits only for a limited time for actions. I will try to investigate, but maybe the duration of the action is not dependent on dunst

[eliabntt]

I believe the action stays on only for the normal duration of the notification. If we override the dbus timeout (necessary to get the notification more persistent) the action "disappears". Could these be related?

[bynect]

I believe the action stays on only for the normal duration of the notification. If we override the dbus timeout (necessary to get the notification more persistent) the action "disappears". Could these be related?

yes this is exactly the problem. because when we override the dbus timeout, the application does not know that. so the notifications stays up, but the actions are not listened anymore. maybe you could change the timeout directly from the browser/gmail?

[eliabntt]

I checked and I don't think I can. Can we override the action timeout as well?

…

On Thu, Jul 10, 2025, 13:45 nect ***@***.***> wrote: *bynect* left a comment (dunst-project/dunst#1488) <#1488 (comment)> I believe the action stays on only for the normal duration of the notification. If we override the dbus timeout (necessary to get the notification more persistent) the action "disappears". Could these be related? yes this is exactly the problem. because when we override the dbus timeout, the application does not know that. so the notifications stays up, but the actions are not listened anymore. maybe you could change the timeout directly from the browser/gmail? — Reply to this email directly, view it on GitHub <#1488 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEXDUJXE4KMO6SHDLKVTRXL3H2RFDAVCNFSM6AAAAAB7UBNHFCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANJYGM3DSOBWHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bynect]

the action timeout is decided by the application sending the notification. I don't think it is possible to forcefully change it

[eliabntt]

Can we inspect/expose it? In this case is probably something like open a pop up at this url but dunst print doesn't show it

…

On Thu, Jul 10, 2025, 18:15 nect ***@***.***> wrote: *bynect* left a comment (dunst-project/dunst#1488) <#1488 (comment)> the action timeout is decided by the application sending the notification. I don't think it is possible to forcefully change it — Reply to this email directly, view it on GitHub <#1488 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEXDUJT7V4WNRDMOGIKTE233H3QZXAVCNFSM6AAAAAB7UBNHFCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANJZGI4DQMBYGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bynect]

i searched a bit and it seems like chrome does not allow for notification timeout to be specified. this is something not in our control unfortunaetly

[eliabntt]

Yes that's what my searched yielded too - that's why I am overriding the dbus timing.
My question would be now if we can access the "action" info part of the notification (e.g. through scripting). At that point I might be able to work something out for my use-case (e.g. expose a different link by republishing the notification)

[eliabntt]

Is there a way to "capture" the original action? e.g. through scripting?

[bynect]

you can save the action name, but in the end the action is handled by the browser. so if the browser stops listening for the action result, even if you make the notification last longer or create a fake copy, the browser will still not execute the action.
unfortunately this problem seems to be out of our hands. the only thing I see if there is some experimental option/plugin for the browser to make notification last longer on their end

[bynect]

maybe you could create a script that gets triggered when you receive an email and has the reply button, but lasts for a longer timeout?