dunst - a dmenu-ish notification daemon

zeltak · 2013-01-13 15:08:12

I would really love that as well, a different border color for diffrent notification levels (low,normal etc)

Z

Last edited by zeltak (2013-01-13 15:15:35)

Army · 2013-01-13 17:20:06

On my multi monitor setup it sometimes (~50%) still happens that dunst popups don't disappear when I use my keyboard shortcut. I use dwm. Everything's fine if I only use the laptop's display = with a single monitor setup.

I'm pretty sure this has been mentioned here before, but seemed to be fixed, but for me it's still happening. What can I do to provide useful informations?

Last edited by Army (2013-01-13 17:32:14)

knopwob · 2013-01-13 18:25:27

Army wrote:

On my multi monitor setup it sometimes (~50%) still happens that dunst popups don't disappear when I use my keyboard shortcut. I use dwm. Everything's fine if I only use the laptop's display = with a single monitor setup.
I'm pretty sure this has been mentioned here before, but seemed to be fixed, but for me it's still happening. What can I do to provide useful informations?

You can run dunst in a terminal and see wether it prints any errors when this happens.

Army · 2013-01-13 19:15:51

That's a reasonable starting point. I'm really tired today, that's my aliby for not coming up with this idea by myself

knopwob · 2013-01-16 18:32:10

Army wrote:

That's a reasonable starting point. I'm really tired today, that's my aliby for not coming up with this idea by myself

any update on this?

@all
I'm collecting screenshots for an update of the homepage. So if you have a cool setup using dunst that you want to share please post them here or send me an email and also please add a short note that you allow me to rehost the image on my side and wether you want to be credited on the screenshot and want to have a link to your homepage or to your github with your config files

Many thanks in advance

Army · 2013-01-17 07:55:29

knopwob wrote:

Army wrote:
That's a reasonable starting point. I'm really tired today, that's my aliby for not coming up with this idea by myself
any update on this?

Nope, all of a sudden there don't seem to be any problems anymore But I'll post here as soon as I have an output message.

Reded · 2013-01-17 15:21:49

knopwob wrote:

@all
I'm collecting screenshots for an update of the homepage. So if you have a cool setup using dunst that you want to share please post them here or send me an email and also please add a short note that you allow me to rehost the image on my side and wether you want to be credited on the screenshot and want to have a link to your homepage or to your github with your config files
Many thanks in advance

http://tinypic.com/r/2a0ja81/6

Go ahead and use it, credit me if you want but I don't have a homepage/github

By the way, is there any way to set the transparency on JUST the background colour of dunst? I think it'd look awesome if it were just the text floating!

chneukirchen · 2013-01-26 15:33:30

I noticed I get loads of messages like "Xlib: extension "XINERAMA" missing on display ":0.0"." while the popup is shown.

Army · 2013-01-26 23:08:33

Army wrote:

knopwob wrote:
Army wrote:
That's a reasonable starting point. I'm really tired today, that's my aliby for not coming up with this idea by myself
any update on this?
Nope, all of a sudden there don't seem to be any problems anymore But I'll post here as soon as I have an output message.

Ok, now the notification doesn't go away. There's no output when I just run dunst. I configured int verbosity and changed it to int verbosity = 1; (what value means what?). Using gdb doesn't show anything (it doesn't crash). Running "strace dunst" shows a hell lot of stuff. I don't know if this helps in any way, but here's the file (>2000 lines). I created it running

strace dunst 2>dunst

Then I did an action which shows up a notification, tried to remove the notification with the keyboard several times, then clicked on it, which made it disappear and then hit ctrl+c to kill dunst.

edit: If it helps, here's my config.h dunst was built with when I ran it through strace.

Hope these infos are okay and helpful.

Last edited by Army (2013-01-26 23:12:08)

knopwob · 2013-01-27 08:40:48

Army wrote:

Army wrote:
knopwob wrote:
any update on this?
Nope, all of a sudden there don't seem to be any problems anymore But I'll post here as soon as I have an output message.
Ok, now the notification doesn't go away. There's no output when I just run dunst. I configured int verbosity and changed it to int verbosity = 1; (what value means what?). Using gdb doesn't show anything (it doesn't crash).

We're still talking about a dual-head setup, right? Does it make a difference whether the keyboard focus and/or mouse is on the monitor on which the notifications are displayed or on the other one?
BTW. the verbosity currently does nothing.

Army wrote:

Running "strace dunst" shows a hell lot of stuff. I don't know if this helps in any way, but here's the file (>2000 lines). I created it running
strace dunst 2>dunst
Then I did an action which shows up a notification, tried to remove the notification with the keyboard several times, then clicked on it, which made it disappear and then hit ctrl+c to kill dunst.

I saw in the strace output that it's looking for a dunstrc. Even though it looks like it didn't find any, I recommend to compile dunst with

CFLAGS=-DSTATIC_CONFIG make

when configuring dunst via config.h. This disables parsing of config files completly, which might prevent confusion when there's a dunstrc somewhere where dunst looks for its config files, since the dunstrc overrides stuff from the config.h.

Army wrote:

edit: If it helps, here's my config.h dunst was built with when I ran it through strace.
Hope these infos are okay and helpful.

Hmm, I didn't find anything (obviously) wrong in your config.h

Army · 2013-01-27 09:39:24

> We're still talking about a dual-head setup, right?
Yes This never happened when I'm away from home only using my laptop display.

> Does it make a difference whether the keyboard focus and/or mouse is on the monitor on which the notifications are displayed or on the other one?
To be honest, I'm not 100% sure, because I never tested this deliberately. I will do next time it happens. I configure my dual head setup with

xrandr --output LVDS --auto --primary --output HDMI-0 --auto --right-of LVDS --output VGA-0 --off

So my laptop display is the primary display (dunst shows up there) and my big screen, which I'm working on, is secondary. It might solve the problem if I changed that a bit, but most of the time everything works perfect. I don't know what's different when all of a sudden the key combination doesn't make dunst vanish. In fact I don't think anything's different, but well ...

> I saw in the strace output that it's looking for a dunstrc
Yes it does. When I launch dunst, it shows me

no dunstrc found -> skipping

I configure dunst via config.h. I'll compile it with "CFLAGS=-DSTATIC_CONFIG make" and will change the package in the AUR.

Is there anything else I can do to debug this?

knopwob · 2013-01-27 09:45:38

Army wrote:

> We're still talking about a dual-head setup, right?
Yes This never happened when I'm away from home only using my laptop display.
> Does it make a difference whether the keyboard focus and/or mouse is on the monitor on which the notifications are displayed or on the other one?
To be honest, I'm not 100% sure, because I never tested this deliberately. I will do next time it happens. I configure my dual head setup with
xrandr --output LVDS --auto --primary --output HDMI-0 --auto --right-of LVDS --output VGA-0 --off
So my laptop display is the primary display (dunst shows up there) and my big screen, which I'm working on, is secondary. It might solve the problem if I changed that a bit, but most of the time everything works perfect. I don't know what's different when all of a sudden the key combination doesn't make dunst vanish. In fact I don't think anything's different, but well ...
> I saw in the strace output that it's looking for a dunstrc
Yes it does. When I launch dunst, it shows me
no dunstrc found -> skipping
I configure dunst via config.h. I'll compile it with "CFLAGS=-DSTATIC_CONFIG make" and will change the package in the AUR.

But I suggest that you just put the "CFLAGS=-DSTATIC_CONFIG make" only in a comment, since otherwise it will break for everybody who uses a dunstrc.

Army wrote:

Is there anything else I can do to debug this?

Currently I have no other idea. I'll let you know when I can think of anything.

Army · 2013-01-27 10:09:02

knopwob wrote:

But I suggest that you just put the "CFLAGS=-DSTATIC_CONFIG make" only in a comment, since otherwise it will break for everybody who uses a dunstrc.

Ok, I just worked on the PKGBUILD and already uploaded it to the AUR. Do you mind taking a quick look at it? I think the way I did it won't cause any trouble. I suppose that only people, who configure dunst with the config.h have this header file in their $startdir. For all the others dunst is built the usual way.

edit: While I'm on it, namcap showed me, that libxdg-basedir isn't a dependency, but dunst won't build without it. So I moved it to the makedepends and yes, dunst runs just fine without it. I guess the (make)dependencies are right?

Last edited by Army (2013-01-27 10:10:39)

knopwob · 2013-01-27 10:25:41

Army wrote:

knopwob wrote:
But I suggest that you just put the "CFLAGS=-DSTATIC_CONFIG make" only in a comment, since otherwise it will break for everybody who uses a dunstrc.
Ok, I just worked on the PKGBUILD and already uploaded it to the AUR. Do you mind taking a quick look at it? I think the way I did it won't cause any trouble. I suppose that only people, who configure dunst with the config.h have this header file in their $startdir. For all the others dunst is built the usual way.

aah, yes. This looks fine.

Army wrote:

edit: While I'm on it, namcap showed me, that libxdg-basedir isn't a dependency, but dunst won't build without it. So I moved it to the makedepends and yes, dunst runs just fine without it. I guess the (make)dependencies are right?

It's a runtime dependency when it's build with dunstrc support e.g. when not using a custom config.h. So when testing with namcap you should test both cases

Army · 2013-01-27 14:23:40

knopwob wrote:

It's a runtime dependency when it's build with dunstrc support e.g. when not using a custom config.h. So when testing with namcap you should test both cases

Damnit

zeltak · 2013-02-09 20:58:28

hya

ok ill be the one to put it out there (and maybe get flammed), can we have click through in dunst (when trasnparency is used) ala the ubuntu ones i keep finding it could be usefull alot of times

best

thx

knopwob · 2013-02-25 21:42:19

kJust a heads up:

I just pushed a bunch of new changes to git. Most importantly I changed the drawing backend to pango. Since pango uses different font strings you might have to update your dunstrc. "Monospace-12" becomes "Monospace 12" for example.

other changes include

markups: If you set "allow_markup = yes" in your global section, you can use a small subset of html markups like <b>bold</b> <i>italic</i> etc.
actions: Actions are buttons in other notification daemons. if a notification begins with "(A)" it contains an action and you can "push that button" with your context-menu shortcut
Use of blocking IO. instead of polling dbus multiple times per seconds dunst now only wakes up if there's actually something to do. This might slightly improve powerconsumtion, since the cpu can sleep longer. Assuming that there are no other processes keeping the cpu awake.

Since this was a huge set of changes there might be some breakage introduced, sorry in advance for that.
Please let me know if you encounter any issues.

progandy · 2013-02-26 00:53:13

There is some code missing you need to display progress values. The value hint is not extracted anymore, and the variable progress in dbus.c should be initialised with -1.

PS: if you want you could add support for icon fonts. Add a new placeholder and a translation table from icon name to unicode symbol for e.g. fontawesome (as a section in dunstrc)
With pango, the following font definition will then display all icons available in the icon font and everything else from unifont:
font = "fontawesome,unifont 16px"

Army · 2013-02-26 08:55:57

Just updated the PKGBUILD for the AUR. Looks like pango is a dependency now. Please update the INSTALL file.

I'm not sure if this has always been that way, but when I build dunst-git without a custom config.h, but don't have a dunstrc, dunst segfaults.

% dunst
no dunstrc found -> skipping

[1]    6486 segmentation fault (core dumped)  dunst

knopwob · 2013-02-26 23:23:34

progandy wrote:

There is some code missing you need to display progress values. The value hint is not extracted anymore, and the variable progress in dbus.c should be initialised with -1.

I knew I forgot something :-/
Anyway this should be fixed now.

Army wrote:

Just updated the PKGBUILD for the AUR. Looks like pango is a dependency now. Please update the INSTALL file.

Good call. I probably would have missed that for the next release.

Army wrote:

I'm not sure if this has always been that way, but when I build dunst-git without a custom config.h, but don't have a dunstrc, dunst segfaults.

No, this was a regression that should be fixed now.

Thanks everybody for testing.

Markus00000 · 2013-02-27 13:37:03

knopwob, where does the string "(U)" come from? It happened first in 20130227, I think.

Try this:

notify-send "x.y"

Result here:

(U) x.y

I haven't found anything related to this string in the docs. Does it have something to do with urgency?

progandy · 2013-02-27 13:50:54

The U is AFAIK a bug. It should only show when the notification has associated URLs.
Edit: It is a bug in the URL detection. x.y is recognized as a link.

Last edited by progandy (2013-02-27 13:57:43)

knopwob · 2013-02-27 15:32:22

progandy wrote:

The U is AFAIK a bug. It should only show when the notification has associated URLs.
Edit: It is a bug in the URL detection. x.y is recognized as a link.

exactly. an A would indicate an Action btw. and the url detection gets a lot of false positives. But I prefer this to wrongly not detected urls. patches that improve url detection are welcome. see menu.c

progandy · 2013-02-27 16:07:14

knopwob wrote:

progandy wrote:
The U is AFAIK a bug. It should only show when the notification has associated URLs.
Edit: It is a bug in the URL detection. x.y is recognized as a link.
exactly. an A would indicate an Action btw. and the url detection gets a lot of false positives. But I prefer this to wrongly not detected urls. patches that improve url detection are welcome. see menu.c

I added a commit to my fork, feel free to cherry-pick this and any other commit I made
https://github.com/progandy/dunst/commi … 78ee1a0288

Last edited by progandy (2013-03-01 14:14:05)

zequav · 2013-03-02 14:49:58

The last version in git broke -geometry x1 for me (no longer spans the whole monitor, behaves like 0x1).

(edit)

Also, my script to send notifications via dbus doesn't work anymore (it works fine on 0.5.0):

dbus-send --type=method_call --dest='org.freedesktop.Notifications' \
/org/freedesktop/Notifications org.freedesktop.Notifications.Notify \
string:"$1" \
uint32:1 \
string:'' \
string:"$2" \
string:"$3" \
array:string:'' \
dict:string:string:'','' \
int32:1000

I use it instead of notify-send because with this script if there is a notification showing, it gets replaced by the new one (it doesn't wait for timeouts). Useful for volume/brightness shortcuts . In the git version I don't see anything when I use this script.

Last edited by zequav (2013-03-02 15:06:53)

Arch Linux

#301 2013-01-13 15:08:12

Re: dunst - a dmenu-ish notification daemon

#302 2013-01-13 17:20:06

Re: dunst - a dmenu-ish notification daemon

#303 2013-01-13 18:25:27

Re: dunst - a dmenu-ish notification daemon

#304 2013-01-13 19:15:51

Re: dunst - a dmenu-ish notification daemon

#305 2013-01-16 18:32:10

Re: dunst - a dmenu-ish notification daemon

#306 2013-01-17 07:55:29

Re: dunst - a dmenu-ish notification daemon

#307 2013-01-17 15:21:49

Re: dunst - a dmenu-ish notification daemon

#308 2013-01-26 15:33:30

Re: dunst - a dmenu-ish notification daemon

#309 2013-01-26 23:08:33

Re: dunst - a dmenu-ish notification daemon

#310 2013-01-27 08:40:48

Re: dunst - a dmenu-ish notification daemon

#311 2013-01-27 09:39:24

Re: dunst - a dmenu-ish notification daemon

#312 2013-01-27 09:45:38

Re: dunst - a dmenu-ish notification daemon

#313 2013-01-27 10:09:02

Re: dunst - a dmenu-ish notification daemon

#314 2013-01-27 10:25:41

Re: dunst - a dmenu-ish notification daemon

#315 2013-01-27 14:23:40

Re: dunst - a dmenu-ish notification daemon

#316 2013-02-09 20:58:28

Re: dunst - a dmenu-ish notification daemon

#317 2013-02-25 21:42:19

Re: dunst - a dmenu-ish notification daemon

#318 2013-02-26 00:53:13

Re: dunst - a dmenu-ish notification daemon

#319 2013-02-26 08:55:57

Re: dunst - a dmenu-ish notification daemon

#320 2013-02-26 23:23:34

Re: dunst - a dmenu-ish notification daemon

#321 2013-02-27 13:37:03

Re: dunst - a dmenu-ish notification daemon

#322 2013-02-27 13:50:54

Re: dunst - a dmenu-ish notification daemon

#323 2013-02-27 15:32:22

Re: dunst - a dmenu-ish notification daemon

#324 2013-02-27 16:07:14

Re: dunst - a dmenu-ish notification daemon

#325 2013-03-02 14:49:58

Re: dunst - a dmenu-ish notification daemon

Board footer