0.7.2dev does not dock Thunderbird

[Léo Studer]

I've build 0.7.2dev on my machine in order to check if the previous bugs with TB 2 were fixed. However this development release appears to systematically fail docking TB with the following output:

The program 'alltray' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadWindow (invalid Window parameter)'.
  (Details: serial 80671 error_code 3 request_code 20 minor_code 0)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.

[Michael B. Trausch (mtrausch)]

sktop environment (and version) are you using, and on which distribution (and release)? Also, what window manager are you using?

If you attempt to run AllTray with the following command, what output do you see:

 $ ALLTRAY_DEBUG=ALL alltray -D thunderbird

Thanks for taking the time to report this issue!

[Léo Studer (leo-studer)]

I am using Ubuntu Intrepid x86_64, with compiz running on Gnome.

$ ALLTRAY_DEBUG=ALL alltray -D thunderbird
[2009-04-08 14:18:13 CEST] INFO/MISC: Debugging subsystem setup complete.
[2009-04-08 14:18:13 CEST] INFO/CL: Command line options parsed.
[2009-04-08 14:18:13 CEST] INFO/MISC: Set WNCK Client Type to PAGER
[2009-04-08 14:18:13 CEST] INFO/PROCESS: Child process 16015 (thunderbird) now running.
Registering Enigmail account manager extension.
Enigmail account manager extension registered.
The program 'alltray' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadWindow (invalid Window parameter)'.
  (Details: serial 30247 error_code 3 request_code 20 minor_code 0)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)

If I launch that with metacity instead of compiz I have no icon docked but the error message does not display:

$ ALLTRAY_DEBUG=ALL alltray -D thunderbird
[2009-04-08 14:48:57 CEST] INFO/MISC: Debugging subsystem setup complete.
[2009-04-08 14:48:57 CEST] INFO/CL: Command line options parsed.
[2009-04-08 14:48:57 CEST] INFO/MISC: Set WNCK Client Type to PAGER
[2009-04-08 14:48:57 CEST] INFO/PROCESS: Child process 19746 (thunderbird) now running.

[Michael B. Trausch (mtrausch)]

Alright, thanks. Would you feel comfortable trying the version currently at the lp:alltray branch to see if the issue persists then?

[Léo Studer (leo-studer)]

Okay I've built the development branch using bzr branch, now with compiz, TB is loading but still no icon:

$ ALLTRAY_DEBUG=ALL alltray -D thunderbird
[2009-04-08 16:34:08 CEST] INFO/MISC: Debugging subsystem setup complete.
[2009-04-08 16:34:08 CEST] INFO/CL: Command line options parsed.
[2009-04-08 16:34:08 CEST] INFO/MISC: Set WNCK Client Type to PAGER
[2009-04-08 16:34:08 CEST] INFO/PROCESS: Child process 4331 (thunderbird) now running.
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9266 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9366 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9424 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9131 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9130 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9130 == 4331? false
[2009-04-08 16:34:08 CEST] INFO/APP: maybe_setup called: 9612 == 4331? false
Registering Enigmail account manager extension.
Enigmail account manager extension registered.
Registering Enigmail account manager extension.
Enigmail account manager extension registered.
[2009-04-08 16:34:11 CEST] INFO/APP: maybe_setup called: 4347 == 4331? false

and with metacity also no icon is docked:

$ ALLTRAY_DEBUG=ALL alltray -D thunderbird
[2009-04-08 16:35:28 CEST] INFO/MISC: Debugging subsystem setup complete.
[2009-04-08 16:35:28 CEST] INFO/CL: Command line options parsed.
[2009-04-08 16:35:28 CEST] INFO/MISC: Set WNCK Client Type to PAGER
[2009-04-08 16:35:28 CEST] INFO/PROCESS: Child process 4436 (thunderbird) now running.
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9366 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9612 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9424 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9266 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9130 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9130 == 4436? false
[2009-04-08 16:35:28 CEST] INFO/APP: maybe_setup called: 9131 == 4436? false
[2009-04-08 16:35:31 CEST] INFO/APP: maybe_setup called: 4452 == 4436? false

[Michael B. Trausch (mtrausch)]

D'oh!

I assumed that since Firefox's startup script gets things right and re-uses the PID for the firefox instance, that Thunderbird would, too. This is not the case, in effect making this bug a dupe of 357468. This can be confirmed by running “xprop” and then clicking on the Thunderbird window, it will show that _NET_WM_PID is not what AllTray is looking for. Looks like I need to remember to ask for the output of that in the future to look for issues like this and recognize them. The _good_ news: The next source drop (0.7.3dev) of AllTray will fix this; in the meantime, hang on for a little bit and I'll have a workaround within the next couple of hours. I will let you know when that workaround exists for Thunderbird.