Re: [Pan-devel] Recent font problem

[walt]

On 12/13/2011 12:26 AM, Heinrich Mueller wrote:

> Hi walt. Next time run pan with gdb and do a backtrace. Also do an
> "export G_DEBUG="fatal-warnings" before running it.

Very disappointing:

Starting program: /home/wa1ter/bin/pan
[Thread debugging using libthread_db enabled]
[New Thread 0xb5590b70 (LWP 21240)]
[New Thread 0xb4d8fb70 (LWP 21241)]
[New Thread 0xb458eb70 (LWP 21242)]
[New Thread 0xb3d8db70 (LWP 21243)]
1234
1234
1234
update sig -1
pan: cairo-surface.c:1287: cairo_surface_set_device_offset: Assertion `status 
== CAIRO_STATUS_SUCCESS' failed.

Program received signal SIGABRT, Aborted.
0xb7fe0424 in __kernel_vsyscall ()
(gdb) bt
#0  0xb7fe0424 in __kernel_vsyscall ()
#1  0xb716a191 in raise () from /lib/libc.so.6
#2  0xb716b9dc in abort () from /lib/libc.so.6
#3  0xbfffe42c in ?? ()
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

I think I need a breakpoint at cairo-surface.c:1287, but I'll need to
stop and review gdb a bit first.

Meanwhile I found a 100% reproducible way of hitting the assertion :)

I've been using gmane.org via ssl since you added that feature to pan.
If I use pan's cert editor to delete gmane's cert and restart pan, I
get a popup from pan asking me to (re)accept gmane's cert (because I
just deleted it, of course).

When I click 'accept' on that dialog, pan immediately hits the assert
that I've pasted above.  Never fails.

Can anyone else reproduce?

BTW, Duncan, I'm still using gtk2 to build pan.  Maybe gtk3 will work
better :)  I'll try it.