Ekiga crashes upon startup

[Fibonacci]

Binary package hint: ekiga

As the title says, Ekiga crashes upon startup. I see no window, no configuration menu, no nothing.
When running ekiga from a terminal, all I see is "Segmentation fault".

I've obtained a full backtrace (thread apply all bt full) on GDB.

I don't know if this is the same as bug #380146 or not, but I think it's not because the OP for that bug did manage to see the window for half a second. That doesn't even happen in my case.

[Fibonacci (fibonacci-prower)]

This is what I see when running ekiga w/ maximum debug level:

$ ekiga -d 4
2009/06/13 02:47:01.340 0:00.395 Version 3.2.0 by on Unix Linux (2.6.28-11-generic-i686) with PTLib (v2.6.1) at 2009/6/13 2:47:01.340
2009/06/13 02:47:01.343 0:00.399 Ekiga SVN revision: unknown
2009/06/13 02:47:01.356 0:00.411 Ekiga registered on D-Bus: org.ekiga.Ekiga
2009/06/13 02:47:01.357 0:00.428 PWLib File handle high water mark set: 17 Thread unblock pipe
2009/06/13 02:47:01.387 0:00.442 PWLib Thread high water mark set: 2
2009/06/13 02:47:01.387 0:00.442 PWLib File handle high water mark set: 19 Thread unblock pipe
2009/06/13 02:47:01.417 0:00.472 PWLib Thread high water mark set: 3
2009/06/13 02:47:01.417 0:00.472 PWLib File handle high water mark set: 21 Thread unblock pipe
2009/06/13 02:47:01.433 0:00.488 PWLib File handle high water mark set: 23 Thread unblock pipe
2009/06/13 02:47:02.036 0:01.091 PWLib Thread high water mark set: 4
2009/06/13 02:47:02.038 0:01.093 HalManager_dbus Initialising HAL Manager
2009/06/13 02:47:02.074 0:01.129 HalManager_dbus Populating device list
2009/06/13 02:47:02.309 0:01.364 HalManager_dbus Detected V4L capabilities on /dev/video0 name: UVC Camera (046d:09a1)
2009/06/13 02:47:02.310 0:01.365 HalManager_dbus Detected V4L2 capabilities on /dev/video0 name: UVC Camera (046d:09a1)
2009/06/13 02:47:02.816 0:01.871 HalManager_dbus Populated device list with 21 devices
2009/06/13 02:47:02.817 0:01.872 HalManager_dbus Populating interface list
2009/06/13 02:47:02.826 0:01.881 HalManager_dbus Populating full interface list failed - Method "getDevices" with signature "" on interface "org.freedesktop.NetworkManager" doesn't exist

2009/06/13 02:47:02.827 0:01.882 Detecting V4L2 devices
2009/06/13 02:47:02.831 0:01.886 PV4L2Plugin detected device metadata at /sys/class/video4linux/
2009/06/13 02:47:03.101 0:02.157 PWLib File handle high water mark set: 28 Thread unblock pipe
2009/06/13 02:47:03.104 0:02.159 PWLib Thread high water mark set: 5
2009/06/13 02:47:03.106 0:02.161 OpalMan Created manager.
2009/06/13 02:47:03.117 0:02.173 OpalMan Registered endpoint with prefix pc
2009/06/13 02:47:03.118 0:02.173 OpalEP Created endpoint: pc
2009/06/13 02:47:03.236 0:02.291 PCSS Created PC sound system endpoint.
Players:
Default
EKIGA
*.wav
Intel ICH5
Intel ICH5 (1)
Recorders:
Default
EKIGA
*.wav
USB Device 0x46d:0x9a1
Intel ICH5
Intel ICH5 (1)
Intel ICH5 (2)
Intel ICH5 (3)

2009/06/13 02:47:03.242 0:02.297 OPAL SetMediaFormatOrder()
2009/06/13 02:47:03.242 0:02.297 ...

[Fibonacci (fibonacci-prower)]

The cause of this bug is the same as for bug #386791 - namely, that getaddrinfo segfaults when /etc/hosts contains too long lines. Upon removing long lines from /etc/hosts, ekiga runs fine again.

So to reproduce this bug, just add a 4000+ line to /etc/hosts, then run the attached C program to get a segfault.

[Fibonacci (fibonacci-prower)]

I meant, a 4000+ *character* line.

[Yannick Defais (sevmek)]

Hi,

Thank you, pushing it upstream.

Best regards,
Yannick