After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 324618 - g-v-m-1.5.7 crashes
g-v-m-1.5.7 crashes
Status: RESOLVED FIXED
Product: gnome-volume-manager
Classification: Deprecated
Component: general
1.5.x
Other All
: Normal critical
: ---
Assigned To: Gnome volume manager maintainers
Gnome volume manager maintainers
Depends on:
Blocks:
 
 
Reported: 2005-12-20 15:39 UTC by Joseph Sacco
Modified: 2006-01-05 18:12 UTC
See Also:
GNOME target: ---
GNOME version: 2.13/2.14

Attachments
fix for crash (854 bytes, patch)
2006-01-05 01:52 UTC, John (J5) Palmieri 		none Details | Review

Description Joseph Sacco 2005-12-20 15:39:24 UTC 
Steps to reproduce:
Three ways to reproduce the crash:

(1) Launch GARNOME-2.13.3 desktop 
    ==> g-v-m silently crashes.

(2) bring up "removable drives and media" dialog. click on "multimedia" tab 
    => crash is trapped by bug-buddy.

(3) start g-v-m from gdb 
    ==> g-v-m crashes.  
    stack trace attached.

Stack trace:
Backtrace was generated from '/opt/garnome/bin/gnome-volume-manager'

Using host libthread_db library "/lib/tls/libthread_db.so.1".
[Thread debugging using libthread_db enabled]
[New Thread 808544032 (LWP 24507)]
0x0fc5f2b8 in __waitpid_nocancel () from /lib/tls/libpthread.so.0

+ Trace 64725

Thread 1 (Thread 808544032 (LWP 24507))

  • #0 __waitpid_nocancel
    from /lib/tls/libpthread.so.0
  • #1 libgnomeui_segv_handle
    at gnome-ui-init.c line 786 


Other information:
System
------
* G4 PPC running YDL-4.0.1+
* GARNOME-2.13.3
* dbus-0.60
* hal-0.5.5.1

Note: g-v-m-1.5.5 does *not* crash when built in this environment
Comment 1 William Lovaton 2005-12-26 15:43:46 UTC 
Would this be the same bug #176445 reported in Red Hat's bugzilla?
[] https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=176445

I am using the same version of g-v-m on a fedora devel box using VMWare, more information on the link provided above.
Comment 2 Jeffrey Stedfast 2006-01-04 19:49:15 UTC 
this doesn't make sense to me... the bt suggests that the mount_point argument is NULL (or otherwise an invalid string such that strlen crashes), yet this cannot possibly be true as mount_point is hard-coded to ""
Comment 3 Jeffrey Stedfast 2006-01-04 19:53:25 UTC 
fwiw, I'm unable to reproduce with g-v-m 1.5.7 & dbus 0.50
Comment 4 John (J5) Palmieri 2006-01-05 01:50:36 UTC 
pointers and arrays are not addressed same in C so when you get the address of the array (&options) *(&options) will derefrence to "sync" and not {"sync", NULL}.  Consiquently when d-bus gets the first element of "sync" it gets a out of range address and causes a crash.  Whenever passing static arrays to d-bus you must do this:

char *array[2] = {"foo", "bar"};
char **parray = array;

dbus_message_append_args (message, DBUS_TYPE_ARRAY, DBUS_TYPE_STRING, &parray, 2, DBUS_TYPE_INVALID);
Comment 5 John (J5) Palmieri 2006-01-05 01:52:23 UTC 
Created attachment 56793 [details] [review]
fix for crash

Assign array of char * to a char ** and send that to d-bus
Comment 6 Joseph Sacco 2006-01-05 03:13:05 UTC 
Good catch. The patch worked.

-Joseph
Comment 7 Jeffrey Stedfast 2006-01-05 18:12:05 UTC 
gah, missed that. thanks j5