Summary: | gnome-base/nautilus and gtk file picker segfault if blank cd is in cd drive | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Michael Arntzenius <daekharel> |
Component: | [OLD] GNOME | Assignee: | Gentoo Linux Gnome Desktop Team <gnome> |
Status: | RESOLVED NEEDINFO | ||
Severity: | major | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
gdb backtrace of the nautilus crash
backtrace after recompilation -O2 |
Description
Michael Arntzenius
2008-06-30 16:41:48 UTC
Created attachment 158977 [details]
gdb backtrace of the nautilus crash
erm, worksforme. The backstrace seems to be missing some important elements so you'd probably want to rebuild more packages with debugging symbols. Please also try to compile glib/gvfs/nautilus with -O2 not -O0 (I know this looks silly but just try it) Created attachment 159431 [details]
backtrace after recompilation -O2
Here's the backtrace after I recompiled glib/gvfs/nautilus with -O2.
The last call in the stack trace is at address 0x0. This is what happens if you call a function pointer equal to 0 (ie NULL), so I'm not sure it's indicative of something "missing" from the backtrace. And, sure enough, gfile.c:4013 calls through a function pointer:
static void
open_read_async_thread (GSimpleAsyncResult *res,
GObject *object,
GCancellable *cancellable)
{
GFileIface *iface;
GFileInputStream *stream;
GError *error = NULL;
iface = G_FILE_GET_IFACE (object);
// this is line 4013
stream = iface->read_fn (G_FILE (object), cancellable, &error); // and here's the function pointer call
if (stream == NULL)
{
g_simple_async_result_set_from_error (res, error);
g_error_free (error);
}
else
g_simple_async_result_set_op_res_gpointer (res, stream, g_object_unref);
}
If the backtrace is still not good enough, what other packages should I compile with debugging enabled to get a better one?
Ouch sorry for not getting back to you sooner. It seems that the issue is in gvfs if I got you right. Does it still happen with gvfs-1.0.3 or latest 0.2.5 in tree ? Could you report it upstream if so and paste the link here ? please get back to us. |