Re: [Fwd: Re: [dvd::rip] dvd::rip 0.97.6 segfaults on FC4 x86_64]

[Dario Castellarin <dacstp@xxxxxxxxxx>]

Your test program seems to work fine, the windows open well.
For your information, these are the perl modules that I've to install 
for dependencies:
perl-Event
perl-Event-RPC
perl-Gtk2-Ex-FormFactory

perl-Glib
perl-Gtk2
perl-IO-Socket-SSL

the first group being from FreshRPMs, the second from Fedora Extras.
Anyway, I've made some additional investigation: after a reinstall, I 
noticed that the segfaults were actually 2 and not one:
tcmodinfo[7094]: segfault at 0000000000000000 rip 0000003f156dc5ef rsp 
00007fffffb427f0 error 4
dvdrip[7057]: segfault at 00000000ae29cf2c rip 0000003c510a014b rsp 
00007ffffffa87c0 error 6

And tcmodinfo is part of transcode afaik. This sounds even weirder since 
I was using the same version of transcode (1.0.2) with dvd::rip 0.52.6 
with no problems at all.
The tcmodinfo segfault happen only on the first run of dvd::rip, after 
there's just the dvdrip segfault as noticed before.

On the rpm release specs I read this:
- Re-enable auto-requires, so perl(Event) is now required.
So it means that "auto-requires" were disabled on previous versions...

How can I test those other perl modules, since perl-Gtk2 seems to work fine?

Thanks,
Dario


Jörn Reder wrote:

> Dario Castellarin wrote:
>
>  
>
> > So by now, I still have no idea on what's happening, I'm switching back 
> > to 0.52.6, waiting for the 0.97 series to become stable, if you need any 
> > further testing just ask.
> >    
>
> I'm pretty sure it's not dvd::rip which causes the segfault here. For 
> Perl programs in general it's nearly impossible to segfault unless they
> use parts written in C. dvd::rip itself is a pure Perl program, but it 
> uses the Gtk2 module, which has of course many C parts. Probably 
> dvd::rip uses other Perl modules with C parts I can't remember yet...
>
> Although the Gtk2 module loads Ok obviously another sub-module causes 
> this segfault, but currently I have no idea how to isolate this.
>
> I just wrote a simple Gtk2 test program opening a window. Probably that
> can reproduce the crash indicating a Gtk2 problem...
>
> Regards,
>
> Joern
>
>