On 10/19/07, Gerfried Fuchs <[EMAIL PROTECTED]> wrote:
>
> * Filip Van Raemdonck <[EMAIL PROTECTED]> [2007-10-19 09:50:21 CEST]:
> > On 10/18/07, Gerfried Fuchs <[EMAIL PROTECTED]> wrote:
> > > That said, could you pretty please look if your problem still exists in
> > > the experimental release of the campaign?
> >
> > I searched around and found a 1.2-1 version on p.d.o in Andreas
> > Tille's webspace. I also grabbed the 1.3.9-3 version from
> > experimental.
>
> > With the unstable 1:1.2.7-2 version of wesnoth{,-data} both crash, but
> > no longer emit the crash message from the old campaign package.
>
> The unstable version 1:1.2.7-2 doesn't has the -sotbe campaign, so I
> wonder what version of the campaign you refer to here that crashes with
> it.
"both" -- i.e. Andreas' -sotbe package as well as experimental's -sotbe package.
> > However, the crash should not happen this ugly -- wesnoth should
> > rather say it can't load the saved game because of whatever the issue
> > is. (and if we can figure that out, maybe it's possible to fix the
> > campaign so I can continue :)
>
> I'll forward your problem to upstream, and from your report I read that
> you used 1.2.7 of the wesnoth packages, right?
Right. That is, of binaries and -data and whatever campaigns are
currently in unstable.
And misc versions of the -sotbe campaign.
> And about fixing the campaign, I see no problem with the campaign, just
> with your mixture of packages. Propably I should narrow the dependencies
> quite a bit to not allow a mixture of versions between campaigns and
> main program.
I agree that the problem about not being able to run -sotbe is the
versions that are off.
I disagree that stricter versioning is the end solution. Consider
this: a (malicious?) external party finds a way to reproduce and
exploit the segfault through a custom scenario (or campaign), and
starts distributing said exploit.
Unless wesnoth is fixed to handle the issue that causes the segfault
graciously, like with an error message "oops can't start this scenario
somethings broken with it" you now have a security problem.
KR,
Filip
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
