Bug#443822: I've found a fix (for: hex-a-hop: Got bt from a segfault :D)

[Jens Seidel]

On Fri, Nov 21, 2008 at 01:03:21PM +0100, Gerfried Fuchs wrote:
> * Jens Seidel <jensseidel at users.sf.net> [2007-12-17 10:49:49 CET]:
> > tags 443822 +pending
> > thanks
> > 
> > On Sun, Dec 16, 2007 at 02:11:52PM +0100, Roland Clobus wrote:
> > > The attached patch is written on the fully patched source code.
> > 
> > > --- hex_puzzzle.cpp	2007-12-16 13:19:47.000000000 +0100
> > > +++ hex_puzzle.fixed.cpp	2007-12-16 13:19:05.000000000 +0100
> > > @@ -602,6 +602,9 @@
> > >  		{
> > >  			while (numStages > 0 && time[numStages-1] >= t)
> > >  				numStages--;
> > > +			if (currentStage > 0 && currentStage >= numStages) {
> > > +				currentStage = numStages - 1;
> > > +			}
> > 
> > Is this also save if numStages==0? I will try to check it during the
> > next weekend and play some levels.
> > 
> > Thanks again for the help and expect an upload after Christmas. I
> > suggest to build now a new package from trunk and to test it ...
> 
>  Hmm, have you checked? The segfault still happens to me regularly, and
> strace isn't really helpful here, even gdb is acting strange to me:

Sorry, I do not remember it well. At least it seems that I cared about the
case numStages==0 (according to the bug log).

Clearly this package needs more care which requires more time. At least I
will try to increase both soon but I cannot promise it.
 
> gdb> bt
> #0  0x0fbd68dc in ?? () from /lib/libc.so.6
> #1  0x0fbd5a20 in memmove () from /lib/libc.so.6
> #2  0x00000000 in ?? ()
> 
>  As I'm not too familiar with gdb I really would appreciate some help to
> debug this further and offer more helpful informations...

This happens on PowerPC, right? I have access to i386, mips, mipsel only.
I'm sure valgrind would help (valgrind hex-a-hop) but this doesn't support
exotic architectures and I could not reproduce this bug with it in the past.

Maybe a memory checker such as efence would help as well, not sure.

Jens