Re: GNU Octave 3.6.3 Released

[Sergei Steshenko]

--- On Wed, 9/5/12, Jordi Gutiérrez Hermoso <address@hidden> wrote:

> From: Jordi Gutiérrez Hermoso <address@hidden>
> Subject: Re: GNU Octave 3.6.3 Released
> To: "Sergei Steshenko" <address@hidden>
> Cc: address@hidden, "C.H.M. Broeders" <address@hidden>
> Date: Wednesday, September 5, 2012, 12:35 PM
> On 5 September 2012 15:26, Sergei
> Steshenko <address@hidden>
> wrote:
> >  liboctave/Sparse.cc
> ....................................panic: Segmentation
> fault -- stopping myself...
> 
> I suppose it's futile to ask you for a stack trace, right?
> 
> > Did the developers run the full build session just
> before the release ?
> 
> No, but the problem seems intermittent. It doesn't show in
> my system
> nor on others'. There are memory problems with valgrind, but
> make
> check doesn't segfault for me.
> 
> - Jordi G. H.
> 

"No" - Jordi, as I stated on various occasions, Octave development model is 
_broken_.

Regarding stack trace - I do not build Octave manually. If you tell me a "one 
button" solution, i.e. a sequence of command lines assuming that Octave tarball 
produced

/home/qemu/AFSWD/build/octave-3.6.3/

directory when unpacked (that's the location in my case), I can try.

...

By the way, I think Octave tarball for such occasions should contain a script 
which assumes 'gdb' is in $PATH and that can be called something like

/path/to/produce_stack_trace /path/to/segfaulting_test.m

.

...

My 'gcc' is gcc-4.5.3, 32 bit Linux OpenSuse-11.4, but it doesn't matter since 
practically all he libraries are self-built.

Regards,
  Sergei.