[Pkg-corba-devel] progress on the omniorb4 FTBFS on arm

[Floris Bruynooghe]

On Fri, Feb 22, 2008 at 09:41:43PM +0100, Thomas Girard wrote:
> On Fri, Feb 22, 2008 at 06:04:08PM +0000, Floris Bruynooghe wrote:
> > On Sun, Feb 10, 2008 at 04:18:52PM -0000, tgg at users.alioth.debian.org wrote:
> > > +ifeq ($(DEB_HOST_ARCH_CPU),arm)
> > > +  confflags += --disable-alloca
> > >  endif
> > 
> > I've finally managed to find time to get qemu running as an ARM
> > machine.  I tried both before and after this fix and in both cases
> > omnicpp segfaults.
> 
> :-(

Had some more time today.  So far I can't see a difference between the
--disable-alloca flag and not.

It's not actually omnicpp that segfaults, it seems it's python that
segfaults inside the _omniidlmodule.so.4.1 extension module (_omniidl
in python).  So far I have no idea why, it segfaults as soon as it
tries to enter void Prefix::newFile() in
src/tool/omniidl/cxx/idlrepoId.cc.  It doesn't even get to the first
statement of that function.  All I can imagine right now is that
somehow something on the stack goes wrong, but I have no idea how we
can have any influence on that.

I haven't really got much experience with this sort of problem, doing
most of my programming in pure python these days.  But I'll keep
digging around.

Regards
Floris

-- 
Debian GNU/Linux -- The Power of Freedom
www.debian.org | www.gnu.org | www.kernel.org