problem in failures handling in idLift

14 views
Skip to first unread message

Charles Bouillaguet

unread,
Jan 1, 2013, 2:20:21 PM1/1/13
to libsingu...@googlegroups.com
Hi libsingular-devel,

I would like to draw your attention to http://trac.sagemath.org/sage_trac/ticket/13714

I couldn't fix this by myself. The problem is that whenever idLift is called with a polynomial that doesn't belong to the ideal, an error is raised and we must not deal with it properly. It makes subsequent calls to primdecSY crash.

In older Sage releases, we did not even reset the "errorreported" variable. Now we do, but apparently this is not enough.

Any help would be appreciated...

Cheers,
--
Charles Bouillaguet

han...@mathematik.uni-kl.de

unread,
Jan 30, 2013, 4:03:51 AM1/30/13
to libsingu...@googlegroups.com
Additional to "errorreported" also the variable "libfac_interruptflag"
needs to be reset to 0 (otherwise all calls to routines from libfac
will immediately return - which happens for example in primdecSY).
See kernel/febase.cc, line 943 ff for what happens in an error case
and needs to be reset:

errorreported = TRUE;
#ifdef HAVE_FACTORY
// libfac:
extern int libfac_interruptflag;
libfac_interruptflag=1;
#endif

Hans Schoenemann

Charles Bouillaguet

unread,
Jan 31, 2013, 4:37:13 AM1/31/13
to libsingu...@googlegroups.com
Thanks Hans,

I could fix the bug I had in Sage with this extra piece of information :)

Cheers,
---
Charles Bouillaguet
http://www.lifl.fr/~bouillaguet/
> --
> You received this message because you are subscribed to the Google Groups "libsingular-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to libsingular-de...@googlegroups.com.
> To post to this group, send email to libsingu...@googlegroups.com.
> Visit this group at http://groups.google.com/group/libsingular-devel?hl=en.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

Reply all
Reply to author
Forward
0 new messages