On Jan 30, 6:25 am, Jaap Spies <j.sp...@hccnet.nl> wrote:
> mabshoff wrote:
<SNIP>
> > The problem is likely the same as Carl Witty and I tracked down above,
> > i.e. brk() does not extend the heap and Singular goes boom even though
> > it did get the memory. Note that in case of failure the amount
> > allocated seems to be sub-page in size, but that is only speculation
> > at this point if that is really a contributing factor.
>
> Great job! I see this behaviour only on Fedora 9, not on Fedora 10.
Well, it isn't fixed yet, but my guess is that it only happens on FC9
and higher since the way brk randomization is done has changed in the
kernel. It is only a question of time until we will hit the same issue
with other distributions. There is still a remote chance the Fedora's
glibc is buggy, but I would not bet money on that one :)
A trivial workaround for that problem would be to make Singular use
malloc instead of mmap or sbrk, but that causes an unrelated bug to
surface that needs to be fixed anyway, but I am clueless about that
one, too, since it seems to involve scoping issues with Cython/
Python.
> Jaap
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
