On Jul 20, 12:58 pm, Willem Jan Palenstijn <w...@usecode.org> wrote: > On Mon, Jul 19, 2010 at 07:54:14PM +0100, David Kirkby wrote: > > I don't know, but suspect so. This page looked interesting, to preload > > a Solaris malloc library with LD_PRELOAD. That seemed to display > > something semi-useful after I inspected the core file with 'mdb' and > > used this 'findleaks command'. There's a lot of pyton and pari stuff > > listed, but sorting the wood from the trees might be hard. > > Memory leaks aren't interesting here; you want to find corruption issues > intead. I have no experience with either libumem or mdb, but some quick > googling shows a ::umem_verify command that may be helpful: > > http://blogs.sun.com/hema/entry/libumem_for_detecting_memory_corruption > > -Willem Jan
Thank you. What is interesting with that is that after I quit Sage (quiting causes a core dump), then using that, it could find no memory leaks, despite there being a core dump. sage: quit Exiting Sage (CPU time 0m0.40s, Wall time 0m4.70s). /rootpool2/local/kirkby/sage-4.5-hacked-for-64-bit-solaris/local/bin/ sage-sage: line 206: 8655 Segmentation Fault (core dumped) sage- ipython "$@" -i kir...@t2:[~/sage-4.5-hacked-for-64-bit-solaris] $ mdb core Loading modules: [ libumem.so.1 libc.so.1 libuutil.so.1 ld.so.1 ] > ::umem_verify Cache Name Addr Cache Integrity umem_magazine_1 10010e028 clean umem_magazine_3 100114028 clean umem_magazine_7 10011a028 clean etc etc (no memory corruption detected). That suggests to me that whatever memory gets corrupted, it was not allocated by malloc. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
