On Aug 22, 2:16 pm, William Stein <wst...@gmail.com> wrote: > Nils -- why don't you build everything in your (presumed) account on > sage.math.washington.edu? Then you can easily given Juanjo access to that, > since he also has an account on sage.math.washington.edu.
Done. The following is now available on sage.math: /scratch/nbruin/sage-4.1.1-ecl/ --- a working sage tree with an ecl- library extension /home/nbruin/sage-4.1.1-ecl-x86_64-Linux.tar.gz --- the above tree "- bdist"-ed /home/nbruin/ecl-crash.sage ---- a script to effect the crash /home/nbruin/use-ecl-maxima.sage ---- a script demoing interfacing with maxima the crash is completely predictable and reproducible. This does the trick: /scratch/nbruin/sage-4.1.1-ecl/sage < /home/nbruin/ecl-crash.sage I'm looking forward to any insights people might have. In fact, I think I already know the cause of the error: the GMP library does have some global state in the form of mp_set_memory_functions. cl_boot does call that via init_number to set ECLs memory allocation routines. Of course, sage uses GMP as well, and allocates some integers right from the start. So it's no surprise at all that integer_free_integer_pool seems to call ecl_dealloc, because it does (via some gmp macro for deallocating a gmp integer), and since that integer was allocated by another memory manager, a segfault is entirely expected. I think this is a complete showstopper for including ecl as a library. Should we call mp_set_memory_functions before and after each call to ecl? And then just be very hygienic about which gmp integer goes where? Not letting ecl set its memory management for gmp would not be an option, because only ecl knows if the memory can be freed. I thought the fact that ecl also using gmp would make it very compatible with sage. Turns out it's a major stumbling block. --- Concerning the safe-eval: Ah, of course. You are just catching the error on lisp-level. Now I can write my own error handler if that's necessary. Concerning signals: Is it necessary for ECL to be notified of signals? I think it's good if sage gets them, so "sigaction" might be the better way to go. --~--~---------~--~----~------------~-------~--~----~ 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 URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
