Em 2 de junho de 2011 11:15, Paulo César Pereira de Andrade <paulo.cesar.pereira.de.andr...@gmail.com> escreveu: > 2011/6/2 Francois Bissey <francois.bis...@canterbury.ac.nz>: >>> Em 1 de junho de 2011 18:22, Paulo César Pereira de Andrade >>> >>> <paulo.cesar.pereira.de.andr...@gmail.com> escreveu: >>> > If I understand it correctly, at least for this particular case: >>> > >>> > >>> > sage/libs/singular/groebner_strategy.pyx:GroebnerStrategy(SageObject)._pa >>> > rent => >>> > sage/rings/polynomial/multi_polynomial_libsingular.pyx:MPolynomialRing_l >>> > ibsingular(MPolynomialRing_generic) >>> > >>> > and the _parent field has a __dealloc__ as: >>> > singular_ring_delete(self._ring) while the GroebnerStrategy >>> > __dealloc__ deferences _parent._ring. >>> > >>> > So, again if I understand correctly, the crash is happening because >>> > MPolynomialRing_libsingular:__deallocate__ should be called >>> > after pyx:GroebnerStrategy:__deallocate__, but in python 2.7 >>> > this is not happening, as the order is not guaranteed, and >>> > apparently just happens to be in the correct order with >>> > python 2.6. >>> >>> Sorry for replying to myself, but actually, it should be quite >>> simple to have it functional, just that it must be done in a case >>> by case approach. Something like: >>> >>> GroebnerStrategy(): >>> __new__: >>> ... >>> the_ring_handle = ... >>> incref(the_ring_handle) >>> __del__: >>> ... >>> decref(the_ring_handle) >>> del the_ring_handle ## requirement of this depends on semantics >>> >>> MPolynomialRing_libsingular(): >>> __new__: >>> ... >>> the_ring_ptr = ... >>> __del__: >>> if (refcnt(self) <= 0) ## how to do this in python? check if >>> == 1? or do not use python reference counter? >>> free(the_ring_ptr) >>> >>> This should be easy for python internal experts to >>> implement (I can try to experiment with that, but I do >>> not know much of python and only do mark&sweep :-) >>> >> Quoting Volker: >> Just to clarify, right now Cython extension classes do not call the Python >> destructor __del__ upon finalization, so Sage can't rely on this mechanism. >> I'm >> not quite sure if it is an intentional thing or if the Cython devs just >> haven't gotten around to implementing it. > > I started doing some experiments mostly for the sake of testing :-) > But my python/cython kung-fu is too weak, as well as general > knowledge of the sage "object model" and my experiments > ended in the same backtrace, so I did not work on the correct > point or something, and for now I give up. > >> So we need cython to catch up with that one. > > Cython probably will also not understand the context of > memory being allocated and released outside its control. > The proper approach should be to ensure that python does > not see reachable sage objects as a looping cycle with 0 > references, and then "breaks" the cycle in unpredictable > ways. > > In my basic understanding of sage classes, it should > release things somewhat like: > >>>> del a > ===> > list = <set of objects reachable from a> > while length(list): > foreach item in list: > if item._parent == item: > <blame the user> > item._parent = None > test = map(list, _._parent == item) > if length(test) == 0: > item.__dealloc__() > remove(list, item)
Err, ignore it, this is bogus for some "complex" cycles, and should be the case here, where somewhere down in MPolynomialRing_libsingular it may reference again GroebnerStrategy. I will become silent now :-) Hopefully at least this brought the problem to attention of someone that knows what is going on. >> Francois Paulo -- 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
