On Mar 28, 8:59 am, William Stein <wst...@gmail.com> wrote: > I know of one place in sage where objects cache their hash for > efficiency reasons (e.g., I think Sage matrices do). I hadn't thought > about the fact that pickling, moving to an object to a platform where > the hashes are different, and unpickling, would result in the subtle > issue above, but that makes sense. Here is an example: > > On a 32-bit platform do this: > > sage: a = matrix(ZZ,2,[1,-2,4,1993938292]); a.set_immutable(); b = > {a:5}; save(b,'/Users/wstein/b.sobj') > > Then load b on a 64-bit platform (e.g. ,sage.math): > > sage: a = matrix(ZZ,2,[1,-2,4,1993938292]); a.set_immutable(); b = > load('b.sobj'); b[a] > boom! KeyError ... > sage: sage: b.keys()[0] == a > True > > The fix here is that right before pickling the cached hash of the > matrix should be deleted. > > I don't know any other places in Sage that do the above.
This also happens (at least) in combinat/; see the __hash__ implementation in combinat/combinat.py > This is now > > http://trac.sagemath.org/sage_trac/ticket/5624 I've mentioned this on the ticket, as well. Carl --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---