On Fri, Mar 13, 2009 at 12:55 PM, Nicolas M. Thiery <nicolas.thi...@u-psud.fr> wrote: > > On Fri, Mar 13, 2009 at 11:37:34AM -0700, Carl Witty wrote: >> >> On Fri, Mar 13, 2009 at 1:03 AM, Nicolas M. Thiery >> <nicolas.thi...@u-psud.fr> wrote: >> > But then, a second problem: the code of cPickle (around line 1977) >> > tries to lookup the attribute "Bla.Ble" in __main__, instead of >> > successively looking up "Bla" in __main__, and then "Ble" in the result: >> >> May I suggest a different workaround? (It's gross, but maybe less >> gross than changing cPickle.) >> >> sage: __main__.__dict__['Bla.Ble'] = Bla.Ble >> sage: loads(dumps(x)) >> <__main__.Bla.Ble instance at 0x1a2e7eac> > > Well, that's a bug in python that should eventually be fixed in > python. But yeah, your workaround could be a good way out.
I certainly wouldn't call it a bug -- at the most, a missing feature. I'm not even convinced that a patched version would be an improvement. (For one thing, you would need to also patch the unpickle side, which means that you can't write a pickle using your patch and read it in older versions of Python. We don't care much about that in Sage, but I think it's important to the Python people.) Carl --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
