Hi! At #15247, I am introducing a class SingletonClass: If you write a class and add SingletonClass to the list of bases, then your class will have a unique instance plus fast comparison and hash. We have a similar thing already, namely sage.categories.category_singleton.Category_singleton.
It would be natural to apply this to RationalField: There is a unique instance, namely QQ. In fact, using #15247, I get a considerable speed-up of common operations such as "RationalField()" (this used to be particularly slow, since it used to call RationalField.__init__ *again* on the unique instance QQ, which is of course bad!), "hash(QQ)" and "QQ==ZZ". Problem: It breaks the pickle jar, and I don't understand why this is happening. Namely, in vanilla sage-5.12, QQ.__reduce__() returns (RationalField,()), which is all that is needed to correctly unpickle QQ. In fact I tested that a pickle of QQ or 1/2 created with vanilla sage-5.12 can be correctly unpickled after applying the branch from #15247. However, sage.structure.sage_object.unpickle_all() fails. It seems that some old pickles do not store QQ via (RationalField,()), but somehow result in calling RationalField.__new__(RationalField). This used to work, since RationalField had a custom __new__ method that was responsible for the unique instance behaviour in the past. So, can you please explain to me what I can do here? Is there a way to use unpickle_override to replace the call to RationalField.__new__(RationalField) by RationalField.__classcall__(RationalField)? Do you have a different idea to solve the problem? Best regards, Simon -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
