Ticket #5419 (new defect)
moving fraction field to new coercion broke old pickles
| Reported by: | craigcitro | Owned by: | craigcitro |
|---|---|---|---|
| Priority: | major | Milestone: | sage-5.10 |
| Component: | misc | Keywords: | |
| Cc: | burcin | Work issues: | |
| Report Upstream: | Reviewers: | ||
| Authors: | Merged in: | ||
| Dependencies: | Stopgaps: |
Description
At some point, fraction fields were moved over to the new coercion model, which is good -- except that it broke all the old pickles. This thread on sage-support is about someone having a problem with them: http://groups.google.com/group/sage-support/browse_thread/thread/b5519db45a141819
The problem is that the old pickles don't have _element_class or _element_constructor fields, and there was no factory function in place -- so unpickling tries to directly create the object, which totally fails. Putting the old __call__ method back in place is an ugly hack to get these to load, but it's not a good permanent solution.
Change History
Note: See
TracTickets for help on using
tickets.

Hmm,
how did this escape the pickle jar doctest?
Cheers,
Michael