Changes between Version 1 and Version 2 of Ticket #9265, comment 24


Ignore:
Timestamp:
Aug 15, 2012, 1:20:39 PM (10 years ago)
Author:
Andrew Mathas
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #9265, comment 24

    v1 v2  
    11Thanks Anne!
    22
    3 The pickling error confused me no end:) It seems that sage keeps pickles of old objects and then checks that new code is still able to unpickle the saved pickles. One way to fix this error would be be to make a new pickle jar which would remove the references to these renamed classes and presumably solve this problem. Another option would be to include deprecation warnings for all of the old class names, which would probably lead to doctest warnings instead of failures when this test is run(?) -- I am not sure but would this lead to deprecation warnings during the unpickling meaning that these tests would still fail?
     3The pickling error confused me no end:) It seems that sage keeps pickles of old objects and then checks that new code is still able to unpickle the saved pickles. One way to fix this error would be be to make a new pickle jar which would remove the references to these renamed classes and presumably solve this problem. Another option would be to include deprecation warnings for all of the old class names -- I am not sure but would this lead to deprecation warnings during the unpickling meaning that these tests would still fail?
    44
    55I am happy to deprecate all of the old class name if you like, although it does seem a little strange to deprecate classes that were never officially part of sage. Other the other hand, the patch has been around for a while so people may be using the old names in their own code, so it might be worth doing because of this.
    66
    7 I am happy to refer to your expertise as to what is the best course of action. Please advise.
     7I am happy to defer to your expertise as to what is the best course of action. Please advise.
    88
    99The doc string error is also a little strange. It is caused by a T'_a in the doc string, but as this is part of a (raw) string I would have thought that sphinx would not have a problem with this... Anyway, I have fixed this by changing T'_a into S_a etc which is more readable anyway.