>> Can you explain the rationale for dropping the use of '_marker' in
>> favor of 'None'?
>
> As far as I can remember some tests in ZODB explicitly pass (None, None)
> when iterating over the whole structure, bit I will give it a better
> look later.

OK, for instance:

(MappingStorage.py)
"""
     def iterator(self, start=None, end=None):
         for transaction_record in self._transactions.values(start, end):
             yield transaction_record
"""

which ultimately takes us to:

(___BTree.py)
"""
     def _range(self, min=_marker, max=_marker,
                excludemin=False, excludemax=False):
         if min is _marker:
             start = 0
             if excludemin:
                 start = 1
     # ...
"""

somewhere along the way `None` should be translated to `_marker`, but 
that currently doesn't happen, which causes some of the tests to fail.

Cheers,

Pedro
_______________________________________________
For more information about ZODB, see http://zodb.org/

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
https://mail.zope.org/mailman/listinfo/zodb-dev

Reply via email to