[ZCM] [ZC] 2042/ 7 Comment "ExtensionClass misses ZODB persistent reference optimization"

[Collector: Zope Bugs, Features, and Patches ...]

Issue #2042 Update (Comment) "ExtensionClass misses ZODB persistent reference optimization"
 Status Accepted, Zope/feature+solution medium
To followup, visit:
  http://www.zope.org/Collectors/Zope/2042

==============================================================
= Comment - Entry #7 by tseaver on Jun 15, 2006 6:33 pm

Test mail delivery again.  "This time for sure!"
________________________________________
= Comment - Entry #6 by jim on Jun 15, 2006 6:28 pm

more testing
________________________________________
= Comment - Entry #5 by jim on Jun 15, 2006 6:26 pm

Testing mail delivery :)
________________________________________
= Comment - Entry #4 by jim on Jun 15, 2006 6:11 pm

Testing mail delivery :)
________________________________________
= Comment - Entry #3 by jim on Jun 15, 2006 6:10 pm

Testing mail delivery :)
________________________________________
= Assign - Entry #2 by tseaver on Jun 15, 2006 5:53 pm

 Status: Pending => Accepted

 Supporters added: jim

Assinging to Jim for his evaluation (and to test mail delivery).
________________________________________
= Request - Entry #1 by d.maurer on Mar 9, 2006 5:19 am


Uploaded:  "ec_getnewargs.pat"
 - http://www.zope.org/Collectors/Zope/2042/ec_getnewargs.pat/view
The ZODB implements an important optimization for persistent references:
it can include a class reference in a persistent reference
such that a ghost can be instanciated without loading the object.

In order for this optimization to be possible, the class must
not define "__getnewargs__". Unfortunately, "ExtensionClass.Base"
defines it and therefore prevents this optimization for
ExtensionClass instances (i.e. for all Zope2 objects).

The attached patch removes the "__getnewargs__" definition
from "ExtensionClass.Base", changes the "_reduce" definition
accordingly and updates the tests.

==============================================================