[ZODB-Dev] RE: changing PersistentMapping to PersistentDict

[Jeff Sasmor]

Hi Jeremy,

Sorry to chime in so late on this, but if it will break old pickles then won't
this change also break any Zope products which use PersistentMapping?
I personally have a Zope product that a number of people use which
uses this class. 

In the non-zope use of ZODB this may cause a problem for some, but the
number of people who use ZODB this way is probably small. I can
cope with this change since my app using ZODB is unreleased, but
I wonder if this is a bad change if it makes some data stored by some
Zope products no longer accessible.

While its certainly true that the name change is more expressive of
the functionality, if it breaks existing code then IMHO it's not a good
thing.

Pls correct me if I'm wrong...
jms

#--------------------------------
Jeff Sasmor
jeff@sasmor.com

>>>>> "GW" == Greg Ward <gward@mems-exchange.org> writes:

  GW> On 31 July 2002, Jeremy Hylton said: Content-Description:
  GW> message body text
  >> This question didn't get any response on the now-defunct
  >> zodb4-dev.  Any takers here?  Would anyone object if
  >> PersistentMapping was called PersistentDict in ZODB4.

  GW> +1 from me.  Would cause next-to-zero bother for us, since we
  GW> always call a convenience function instead of constructing a
  GW> PersistentMapping directly.

I've made the change, but be warned that old pickles won't work.  They
refer to a newly non-existent PersistentMapping class.

Jeremy