sorry for the poorly worded question (midnight madness), let me try and restate them a bit better in the fond hope that someone will answer any of them. whats is a jar? as far as i know its just the data representation of the object that gets stored in the zodb, hence its directly related to the __dict__ of the object. but then what is the signifigance of calling methods directly on the jar like i see in some of the transaction machinery, is this an indirect call to the thread's zodb connection?

does zope restrict you from doing metaprogramming with python products? or does zope's own metaprogramming reserve some names __klass__, etc?

i was reading through guido's essays yesterday, and happened upon the metaprogramming essay ( i read it when i was first learning python but it made a lot more sense now:). i was thinking about creating my own caching metaclass, but i know that acquisition (and probably persistence, and zclasses) use some c-based metaclasses to work. i'm just curious what the possibilites and restrictions are of doing metaclasses in zope, so my classes play well with others. in /lib/python/Shared/DC/ZRDB/TM.py, i see the class Surrogate defining methods for __inform_commit__ and __inform_abort__, what are these for? In ZPatterns/Transaction.py I see some talk of transaction boundaries, namely the inability for an object participating in the transaction to know that a transaction is finished. From my understanding of transactions i thought that an object could assume a transaction was over if it has its tpc_finish called on it when its not in a subtransaction. I understand that the Zpatterns problem here might be related to providing attrs to other objects which might be used in a tpc_finish but i just wanted to clarify/make sure. kapil