[Zope] Monkey patching and inheritance

[Kapil Thangavelu]

a man goes to a doctor and the doctor asks whats the problem, 
'doc it hurts when i do this', and the doc replies 'well don't do that'
and the sends the man home.

iotw, its always better not to monkey patch unless you have no
alternative. userfolders are something that can be replaced soley by
product code and migrated, so using monkeypatching here is unesc.
 
wrt to the specific error, its hard to suggest much given the
information given, surrogates in the core zope are generally used for
database connections registering for transactions. by crash i assume you
mean an exception gets raised.

cheers,

-kapil


On Wed, 2004-01-14 at 16:47, Dennis Allison wrote:
> I am using monkey patches to modify Zope.
> 
> I modify UserFolder.py by providing replacements for a number
> of methods.  These methods are simply defined in an __init__.py
> in a module in the product folder, MyUserFolder.
> 
> The UserFolder class is imported, the new overriding methods defined, and,
> at the end, InitializeClass is imported from Globals, and the patched
> procedure names assigned to their correspondiong  names in Userfolder
> namespace, and InitializeClass() called on the UserFolder.
> 
> The new code is fairly cautious about persistence.  Use is made of 
> PersistentMapping and PersistentList for mutable structures.  
> 
> I've been having occasional crashes of the Zope system with a 
> diagnostic 
> 
>    AttributeError: Surrogate instance has no attribute '_p_oid' 
> 
> which I have finally traced to the UserFolder modifications! 
> Apparently I have somehow broken the persistence mechanism
> in my monkeypatching.
> 
> Advise on what the problem might be and how to repair it would be very
> helpful.
> 
> -dra