Hi Dave,
I had to make slight changes to your version of SQLAlchemy adapter to
return the correct results for my project. The patch is:
----------->8------------- _sqlalchemy.patch --------------
>8---------------
--- /home/papagr/_sqlalchemy_old.py 2008-09-18 17:29:06.000000000 +0200
+++ /home/papagr/_sqlalchemy_new.py 2008-09-18 17:29:20.000000000 +0200
@@ -17,14 +17,15 @@
return pyamf.util.native_get_attrs(obj)
attrs = {}
- for property in obj.__dict__:
- if property in sa_properties:
+ mapper = obj._sa_class_manager.mapper
+ for property in mapper.iterate_properties:
+ if property.key in sa_properties:
continue
- if property.startswith('__'):
+ if property.key.startswith('__'):
continue
- attrs[property] = getattr(obj, property)
+ attrs[property.key] = getattr(obj, property.key)
return attrs
def write_SA_collection(obj):
----------->8------------- _sqlalchemy.patch --------------
>8---------------
The reason for using object's mapper to iterate properties instead of
fetching object's __dict__ is because SQLAlchemy initially creates
_CompileOnAttr except if you manually call mapper's compile method.
Expecting your comments,
Cheers,
Nikos.
_______________________________________________
PyAMF dev mailing list -
dev@...
http://lists.pyamf.org/mailman/listinfo/dev
