[IronPython] adodbapi works correctly on RC1.
vernondcole at gmail.com
Tue Oct 20 12:25:25 PDT 2009
Do not hold your breath. I understand that there is a problem with unicode
vs str in django. The PEP 429 specifies a BINARY constructor (which within
adodbapi has different definitions for Python2.x and Python3.x) which a user
can call to get a piece of raw memory. Django does not seem to have similar
functionality, and all that will have to be fixed before porting to Python
3. IronPython uses 'str' as a synonym for 'unicode' -- so it is halfway
there, and porting to IPy is a good step in getting ready for Python 3.
Internally in adodbapi I use the symbols 'makeByteBuffer' and
'memoryViewType' when I need to refer to raw bytes. Django will need
something like that.
Also, the version of adodbapi which is used in the django-MSsql project was
forked off before the IronPython support was added. I have communicated with
the authors and will try to make a new version of adodbapi which has all of
the functionality needed for django, so that this fork can be merged in. I
do not know whether the other dbapi modules for IronPython are complete
enough to support django, but I kind of doubt it. The database access code
in django seems to be quite version specific for database api modules.
Again, ADO should be able to replace any of them -- but it will take some
work to get there.
On Tue, Oct 20, 2009 at 12:48 PM, Josh <josh at globalherald.net> wrote:
> "I Py 2.6 RC1 successfully passes all tests for adodbapi. Congratulations!
> You now have a fully PEP 249 compliant database engine on IronPython."
> Excellent. I wonder if this bodes well for Django out of the box on IPy
> 2.6? Is there still an issue of Django's handling of Unicode for binary
> Users mailing list
> Users at lists.ironpython.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Users