Re: OBJ_* ASN1_OBJECT functions

On Fri, Jul 13, 2007 at 10:50:28AM -0700, Heikki Toivonen wrote:
> It would be great if we didn't need to worry about backwards
> compatibility. I don't think it can be put into M2Crypto at this point.
This interface may be built on top of M2 as an extansion. 
There is no need to replace current code. I'd suggested it as a
subject to discuss :) So correct subject would be 
"[RFC] Native python representation of Name/Name Entry objects"

> (It does have one minor weakness, in that tuple is not as
> self-documenting as a class.)
Tuples have weak points such as hard constructor overloading and
immutability. But they are extremly simple. Name Entry may be
implemented as Python class with oid/data fields and function
that convers this native python representation to libssl objects.

Major point here is that all manipulations are performed in python
and only interaction with other OpenSSL code require conversion.

I'd mentioned reason why it may be useful - new functions may 
be implemented as 'addons' and not library modification involving
SWIG.

				Pavel