New Python block cipher API, comments wanted

[Paul Rubin]

Nick Vargish <nav at adams.patriot.net> writes:
> > It's not a "distribution".  It's not really even "software".  It's
> > code intended to be looked at, not used for real data.  
> 
> Oh, okay, sorry. I thought you were trying to get it into the 2.3
> release, which made me think it was intended for distribution and
> production use.

I do want to get a package into the 2.3 release.  The thing I posted
is a prototype of what that package is supposed to look like for the
user.  There's some urgency to this because I'd like to get a useable
test version (which this isn't) into 2.3a2 (which is going out very
soon).  So I put what I have out for for comments, even in its current
rough form.  The package I want to have in the final 2.3 release will
be written as a C extension and be fast enough for bulk operations in
production systems.

BTW, Bryan Olson kindly supplied me with a version of AES written in
pure Python, which is quite an elegant piece of code, and fast enough
to be useful for some limited purposes (on my p3-750 it encrypts a few
hundred blocks per second, as compared to around 1 block/sec for the
original Python reference implementation).  So I'll incorporate
Bryan's code into my test package, and will attempt some NIST test
vectors.

Anyway, I've gotten some helpful comments; more comments are of course
still solicited.

Thanks.

Paul