[MPlayer-dev-eng] Zattoo, AES, avutil

[Ilja Kamps]

On Mon, Mar 12, 2007 at 07:02:49PM +0100, Michael Niedermayer wrote:
> Hi
> 
> On Mon, Mar 12, 2007 at 05:11:41PM +0100, Ilja Kamps wrote:
> > On Sun, Mar 11, 2007 at 10:03:16PM +0100, Michael Niedermayer wrote:
> > > Hi
> > > 
> > > On Sun, Mar 11, 2007 at 09:16:06PM +0100, Reimar D?ffinger wrote:
> > > > Hello,
> > > > On Sun, Mar 11, 2007 at 08:58:34PM +0100, Michael Niedermayer wrote:
> > > > > On Sun, Mar 11, 2007 at 05:31:47PM +0100, Vladimir Serbinenko wrote:
> > > > > > It's problematic to use AES from libavutil because it doesn't support
> > > > > > padding and padding is necessary for zattoo
> > > > > 
> > > > > well so add 3 lines of c code to your zatoo code to pad the last
> > > > > block
> > > > 
> > > > He is talking about OEAP, for which there only seems to be only the overly
> > > > complicated and incorrect mathematical paper http://www-cse.ucsd.edu/users/mihir/papers/oae.pdf
> > > > and some similarly overly complicated code as documentation.
> > > 
> > > hmm, then why dont you (or the author of the patch) explains it in a 
> > > few words?
> > > 
> > > 
> > > > It at least needs a SHA-implementation though.
> > > 
> > > so from all the tiny fragments of information on the ML it seems we need 
> > > a SHA-1 and a RSA implementation?
> > > which then somehow get connected ...
> > >
> > 
> > SHA1 isn't much, see attached file
> 
> RSA isnt more either ...
> 
> anyway both RSA and SHA-1 should be added to libavutil, but these must be
> clean, small, simple and fast implementations ...
> 
> the one below fails in clean and small but can certainly serve as a starting
> point for a nicer implementation
> the most obvious problem it has is the use of macros and manually unroled
> loops instead of inline functions and normal for()
> the rest is simple plain stupidity, like writing to variables which are
> never read afterwards and doing otherwise silly things ...
> 
> [...]
>

Agreed, it could take some work to clean up further, I just did it to
the point where I could actually use it in another project, I actually
have a public domain RSA implementation aswell somewhere, but I couldn't 
find it.

Actually I think all this crypto stuff is also in gnutls or some
similair GNU project, so you could use that or code from that.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20070312/1af95a2d/attachment.pgp>