Crypto timings

[Eric S. Raymond]

Hal Murray <hmurray at megapathdsl.net>:
> There shouldn't be any MD5 code.  Both MD5 and SHA1 should use a generic 
> routing to calculate the digest.

They both do, now that everyting goes through OpenSSL.  What I meant by the MD5
code was the calls in a_md5encrypt.c, as opposed to the ones in ntp_leapsec.c.

> I copied some code to make the timings.  It took me longer than expected to 
> find the code I wanted.  I think part of the confusion was that some routines 
> had MD5 in their names when they actually weren't MD5 specific.  Mumble.  I 
> think that area could be cleaned up.

/me looks

Oh dear Goddess you are right.  I think I noticed that before but spaced it.
I'll fix that up once I've had some sleep.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>