Re: Hash Function Availability (sha / md5)

[Mike Frysinger <vapier at gentoo dot org>]

On 04 Nov 2016 13:14, Florian Weimer wrote:
> On 11/04/2016 02:11 AM, Patrick Herbst wrote:
> > The glibc crypt function uses sha and/or md5, but it doesn't appear
> > these routines are in a published header for application programming.
> 
> That's the intent, yes.
> 
> > Am I wrong on this?  Are there headers/libraries that are part of
> > glibc that offer these hash functions natively?
> >
> > If not, why?
> 
> The hash functions are an internal implementation detail.  The 
> implementations we have today are not optimized and do not use hardware 
> acceleration (except on sparc64).
> 
> I know that there is no compelling or canonical source of hash functions 
> right now.  All existing cryptographic libraries are weird in some way, 
> perhaps with the exception of GNUTLS (but it's way too large if you just 
> need a hash function).

fwiw, libmd attempts to fill that lightweight gap:
https://www.hadrons.org/software/libmd/
-mike

Attachment: signature.asc
Description: Digital signature