[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

RE: ITP libmd5 and libmd5-dev


On 06-Feb-2000 Russell Coker wrote:
> I've just spend too much time (again) trying to work out how to implement MD5
> in my C programs.
> Currently it seems that the only way to write MD5 programs is to grab the
> source of something that does MD5 (md5sum or any RADIUS server), grab the
> md5.{c,h} from that and then try and figure them out.  There seems to be no
> documentation for this and there are no decent source code comments or good
> header files.

There is: 

1. There is some source in glibc which has MD5 but I do not think that it is
very generic.

2. Check OpenBSD sources. 

Contrarily to most of the the sources floating around (skey for example) which
are derived from the RSA implementation and have endian/64 bit problems this one
has gone through a complete 64 bit cleannes overhaul. Contains MD2,
MD4, MD5 and RIPE-160.

Cheers ;-)

> I propose to create man pages, a header file that can be included in C or C++
> programs, and maybe a -dev package.
> Is it worth making it a shared object?  As an object file created with
> gcc -c -O2 md5.c
> is only 3484 bytes in size should I just have the md5.o file in the package
> and have programs statically link (which means they won't require a
> dependency on libmd5).  If I do that should I just have a libmd5-dev package
> with no corresponding libmd5?


> My reason for wanting to do this is that the pain that I've gone through
> twice now to work out how to calculate md5's is greater than that which would
> occur from writing a proper Debian package once.

Cheers ;-)

- ----------------------------------
Anton R. Ivanov
IP Engineer Level3 Communications
RIPE: ARI2-RIPE      E-Mail: Anton Ivanov <aivanov@eu.level3.net>
@*** Golden Rules of Indulgence  (No 3 of 4)***
      Moderation is for monks.

- ----------------------------------
Version: GnuPG v1.0.0 (GNU/Linux)
Comment: For info see http://www.gnupg.org


Reply to: