* Darren J. Moffat <Darren.Moffat at Sun.COM> [2006-12-14 07:00]:
> The md5sum and sha*sum commands don't clash today but there is overlap
> with the as yet not putback PSARC/2005/530 which actually proposes
> shipping /usr/bin/md5sum as a link to /usr/bin/digest - see the case
> it would produce slightly different output in that case.
>
> I have no problem with shipping /usr/gnu/bin/md5sum but we should
> consider that these particular md5/sha commands to be a possible
> conflict. The other consideration here is that we have put a great
> deal of work in to things like libmd which contain platform and cpu
> optimisations for these hash algorithms, adding the GNU variants would
> add more implementations of the md5/sha code when the goal has been
> reduction. However that isn't too big a concern if it is just
> isolated to the commands but if they are implemented via a library I
> have a big concern - particularly since libmd was designed to be API
> compatible with libmd on other platforms.
Well, this case would be the "existing or anticipated components"
portion of 2.4--good.
These implementations use libm and libc:
$ ldd sha1sum md5sum
sha1sum:
libc.so.1 => /lib/libc.so.1
libm.so.2 => /lib/libm.so.2
md5sum:
libc.so.1 => /lib/libc.so.1
libm.so.2 => /lib/libm.so.2
So I'll move the md5sum and sha*sum commands to /usr/gnu/bin, due to
anticipated conflicts.
- Stephen
--
Stephen Hahn, PhD Solaris Kernel Development, Sun Microsystems
stephen.hahn at sun.com http://blogs.sun.com/sch/
