Re: rln, a potential new core util

[Paul Eggert]

Shigio Yamaguchi proposed something similar four years ago.  In his
approach, "ln -a -s" always creates an absolute symlink, even if the
source is relative, and "ln -r -s" always creates a relative symlink,
even if the source is absolute.

He publishes the complete source for such an ln implementation (based
on BSD ln) in his PathConvert project
<http://tamacom.com/pathconvert.html>.  This has not found its way
into the BSDs.  He also proposed a patch along these lines for GNU ln in
<http://lists.gnu.org/archive/html/bug-fileutils/2002-03/msg00093.html>
but it didn't get accepted.

I sort of like the idea myself.  I guess.  (Possibly because I haven't
thought through all the consequences....)