Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!BRAUER.HARVARD.EDU!primer
From: primer@BRAUER.HARVARD.EDU (Jeremy Primer)
Newsgroups: gnu.emacs.bug
Subject: absolute paths in GNU emacs (18.55-berkeley unix)
Message-ID: <9002240247.AA13575@brauer.harvard.edu>
Date: 24 Feb 90 02:47:17 GMT
Sender: daemon@tut.cis.ohio-state.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 24

It annoyed me somewhat that putting

(setq load-path (cons "~/emacs/lisp" load-path))

in my .emacs file failed to cause emacs to search my own lisp directory for
elisp code.  Reason: emacs only likes to use absolute paths internally,
objected to the use of "~" in a directory name.  In short, I became the
"luser" in the function "openp()" in lread.c.

I can live with that.  If emacs wants its internal paths to be absolute,
then that makes sense to me.  What bothered me much more was that the same
treatment was given to paths specified in the environment, i.e.

setenv EMACSLOADPATH "~/emacs/lisp:/usr/local/lib/emacs/lisp"

also luses for the same reason.  This is bad, since people have been
specifying relative paths in their .login's for years.  Please hack
decode_env_path() to expand relative paths in the next release.

Or, alternatively, if I am mistaken and emacs is supposed to deal with
relative paths internally, then I guess openp() needs fixing.

Jeremy Primer
primer@math.harvard.edu