Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!rutgers!mephisto!prism!gt0178a
From: gt0178a@prism.gatech.EDU (BURNS,JIM)
Newsgroups: comp.unix.shell
Subject: Re: command line prompt (need current directory)
Message-ID: <13717@hydra.gatech.EDU>
Date: 16 Sep 90 14:52:47 GMT
References: <1990Sep14.135525.8981@usenet.ins.cwru.edu>
Distribution: comp
Organization: Georgia Institute of Technology
Lines: 23

in article <1990Sep14.135525.8981@usenet.ins.cwru.edu>, chet@cwns1.CWRU.EDU (Chet Ramey) says:
> This will not export the functions to separate invocations of ksh.  Only
> shells `spawned' by ksh to execute scripts will inherit these function
> definitions.  There is no way to export functions to separate invocations
> of the shell; you must do it via the $ENV file.

Actually, I do put my functions in .kshrc. I include the typeset also
because I find vi is funny about what it includes in a subshell, and
wasn't defining my ll() function w/o the typeset. As a matter of fact,
most of what is in my .kshrc is there instead of in .profile, not because
I want to maintain compatibility w/ /bin/sh, but because of vi's quirks.

> The Korn Shell and Bash will do alias substitution first.  It's really the
> last stage of token recognition rather than part of command execution.  For

Sorry, of course you're right. I tested his function (calling it cp) to
be sure of the evaluation order, but then my mind short circuited when I
wrote that builtins executed before aliases as well. :-)
-- 
BURNS,JIM
Georgia Institute of Technology, Box 30178, Atlanta Georgia, 30332
uucp:	  ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt0178a
Internet: gt0178a@prism.gatech.edu