Subject: Re: Proposed rc.d changes....
To: Todd Vierling <>
From: Greywolf <>
List: tech-userlevel
Date: 02/21/2001 08:13:04
On Wed, 21 Feb 2001, Todd Vierling wrote:

# On Wed, 21 Feb 2001, Greywolf wrote:
# : # converting the startup scripts to use that shell instead of the
# : # legacy sh would have the immediate benefits that some of the
# : # problems which are currently being discussed would be addressed
# : # right away (like regular-expression pattern-matching with builtins,
# : # avoidance of external commands for better performance, etc.)
# :
# : I will personally shoot on sight the person who converts these to ksh.
# : Wading thru ksh kode is the _worst_ pain in the ass.
# Using ksh doesn't necessarily imply writing shell functions and such using
# the ksh syntax.  It *is* Bourne shell compatible, after all.

It's not the functions I object to (bourne shell has them, after all).
I've had the nightmare of wandering through scripts which use every single
feature that ksh has and porting it to Perl (all the while fighting with a
dolt who insisted that messages and code did not have to be formatted to
80 (or fewer) columns wide (line breaks were OK)...his coding style, while
technically correct, was also, stylistically and syntactically, enough to
send one in to anapleptic/epileptic shock (roll a die).  But I digress).

I, for one, do not want to see every single feeping creature from ksh
make its way into our startup scripts.

I think the argument was that case was more efficient/flexible than test,
even on a builtin level, and I'm wont to agree with this.

# -- Todd Vierling <>  *  Wasabi NetBSD:  Run with it.
# -- NetBSD 1.5 now available on CD-ROM  --

*BSD: We Suck Less