Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site pur-ee.UUCP
Path: utzoo!linus!philabs!cmcl2!floyd!harpo!ihnp4!inuxc!pur-ee!ecn-ee!ks
From: ks@ecn-ee.UUCP
Newsgroups: net.bugs.4bsd
Subject: Re: 4.xBSD /bin/csh bug fix wanted.. - (nf)
Message-ID: <1569@pur-ee.UUCP>
Date: Fri, 10-Feb-84 21:35:40 EST
Article-I.D.: pur-ee.1569
Posted: Fri Feb 10 21:35:40 1984
Date-Received: Sat, 11-Feb-84 23:37:18 EST
Sender: notes@pur-ee.UUCP
Organization: Electrical Engineering Department , Purdue University
Lines: 26

#R:ecn-ee:16700002:ecn-ee:16700004:000:659
ecn-ee!ks    Feb 10 00:28:00 1984

Well, the output of tset -s looks something like:

set noglob;
setenv TERM adm3a ;
setenv TERMCAP 'la|adm3a|3a:am:cr=^M:do=^J:nl=^J:bl=^G:le=^H:bs:cm=\E=%+\040%+\040:cl=1^Z:co#80:ho=^^:li#24:ma=^K^P:nd=^L:up=^K:';
unset noglob;

noglob should be set while doing this.
Something is screwed up because it is in a line like

eval `tset -s`

The peculiarities:
	It works as long as there are no '*'s in the TERMCAP (maybe '?'s too).
	It fails to work every second time you try it.

The two suspect csh mechanisms are `back-quoting` and "eval".
My best guess is "eval" is messed up...
I really haven't dug into the sources (yet)...

					Kirk Smith
					Purdue EE