Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com From: "Robert A McDougall" Organization: Agricultural Economics-Purdue To: cygwin AT cygwin DOT com Date: Wed, 29 Aug 2001 10:56:57 EST MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: teTeX mktexlsr problem Message-ID: <3B8CCA78.31303.4B400B@localhost> X-mailer: Pegasus Mail for Win32 (v3.12c) mktexlsr misbehaves under Cygwin when $CYGWIN is unset or of zero length. Typical symptoms are that the web2c filename databases don't get made, and TeX, LaTeX etc. can't find their "fmt" files. The error is in the test mktexlsr uses to decide whether to use Un*x- or W*nd*ws-style path separators: if test -z "$COMSPEC" && test -z "$ComSpec" || test -n "$CYGWIN"; then SEP=':' else SEP=';' fi This is wrong because it's possible (and reasonable) to use Cygwin without setting the $CYGWIN variable. Better is: if test -z "$COMSPEC" && test -z "$ComSpec" || test "x$OSTYPE" = "xcygwin"; then SEP=':' else SEP=';' fi From the mail archive, it seems likely that this was William D. Kirby's problem (http://sources.redhat.com/ml/cygwin/2001-04/msg00471.html). Also that Kenta MURATA had the same problem, and proposed an alternative solution, namely, unset $COMSPEC (http://sources.redhat.com/ml/cygwin/2001-04/msg00476.html). -- robert mcdougall . center for global trade analysis -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/