Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!sdd.hp.com!elroy.jpl.nasa.gov!jpl-devvax!lwall
From: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Newsgroups: comp.lang.perl
Subject: Re: Is this a bug?
Message-ID: <9105@jpl-devvax.JPL.NASA.GOV>
Date: 10 Aug 90 22:34:55 GMT
References: <1990Aug9.155120.2703@uvaarpa.Virginia.EDU>
Reply-To: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 40

In article <1990Aug9.155120.2703@uvaarpa.Virginia.EDU> worley@compass.com writes:
: I tried to write a program with the following regexp:
: 
: 	/(^\s*$)|(^---)/
: 
: That is, match any line containing only whitespace, or beginning with
: '---'.  (Are ^ and $ allowed other than at the beginning or end of the
: regexp?)  Perl gives the strange error message:
: 
: 	/(^\s*|(^---)/: unmatched () in regexp at ss line 3.
: 
: Where did the missing ')' go?
: 
: Actually, it was probably assumed to be part of a '$)' variable.  (Can
: one use '$/' as a variable is a regexp?)
: 
: What is going on here?  What *should* be going on here?

It was misinterpreting $) in patterns as a variable.  At patchlevel 27
it's interpreted correctly as an end of line check and a terminating paren.
Which means you can't interpolate $) into a pattern directly.

$/ has never been a problem.

By the way, it's more efficient to factor out the ^ to the front:

	/^(\s*$|---)/

The reason for this is that it then knows it doesn't have to start looking
at every single position of the input string.  I suppose I should make it
do this optimization itself...

It's probably also faster to put the literal string before the *:

	/^(---|\s*$)/

This will be less of a problem after patchlevel 27, but it still helps
some, unless almost all your strings are blank.

Larry