Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!cmcl2!rutgers!rochester!PT!k.gp.cs.cmu.edu!bww
From: bww@k.gp.cs.cmu.edu (Bradley White)
Newsgroups: comp.lang.c
Subject: Re: whats wrong with this lex grammer(sic)? (help)
Message-ID: <1238@k.gp.cs.cmu.edu>
Date: Fri, 25-Sep-87 19:19:06 EDT
Article-I.D.: k.1238
Posted: Fri Sep 25 19:19:06 1987
Date-Received: Sun, 27-Sep-87 04:56:59 EDT
References: <443@aucs.UUCP> <8765@mimsy.UUCP>
Distribution: na
Organization: Carnegie-Mellon University, CS/RI
Lines: 17
Summary: correct lex expression for (*...*) comments

In article <8765@mimsy.UUCP>, chris@mimsy.UUCP (Chris Torek) writes:
> The expression
> 
> 	\(\*(\*[^)]|[^*])*\*+\)
> 
> should work.

Clearly the above expression is wrong if Lex selects the longest
matching pattern (as it does).  Consider, for example, the string
"(***)*)".  If Lex took the shortest pattern then \(\*.*\*\) would
suffice.  A correct expression is

	\(\*[^*]*\*+([^*)][^*]*\*+)*\)

-- 
Bradley White <bww@cs.cmu.edu>         +1-412-268-3060
CMU Computer Science Department  40 26'33"N 79 56'48"W