From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <2142d47b30d21e30e6d05a708a0907a8@quintile.net>
From: "Steve Simon" <steve@quintile.net>
Date: Fri, 15 Jul 2011 14:05:42 +0100
To: 9fans@9fans.net
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: [9fans] awk, big REs
Topicbox-Message-UUID: 016d9fe4-ead7-11e9-9d60-3106f5b1d025

I am trying to run configure for a GNU project on plan9.

yes I know, but stay with me...

it mostly works but part of the configure script fails where it runs
awk over a header file and uses a rather sill RE to find all defines/undefines.

/^[\t ]*#[\t ]*(define|undef)[\t ]+[_abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ][_abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789]*([\t (]|$)// { print; }

For some reasone they don't like character class ranges, and if I
replaces the lists with ranges the RE works as I expect,

/^[\t ]*#[\t ]*(define|undef)[\t ]+[_a-zA-Z][_a-zA-Z0-9]*([\t (]|$)/ { print; }

I looked at the awk source and even upped MAXRE - though its
existing value of 512 looks fine to me.

anyone any ideas why awk might not match with an RE like this?

-Steve