Re: match finds wrong space.

[Davide Brini]

On Tue, 6 Jul 2010 22:54:33 +0100 "Chris Willis" <address@hidden>
wrote:

> Hi
> 
> another one, and pretty basic.
> 
> The attached snippet sd respond "9 Mary Ann jane", rather than "5  Mary
> Ann jane" it actually responds with. In actual practice it was
> particularly difficult as the failure caused an infinite loop.

Depending on your locale, it's expected behavior, since [a-z] can match
"A" in some locales, strange as it may sound.

$ LC_ALL=en_US.utf8 awk 'BEGIN {
s = "Mary Ann jane"
n = match( s, /\040[a-z]/ )
print n, s }'
5 Mary Ann jane

$ LC_ALL=C awk 'BEGIN {
s = "Mary Ann jane"
n = match( s, /\040[a-z]/ )
print n, s }'
5 Mary Ann jane

If you don't want to set the locale, use [[:lower:]]:

$ awk 'BEGIN {
s = "Mary Ann jane"
n = match( s, /\040[[:lower:]]/ )
print n, s }'
9 Mary Ann jane

And btw, the versions of awk you're using are old. Latest is 3.1.8.

-- 
D.