Re: follow up to: gawk v 3.1.1 match function bug

[Peter Leggett]

Hi (again),

Apologies. I should have searched the web more thoroughly.

After a search of the newsgroups I now understand a little
more about locales, use of :upper: etc.. and realise it was my locale
en_GB.UTF-8 which was causing the problem. If I set to 'C' problem
goes away.

Sorry to report a red herring.

Pete

Peter Leggett wrote:
> Hi,
>
> I think there is a bug with the match function in gawk 3.1.1 (as 
> released with
> RedHat 9) and case matching.
>
> Here is a simple example of the bug.
>
> address@hidden src]$ uname -a
> Linux wiggen 2.4.20-18.9 #1 Thu May 29 06:54:41 EDT 2003 i686 athlon 
> i386 GNU/Linux
> address@hidden src]$ gawk --version
> GNU Awk 3.1.1
> address@hidden src]$ more testdata
> ABC
> abc
> address@hidden src]$ gawk '{ print match($0,/[A-Z]/) }' < testdata
> 1
> 2
> address@hidden src]$
>
>
> And to show it works on an earlier version
>
> address@hidden 110 % uname -a
> SunOS herman 5.8 Generic_108528-19 sun4u sparc SUNW,Sun-Blade-1000
> address@hidden 111 % gawk --version
> Gnu Awk (gawk) 3.0, patchlevel 0
> address@hidden 112 % more testdata
> ABC
> abc
> address@hidden 113 % gawk '{ print match($0,/[A-Z]/) }' < testdata
> 1
> 0
> address@hidden 114 %
>
> Please accept my apologies in advance if this bug has already been reported
> or is not a bug but a product of a mis-understanding on my part.
>
> Kind regards
>
> Peter