Re: [bug-gawk] Bug with GNU Awk 3.1.4

[Andrew J. Schorr]

Hi,

On Mon, Mar 14, 2016 at 09:32:48PM -0400, Christo M wrote:
> I would like to report a bug with GNU Awk version 3.1.4.

I have 3 comments about your problem:

1. This is a really old version of gawk. Please upgrade to the
current version 4.1.3. You may find that it fixes your problem.
You can download it here:
   http://ftp.gnu.org/gnu/gawk/

2. The behavior of bracket expressions like [A-Z] depends on the locale. 
Do you see the same problem in the C locale?

3. GNU Awk has character classes that can be used to avoid locale-sensitive
differences in behavior. They are documented in the distributed info and man
pages, and also here:
   https://www.gnu.org/software/gawk/manual/html_node/Bracket-Expressions.html

> awk -F'[ :]' --re-interval '$1 ~ /^[A-Z][a-z]{3}$/{print $1}' lab3.data

This is better written as:

   gawk -F'[ :]' --re-interval '$1 ~ /^[[:upper:]][[:lower:]]{3}$/{print $1}' 
lab3.data

> awk -F'[ :]' --re-interval '$1 ~ /^[A-Z]{4}$/{print $1}' lab3.data

This should be:
   gawk -F'[ :]' --re-interval '$1 ~ /^[[:upper:]]{4}$/{print $1}' lab3.data

Also, please note that --re-interval is the default behavior in current gawk,
so need not be specified.

Regards,
Andy