Re: Bug-Report

[Aharon Robbins]

Greetings.

This is undoubtedly a locale issue. If you use

        export LC_ALL=C

at the shell level, things should behave as you expect.

Alternatively, you can use the POSIX bracket expressions to
express the character types

        /[[:alpha:]]/   # alphabetic characters
        /[[:lower:]]/   # lowercase
        /[[:upper:]]/   # uppercase
        /[[:digit:]]/   # digits

and things should also work as you expect.

See the gawk manual.

Thanks,

Arnold

> Date: Sun, 24 Sep 2006 03:04:01 +0800
> From: DsangWonFan <address@hidden>
> Subject: Bug-Report
> To: address@hidden
>
> Greeting:
>
> There is something wrong with gawk 3.1.4. While using some regular
> expressions such as /[1-9]/,
> /[a-z]/, or/[A-Z]/, as a pattern ,the gawk 3.1.4 often makes mistakes.
> For examples, let's suppose there is a file named "temp", and what
> "temp" contains is:
>
> 1234
> 4567
> acpi
>
> i.e. when you execute the command "cat temp" the result of which is:
>
> 1234
> 4567
> acpi
>
> But unfortunately, if we execute a command like this "gawk '{if($0 ~
> /[1-9]/){print $0;}}' temp" ,
> the result is:
>
> 1234
> 4567
> acpi
>
> In fact, the "acpi" shall not be printed!
>
> The version of the compiler which I employed to compile gawk 3.1.4 is
> gcc (GCC) 3.3.5 (Debian 1:3.3.5-13).
>
> Sincerely Yours,
> DsangWonFan