Re: [bug-gawk] Using ENVIRON["LANG"] = "C" instead of --characters-as-bytes

[Hermann Peifer]

On 2015-03-12 20:11, Eli Zaretskii wrote:
> > Date: Thu, 12 Mar 2015 19:14:47 +0100
> > From: Hermann Peifer <address@hidden>
> >
> > I thought that in gawk/master, I could use ENVIRON["LANG"] = "C" instead
> > of --characters-as-bytes, in analogy to changing the time zone via
> > ENVIRON["TZ"] = "UTC". I do however always end up with the lint warning:
> > Invalid multibyte data detected. There may be a mismatch between your
> > data and your locale.
> >
> > Is this a feature ?
> >
> > Hermann
> >
> > # Some code snippet which doesn't work as epxected
> > BEGIN {
> >         # Try to simulate --characters-as-bytes
> >         ENVIRON["LC_ALL"] = "C"
> >         ENVIRON["LANG"] = "C"
> > }
>
> AFAIR, you need to call 'setlocale' after setting these in the
> environment, for it to switch to another locale.  Unless, that is,
> Gawk does that for you when you set these members in ENVIRON[] (which
> it doesn't, AFAICS).
>
> And btw, the assumption that 'setlocale' looks at these environment
> variables is non-portable outside of the Posix world.

Thanks. I am using gawk/master where the NEWS say:

1. If not in POSIX mode, changes to ENVIRON are reflected into
   gawk's environment, affecting any programs run by system()
   or for piped redirections. This can also affect built-in routines, such
   as mktime(), which is typically influenced by the TZ environment 
variable.

# This works as expected (my locale is en_US.UTF-8)
$ LC_ALL=C awk -f myscript.awk ...

Hermann