Re: commit-msg hook

[Eli Zaretskii <eliz@gnu.org>]

> Date: Sat, 11 Apr 2015 12:55:50 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> Digging a bit deeper, I see that the problem with hanging is not due
> to Gawk, but instead due to 'printf': it doesn't produce the UTF-8
> encoding of the percent sign, but some invalid sequence instead.
> 
> If I add an explicit
> 
>     cent_sign = "¢"
> 
> to the Awk part of the script (where the cent sign is UTF-8 encoded),
> then the hook correctly processes log messages with UTF-8 encoded
> non-ASCII letters and doesn't hang.  It doesn't use the [[:print:]]
> alternative, though, probably because UTF-8 encoded characters are not
> recognized as such, and a UTF-8 locale cannot be used on MS-Windows to
> force that.

Would it be too expensive to use Emacs for detecting unprintable
characters, instead of the shell and Awk?  All those issues with
handling UTF-8 encoded characters correctly and portability issues
will then simply disappear.