Hello,
On Tue, 19 Mar 2013 23:44:22 +0000
"brian m. carlson" <sandals@crustytoothpaste.net> wrote:
> > Well, as far as I know, mawk has some sort of terrible UTF-8
> > support, so it's a no way for many applications.
> Could you please explain? And if you haven't filed a bug report,
> could you please do so? Searching Google, the only UTF-8-related
> bugs I found are bugs mandated by POSIX (and one that updating mawk
> to 1.3.4 would fix).
$ echo привет | mawk '{ print length($0) }'
12
$ echo привет | gawk '{ print length($0) }'
6
$ echo привет | mawk '{ printf substr($0, 1, 1) }' | hd
00000000 d0 |.|
00000001
I don't think it's mandated by POSIX (or is it?). Anyway, it's obviously
not what most of the applications want.
> Also, my original post was inspired by the fact that most packages
> depending on gawk invoke "awk" as their binary. In that case, the
> dependency is wrong and unnecessary.
I just always put #!/usr/bin/gawk -f in the beginning of the file and
declare an explicit dependency on gawk. And also, I don't support
running those scriptf on broken awk implementations.
--
WBR, Andrew
Attachment:
signature.asc
Description: PGP signature