Hacker News Reader

Comment by timakro | Hacker News Reader

I believe no distro actually ships this version of awk by default. They ship GNU awk which has Unicode support anyways.

OpenBSD uses "The One True AWK."

  $ awk -V
  awk version 20211208

Kernighan's version is likely used in other places where the GPL is eschewed.

I think the other BSDs do too, including macOS.

Yes, here on macOS 11.6.8

  $ awk -version
  awk version 20200816

On a FreeBSD 12.3-RELEASE box, FWIW:

   # awk -version
   awk version 20210215

Debian:

https://distrowatch.com/table.php?distribution=debian&pkglis...

So it turns out the default on Debian is mawk which does NOT support Unicode. Thanks for pointing that out. This simple test gives different results for gawk and mawk.

  $ echo 'ö' | awk '{print length}'

…only if the current locale is set to use UTF-8 (or some other variable-width encoding). Which nowadays the default locale usually does, but in principle it doesn’t need to be.

Reply on news.ycombinator.com