Re: sed error message reports byte position instead of char position when program contains UTF-8

> From: [email protected] (Paul Jarc)
> Date: Wed, 15 May 2013 14:36:47 -0400
> 
> Eli Zaretskii <[email protected]> wrote:
> > Options to the 's' command are all pure-ASCII single-byte characters,
> > so Sed is correct in this case, I think.
> 
> I don't think so.  The example doesn't violate that rule.  The unknown
> option is "/".  But due to a multibyte character in the pattern, the
> offset is either miscalculated (should say "11" instead of "12") or
> badly described (should say "byte" instead of "char").

"Byte" is the only viable alternative, but that leaves the burden of
counting bytes on the user.