use Encode qw(from_to); use strict;
my $s = join "", map { chr } 128..255;
for my $enc (qw( iso8859_3 ))
{
from_to($s,$enc,"utf-8");
}
The problem is that iso-8859-3 does not assign characters to all octet
values.
What should Encode:: do in such cases:
A. U+FFFD
B. Map octet to Unicode/iso-8859-1
C. Use a "private use" page...
If there is no function hanging on the 'illegal characters' hook, then it
should return U+FFFD. This is pretty clear from the standard, etc. The
other options don't bare thinking about since they change the mapping
definition of iso8859-3.
The fun starts on the return path :) Ideally an encoding should list a
default fallback Unicode fail character for mapping unmapped Unicode
values. I.e. the 8-bit equivalent of U+FFFD.
Martin Hosken
