It works fine for characters in the native charset...

# feeder.pl
$s = join '', map chr, 0..255;
print $s;
[download]

# tester.pl
$s = join '', map chr, 0..255;
1 while $read = read(STDIN, $in, 512, $ofs+=$read);
print($in eq $s ? "pass\n" : "fail\n");
[download]

$ perl feeder.pl | perl tester.pl
pass
$ perl feeder.pl | perl -C6 -pe1 | perl -C1 -pe1 | perl tester.pl
pass
[download]

... But it doesn't work so well for characters not in the native charset.

$ perl -C6 -e 'print "\x{2660}"' | perl -C1 -pe1
Wide character in print, <> line 1.
[junk][junk][junk]
[download]

Compare to

$ perl -C6 -e 'print "\x{2660}"' | perl -C1 -MEncode -pe '$_=encode("i
+so-latin-1",$_);'
?
[download]

Up to the reader to choose if that's good enough.