y/·/./ gives two points instead of one

[rsFalse]

I was translating morse code, and each centered-point was translated to two (not one) ascii dots. Can anybody explain?

use warnings;
use strict;

$_ = '·|·';
print y/·/./r, "\n";
print s/·/./gr, "\n";

OUTPUT:
..|..
.|.

[ikegami]

You think you have

$_ = '·|·';
print y/·/./r, "\n";
print s/·/./gr, "\n";

but you actually have

$_ = '·|·';
print y/·/./r, "\n";
print s/·/./gr, "\n";

This is because you treated the file as UTF-8 when you viewed it, but you didn't tell Perl to do the same by adding use utf8;.

[rsFalse]

Thanks for explanation!

[LanX]

> Can anybody explain?

use utf8 ! :)

update

Without utf8 flag you are forcing perl to transliterate a 2 byte character to a 1 byte character.

My interpretation is that this 1 byte char is doubled to keep the counters in sync.¹

update

same effect here

use warnings;
use strict;
#use utf8;

$_ = '§|§';

print y/§/$/r, "\n";

Careful with such experiments, I was able to make my Perl (5.14) coredump

update

¹) ikegami's explanation is better, you are transliterating the first and second byte for · separately to .

  DB<123> unpack 'H*','·'
 => "c2b7"

Got confused about the tr flags and their defaults.