Hm, it seems that http://perldoc.perl.org/perlsyn.html doesn't include the text you quoted, looks like it's behind. But it's up on http://perldoc.pl/perlsyn.

Yeah I took it from .Pl because my old mobile isn't up to date on certificates and is blocked from Perl.org.

Honestly I was surprised about the new detailed explanation

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery FootballPerl is like chess, only without the dice}

Even if it wasn't magical, it still wouldn't stop at that line. The two-character string being returned, 0␊, is true. The magic only matters when the last line of the file consists of a 0 without a terminating Line Feed. (It could also happen when $/ = \1;.)

(Oops, the main point of this has already been mentioned a couple of times below, though this post still brings in new information.)

(solved, see update)

> (It could also happen when $/ = \1; )

This splits on single characters, including \n.

But I couldn't find this behavior documented, neither in perlvar#$INPUT_RECORD_SEPARATOR nor readline

use strict;
use warnings;
use feature 'say';

$/ = \1;

while ( $_= <DATA>) {
    say "<$_>";
}

__DATA__
1
0
345
[download]

--->

<1>
<
>
<0>
<
>
<3>
<4>
<5>
<
>
[download]

update

oops I misread \1 as "\1" (ASCII 01), but it's a ref to literal 1, so this from perlvar applies:

> Setting $/ to a reference to an integer, scalar containing an integer, or scalar that's convertible to an integer will attempt to read records instead of lines, with the maximum record size being the referenced integer number of characters.

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery FootballPerl is like chess, only without the dice}