Consider the following:

use strict;
use warnings;

local $_ = q{<>^};

for my $i (qw(<> >^)) {
    /\Q$i/p and print "${^MATCH}\n";
}
[download]

Output:

<>
>^
[download]

Remember that "^" can also be used to negate a class match in addition to denoting a match at the beginning of a line, e.g., /^[^aeiou]/i (no vowels at the beginning of a line). Both "<" and ">" are metacharacters. See this Extended regex sequences section.

Hope this helps!

Thanks

> for $i qw(<> >^){

should be deprecated by now, please always put parentheses around your list.

>perl -e"for my $i qw( a b ) { }"
syntax error at -e line 1, near "$i qw( a b )"
Execution of -e aborted due to compilation errors.
[download]

Thanks for discussion and explanations, but I haven't understand why it is a bug - not using parentheses, a from when? And then - why?

And it seems that <> characters are quoted, if I use quotemeta, but why they are meta?

quotemeta doesn't bother being smart about what it escapes, because it doesn't need to, and because a list of metacharacters might not be portable across different versions of Perl. quotemeta just escapes everything that might be special, even if that character isn't necessarily special at that point in the regex.

Always using quotemeta / \Q...\E is a good habit to get into when interpolating variables into regexes. I'd only not use them when I want the string in the variable to be interpreted as a regex, or when I know the variable I'm interpolating contains no special characters (something like /^[A-Za-z_0-9]*$/).