Re^4: perl 5.12 BSD portability (CPAN test result)...print

[perl-diddler]

in reply to Re^3: perl 5.12 BSD portability (CPAN test result)...print
in thread perl 5.12 BSD portability (CPAN test result)...print

But only in the FH->print format? Why does it work in the print FH mode? Aren't they supposed equivalent?

[chromatic]

The parser knows about print as a keyword. It's been a special case as far back as I can remember (Perl 1?). You can look at it as an indirect method call from the language level, but the parser doesn't need heuristics because the grammar has a specific rule for the print $filehandle case.

[ikegami]

But only in the FH->print format? Why does it work in the print FH mode? Aren't they supposed equivalent?

No, they're not equivalent. print FH ... is an instance of a the print operator.

FH->print(...) is a method call. It does whatever FH's class's print method does. If FH is an IO::Handle or IO::File object, its print method calls the print operator.

sub print {
    @_ or croak 'usage: $io->print(ARGS)';
    my $this = shift;
    print $this @_;
}

[perl-diddler]

Ah...I thought it was just an indirect method call...didn't know it was it's own special operator... Does it still have to be it's own special operator? I.e. are there places where the indirect method call wouldn't give the same behavior? (Apparently there were in the past, but what about now?)....

[ikegami]

Assuming you're ok with breaking every program that does print $x, how would you implement IO::Handle::print without the print operator?

There are other issues, but they're not worth mentioning after the above two.