Re: Chained Modifiers

On Mon, Sep 15, 2008 at 12:16:22PM +0200, Moritz Lenz wrote:
: Chris Davaz wrote:
: > I'm not sure if it's anywhere in the specs, but I was expecting to be able
: > to do this:
: > 
: > .say for =$in unless $foo;
: 
: to quote http://perlcabal.org/syn/S04.html#Loop_statements
: 
: :Looping statement modifiers are the same as in Perl 5 except that, for
: :ease of writing list comprehensions, a looping statement modifier is
: :allowed to contain a single conditional statement modifier:
: :
: :    @evens = ($_ * 2 if .odd for 0..100);
: 
: 
: > Where we have multiple modifiers, here "for" and "unless". Is this in the
: > specs? If yes, is it supposed to work in pugs or rakudo? If it's not in the
: > specs can we add it? ;-)
: 
: The "authoritative" source to ask if a piece of text is valid Perl 6 is
: STD.pm. It parses your example, but not as you might expect. Output from
: STD5_dump_match
: 
: :Unknown routines:
: :        unless called at 1
: :comp_unit:
: : statementlist:
: :  statement:
: :   EXPR:
: :    noun:
: :     dotty:
: :      .
: :      dottyop:
: :       methodop:
: :        longname:
: :         name:
: :          identifier:
: :           say
: :
: :   statement_mod_loop:
: :    for
: :    modifier_expr:
: :     EXPR:
: :      pre:
: :       prefix:
: :        =
: :      noun:
: :       variable:
: :        sigil:
: :         sigil:
: :          $
: :        desigilname:
: :         longname:
: :          name:
: :           identifier:
: :            in
: :
: :  statement:
: :   EXPR:
: :    noun:
: :     term:
: :      unless
: :      args:
: :
: :       arglist:
: :        EXPR:
: :         noun:
: :          variable:
: :           sigil:
: :            sigil:
: :             $
: :           desigilname:
: :            longname:
: :             name:
: :              identifier:
: :               foo
: :  ;\n
: 
: So 'unless' is actually parsed as a sub call.

That's obviously a bogus parse, and I even know how it happened
(because "unless" is currently considered a terminator that doesn't
distinguish between expression termination and statement termination),
but I'm not sure how to fix it yet.  It's possible that statement
modifiers should be demoted to very-low-precedence operators, but there
are some problems with that that I don't have enough brain to think
through at the moment.

Larry

• Chained Modifiers by Chris Davaz
• Re: Chained Modifiers by Moritz Lenz
• Re: Chained Modifiers by Larry Wall