You're still assuming they're the "same thing"in some sense. The type of the () postfix operator says it applies to a function or method (more specifically, to something that supports the Callable role), thus to "words" itself. The type of the [] postfix operator says it applies to something that supports the Positional role; as "words" produces a Positional, the [] applies to the *result* of words.
"words" itself doesn't see or know about either of them. It is invoked by the () postfix, and its result is processed by the [] postfix. Do you really think every function that produces a "list" has built into it secret knowledge of how you would index a list? They just produce a "Positional" or "Seq" or "List", etc. and *those* know what to do with []. The function doesn't even need to say it's doing that; that's to produce better error messages, not "wire what a list is into me". On Wed, Sep 26, 2018 at 5:32 PM ToddAndMargo <toddandma...@zoho.com> wrote: > On 9/26/18 11:42 AM, Larry Wall wrote: > > On Wed, Sep 26, 2018 at 01:16:26PM -0400, Parrot Raiser wrote: > > : Would it be correct to say: > > : [ ] aka square brackets, always surround the subscript of an array or > > : list, i.e. here "n: is an integer, [n] always means the nth item, > > : while > > : ( ), round brackets or parentheses, separate and group items to > > : control processing order, and identify subroutine calls, surrounding > > : the argument list, if any? > > > > I'd say there's a bit of confusion here between different syntactic > > categories. Your first definition is assuming only postfix position > > (making [] a "postcircumfix"), while your second definition includes the > > use of () in both term position and postfix position. In term position, > > yes, parentheses control the processing order and group things, but > > [] also has a term-position interpretation of its own, which is as an > > array composer (but not a subscript). In postfix position, [] is always > > a subscript as you indicated, but () is only for passing arguments > > to something (and forcing a function call if that something might be > > construed as a different keyword without the parens). Any grouping > > behavior is incidental to the argument processing, and in fact many > > constructs you can use in normal grouping parens are illegal in an > > argument list: > > > > > p6 'say abs(42 or 43)' > > ===SORRY!=== Error while compiling -e > > Unable to parse expression in argument list; couldn't find final > ')' (corresponding starter was at line 1) > > at -e:1 > > ------> say abs(42⏏ or 43) > > > > but in term position that's fine since the inside is a whole statement: > > > > > p6 'say abs (42 or 43)' > > 42 > > > > The latter cannot be taken as a postfix because of the space. (Unless > > you use Tuxic, which confuses this intentionally. :) > > > > The three main syntax categories you have to be able to track (and > > it's much easier to track in Perl 6 than in Perl 5), are: > > > > 1) when you're expecting a term > > 2) when you're expecting a postfix (or infix if there's no postfix) > > 3) when you're expecting only an infix > > > > The careful distinction of which category the grammar engine is expecting > > is critical to understanding any version of Perl, whether 5 or 6. > > You can't adequately describe any random syntax in Perl without first > > nailing down which of the main grammar rules is going to be parsing it. > > > > Well, I've probably said more than enough. :) > > > > Larry > > > > Hi Larry, > > Wonderful explanation! Thank you! > > Followup question. I am confused. How did they get from > > multi method words(Str:D $input: $limit = Inf --> Positional) > > to what goes in the () and what goes in the []? > > And why are they using the term "$limit" instead of "$selection"? > > Yours in confusion, > -T > > > > -- > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > Computers are like air conditioners. > They malfunction when you open windows > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > -- brandon s allbery kf8nh allber...@gmail.com
