develooper Front page | perl.perl5.porters | Postings from August 2009

Re: A complete design for := (bind)

Thread Previous | Thread Next
August 25, 2009 13:22
Re: A complete design for := (bind)
Message ID:
On Tue, Aug 25, 2009 at 12:48:41PM -0700, Chip Salzenberg wrote:
> On Fri, Aug 21, 2009 at 07:37:05AM -0400, David Golden wrote:
> > Is @b or *@b likely to be the more common case?  If *@b, then I'd
> > rather see those as \@b and @b, respectively so the common case is
> > most like the flattening Perl 5 syntax we know today.
> I'm somewhat sympathetic to the POV that introducing Perl 6 flattening
> semantics and punctuation is a poor fit in Perl 5, which has opposite
> defaults [most of the time].
> There are two competing memes in this field: Perl 5 prototypes (that exist),
> and lvalue refs (proposed long ago).
> The existing Perl 5 prototype syntax spells concepts *similar* to flatten
> and unflatten as @ and \@.  Just changing the bind syntax in that fashion,
> one ends up with:
>     :($a, \@b, \%c) := ...   # should have 3 elements on RHS
>     :($a,  @b,  %c) := ...   # @b ends up with N elements, %c ends up empty
> Good: this isn't as ugly as I expected, and kinda looks like what it does.
> Bad: It means that the backslash here:
>     :(\$a) := ...
> is either illegal or useless -- an unfortunate situation, and a warning sign
> of error, but perhaps tolerable.
> Alternate universe:
> If we were to discard Perl 6 as a syntactic model and take only its useful
> semantics, we could start with a proposal I made long ago but didn't work
> all the kinks out of: lvalue refs.  All of the below, with one exception[1],
> are fairly straightforward semantic extensions to Perl 5, and their meanings
> (if not implementations :-)) should be clear:
>    \$a = \$b;
>    (\@a, \@b) = (\@c, \@d);
>    \my($self, %args) = \(@_);

The above are somewhat obvious. But it begs the question, what do the
following mean? Would it even be legal?

    \$a = $b;    # Runtime error if $b isn't a ref to a scalar?
    \$a = \@a;
    \@a = [@b];

I don't think we need another smartmatch-like table to figure out what
LHS ref vs RHS ref means.

I kind of like:

    my    ($a, @b, %c)
    local ($a, @b, %c)
    our   ($a, @b, %c)
    state ($a, @b, %c)   # Yeah, if only it could do aggregates.
    bind  ($a, @b, %c)

> But where does one put "ro", "rw", or "copy" in this?  And what becomes the
> syntactic marker to allow parsing new keywords in the first place?  And will
> the resulting function prototypes be readable enough to use?  One wonders.
>   [1] The unstraightforward part of the extension is that using \(@a) or
>       \(%a) as an lvalue would erase existing contents of @a or %a; or if
>       not that, would at the very least stretch them if they are too small.
>       Currently, nothing you can do with \(@a) alters the size of @a.
> > I also don't terribly like :() syntax because I find it hard to skim.
> I'm sorry, but "that's too much punctuation" is not a credible sentiment
> from a Perl 5 programmer.  :-,

I like a keyword because it acts a bit like my/local/our/state - but that's
because I'm assume it will have some lexical scoped effect. ':( )' has the
advantage it's currently not legal syntax (AFAIK), and hence wouldn't need
an 'use feature;' to enable it.

But I find the feature to useful and important to spend too much time on
bikeshedding the syntax.

> > >  :($a is ro) := ...
> > >  :($a is copy) := ...
> > 
> > I don't like overloading "is" any more than it is already.
> I'm sorry, but "I have to be aware of context" is not a credible sentiment
> from a Perl 5 programmer, either.  :-,

That I fully agree with.

> >     bind(@a) := @x;
> In fairness, that's pretty.


Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About