"my TYPE VARLIST" isn't good enough

[Ovid]

From perldoc -f my:

    my VARLIST
    my TYPE VARLIST
    my VARLIST : ATTRS
    my TYPE VARLIST : ATTRS

Currently, we can write code like this:

my ( $count, $aref ) = @_;

The TYPE cannot handle that. It assumes one type for all.

If a VAR is defined as VAR or TYPE VAR, we can make this work:

my ( UINT $count, ARRAY $aref ) = @_;

(Or using the attribute syntax, perhaps).

Paging @Perl-Oshun/core-team and @leonerd

Thoughts?

[druud]

I presume this compiles:

(my UINT $count, my ARRAY $aref) = @_;

but I always wondered how the first my doesn't see a VARLIST of 2 elements.

This certainly doesn't compile:

my main ($i, my CORE @r) = @_; 

(ignore the main/CORE usage, they are only there as available classnames)

[Ovid]

Don't know why it doesn't see a VARLIST, but I think people would strongly object to that being the syntax to use for checks.

[perigrin]

(Using the word type here cause that’s how perldocs refer to them)

I’m comfortable with changing the type annotations the way you’ve suggested … this is effectively how it’s going to have to change to allow type annotations in signatures.

If sub ($arg, @rest) {…} is supposed to be equivalent to sub { my ($arg, @rest) = @_; … } then we are gonna need to alter the way type annotations are parsed if we want type annotations in signatures.

I think sub { my ( UINT $count, ARRAY $aref ) = @_; } is a natural expansion of sub ( UINT $count, ARRAY $aref ) {} … so logic dictates if the latter is allowed so is the former.

[zmughal]

Also would align with destructuring assignment.

[perigrin]

Exactly.