Bug #89696 for Type-Tiny: Types::Standard: is it possible to check for an empty ArrayRef/HashRef?

Tue Oct 22 08:22:36 2013 bpj [...] melroch.se - Ticket created

Subject:	Types::Standard: is it possible to check for an empty ArrayRef/HashRef?
Date:	Tue, 22 Oct 2013 14:22:19 +0200
To:	bug-Type-Tiny [...] rt.cpan.org
From:	BPJ <bpj [...] melroch.se>

Is it possible to check for an empty ArrayRef or HashRef with Types::Standard? The problem is I want to make a constraint which accepts either an empty ref or a ref with some specific (type of) contents. It seems now I can only accept either a ref with some specific content (Tuple/Map), a ref with some specific type of content (ArrayRef/HashRef with params) or a ref with any content (ArrayRef/HashRef without params). To be sure I can do something like my $empty_hash = declare 'EmptyHashRef', as HashRef, where { 0 == keys %$_ }; my $foo = union [ $empty_hash, Map[ Str, Str ] ]; but I see myself doing this all the time, so it would be nice if EmptyArrayRef and EmptyHashRef or something with the same effect were bundled in. /bpj

Tue Oct 22 20:17:56 2013 perl [...] toby.ink - Correspondence added

union[$empty_hash,Map[Str,Str]] would be equivalent to Map[Str,Str] anyway, because the empty hash passes Map[Str,Str]. I think perhaps Dict[] (with empty brackets) probably should be an explicit check for an empty hashref; and Tuple[] should correspondingly check for an empty arrayref. But they do not currently work that way - I think that's arguably a bug.

Tue Oct 22 20:17:56 2013 The RT System itself - Status changed from 'new' to 'open'

Wed Oct 23 03:46:51 2013 bpj [...] melroch.se - Correspondence added

Subject:	Re: [rt.cpan.org #89696] Types::Standard: is it possible to check for an empty ArrayRef/HashRef?
Date:	Wed, 23 Oct 2013 09:46:36 +0200
To:	bug-Type-Tiny [...] rt.cpan.org
From:	BPJ <bpj [...] melroch.se>

2013-10-23 02:17, Toby Inkster via RT skrev: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=89696 > > > union[$empty_hash,Map[Str,Str]] would be equivalent to > Map[Str,Str] anyway, because the empty hash passes > Map[Str,Str].

It does? So Map[Str,Str] means "*If* the hashref has any keys/values they should be strings" rather than "the hashref should not be empty and its keys/values should be strings"; in that case I think there ought to be some way to check for the *non*-emptiness of a hash/array ref. Perhaps an optional additional parameter (Boolean)? Show quoted text

> I think perhaps Dict[] (with empty brackets) probably should be > an explicit check for an empty hashref; and Tuple[] should > correspondingly check for an empty arrayref. But they do not > currently work that way - I think that's arguably a bug.

Actually that was how I expected them -- and Map[] -- to work! /bpj

Mon Oct 28 07:33:28 2013 perl [...] toby.ink - Correspondence added

I've written a fix for Dict[] and Tuple[] which are not explicit type constraints for empty hashrefs and arrayrefs respectively. This is in the repo, and will be in the next stable release. If you want an explicit non-empty hash or array you can combine various type constraints using overloaded operators: ( ArrayRef[Int] ) & ~ ( Tuple[] ) would be an ArrayRef of Ints and cannot be empty. Or, if the overloading is a little too magic-looking, you could use the object-oriented style to make it more explicit: Type::Tiny::Intersection->new( type_constraints => [ ArrayRef->parameterize(Int), Tuple->parameterize()->complementary_type, ], ) I'm marking this ticket as "patched" for now; once this change gets into a stable release I'll mark it as "resolved". I probably still ought to write a bit more documentation in this area.

Mon Oct 28 07:33:28 2013 perl [...] toby.ink - Status changed from 'open' to 'patched'

Mon Oct 28 17:53:14 2013 perl [...] toby.ink - Taken

Tue Nov 05 06:38:06 2013 perl [...] toby.ink - Correspondence added

Tuple[] and Dict[] work in 0.032.

Tue Nov 05 06:38:08 2013 perl [...] toby.ink - Status changed from 'patched' to 'resolved'