=pod =encoding utf-8 =head1 NAME Type::Tiny::Manual::UsingWithMoose - how to use Type::Tiny with Moose =head1 MANUAL First read L, L, and L. Everything in those parts of the manual should work exactly the same in Moose. This part of the manual will focus on Moose-specifics. =head2 Why Use Type::Tiny At All? Moose does have a built-in type constraint system which is fairly convenient to use, but there are several reasons you should consider using Type::Tiny instead. =over =item * Type::Tiny type constraints will usually be faster than Moose built-ins. Even without Type::Tiny::XS installed, Type::Tiny usually produces more efficient inline code than Moose. Coercions will usually be a lot faster. =item * Type::Tiny provides helpful methods like C and C that allow type constraints and coercions to be easily tweaked on a per-attribute basis. Something like this is much harder to do with plain Moose types: has name => ( is => "ro", isa => Str->plus_coercions( ArrayRef[Str], sub { join " ", @$_ }, ), coerce => 1, ); Moose tends to encourage defining coercions globally, so if you wanted one B attribute to be able to coerce from B<< ArrayRef[Str] >>, then I B attributes would coerce from B<< ArrayRef[Str] >>, and they'd all do that coercion in the same way. (Even if it might make sense to join by a space in some places, a comma in others, and a line break in others!) =item * Type::Tiny provides automatic deep coercions, so if type B has a coercion, the following should "just work": has xyzlist => ( is => 'ro', isa => ArrayRef[Xyz], coerce => 1 ); =item * Type::Tiny offers a wider selection of built-in types. =item * By using Type::Tiny, you can use the same type constraints and coercions for attributes and method parameters, in Moose and non-Moose code. =back =head2 Type::Utils If you've used L, you may be accustomed to using a DSL for declaring type constraints: use Moose::Util::TypeConstraints; subtype 'Natural', as 'Int', where { $_ > 0 }; There's a module called L that provides a very similar DSL for declaring types in Type::Library-based type libraries. package My::Types { use Type::Library -base; use Type::Utils; use Types::Standard qw( Int ); declare 'Natural', as Int, where { $_ > 0 }; } Personally I prefer the more object-oriented way to declare types though. Since Type::Library 1.012, a shortcut has been available for importing Type::Library and Type::Utils at the same time: package MyType { use Type::Library -base, -utils; ...; } In Moose you might also declare types like this within classes and roles too. Unlike Moose, Type::Tiny doesn't keep types in a single global flat namespace, so this doesn't work quite the same with Type::Utils. It still creates the type, but it doesn't store it in any type library; the type is returned. package My::Class { use Moose; use Type::Utils; use Types::Standard qw( Int ); my $Natural = # store type in a variable declare 'Natural', as Int, where { $_ > 0 }; has number => ( is => 'ro', isa => $Natural ); } But really, isn't the object-oriented way cleaner? package My::Class { use Moose; use Types::Standard qw( Int ); has number => ( is => 'ro', isa => Int->where('$_ > 0'), ); } =head2 Type::Tiny and MooseX::Types L should be a drop-in replacement for L. And L and L should easily replace L and L. That said, if you do with to use a mixture of Type::Tiny and MooseX::Types, they should fit together pretty seamlessly. use Types::Standard qw( ArrayRef ); use MooseX::Types::Common::Numeric qw( PositiveInt ); # this should just work my $list_of_nums = ArrayRef[PositiveInt]; # and this my $list_or_num = ArrayRef | PositiveInt; =head2 C<< -moose >> Import Parameter If you have read this far in the manual, you will know that this is the usual way to import type constraints: use Types::Standard qw( Int ); And the C which is imported is a function that takes no arguments and returns the B type constraint, which is a blessed object in the L class. Type::Tiny mocks the L API so well that most Moose and MooseX code will not be able to tell the difference. But what if you need a real Moose::Meta::TypeConstraint object? use Types::Standard -moose, qw( Int ); Now the C function imported will return a genuine native Moose type constraint. This flag is mostly a throwback from when Type::Tiny native objects I<< didn't >> directly work in Moose. In 99.9% of cases, there is no reason to use it and plenty of reasons not to. (Moose native type constraints don't offer helpful methods like C and C.) =head2 C<< moose_type >> Method Another quick way to get a native Moose type constraint object from a Type::Tiny object is to call the C method: use Types::Standard qw( Int ); my $tiny_type = Int; my $moose_type = $tiny_type->moose_type; Internally, this is what the C<< -moose >> flag makes imported functions do. =head1 NEXT STEPS Here's your next step: =over =item * L How to use Type::Tiny with Mouse, including the advantages of Type::Tiny over built-in type constraints, and Mouse-specific features. =back =head1 AUTHOR Toby Inkster Etobyink@cpan.orgE. =head1 COPYRIGHT AND LICENCE This software is copyright (c) 2013-2014, 2017-2023 by Toby Inkster. This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself. =head1 DISCLAIMER OF WARRANTIES THIS PACKAGE IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. =cut