1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
Post-1.000000 Changes
=====================
* Type::Params should provide compile_named/validate_named functions
tailored for named parameters rather than using the slurpy Dict hack.
* Try to make inlining for ClassName, RoleName, StrMatch, etc more
suitable for things like Mite.
* Inform Type::Coercion when it *should have* been frozen (for example,
the parameterized ArrayRef, HashRef, Dict, and Tuple coercion
generators).
After that, $coercion->add_type_coercions can issue a warning about
that bad things are happening. On Perl 5.10, this can be done with
warnings::warnif. On earlier Perls that mechanism is unreliable, so
just use plain old warn.
* Investigate popular type constraints defined outside Type-Tiny, and
consider which can be added to Types::Standard/Types::Common::*.
These would need to be constraints that can be implemented with no
non-core dependencies.
* Possibly add a NonEmpty type constraint, allowing NonEmpty[Str],
NonEmpty[HashRef[Int]], NonEmpty[ArrayRef], etc. This would be an
interesting one for optimizing.
* CycleTuple[Int,Str,HashRef] should allow [1,foo=>{},2,bar=>{}].
* Perhaps allow Type::Coercion::FromMoose's implementation of
add_type_coercions to succeed?
* to_TypeTiny's _TypeTinyFromGeneric could support inlining?
|
