Bug #128337 for Type-Tiny: Support something similar to Params::ValidationCompiler#named_to

Wed Jan 23 14:43:53 2019 violapiratejunky [...] gmail.com - Ticket created

Subject:

Support something similar to Params::ValidationCompiler#named_to_list

It would be nice if Type::Params supported something like named_to_list in Params::ValidationCompiler: https://metacpan.org/pod/Params::ValidationCompiler#named_to_list

Fri Jan 25 09:00:40 2019 perl [...] toby.ink - Correspondence added

On 2019-01-23T19:43:53Z, srchulo wrote: Show quoted text

> https://metacpan.org/pod/Params::ValidationCompiler#named_to_list

That looks pretty possible. Would only be a few lines of code, I think. Might be able to include it before the next stable release. What should be done with missing optional parameters (with no defaults)? Should it just return undef for those?

Fri Jan 25 09:00:40 2019 The RT System itself - Status changed from 'new' to 'open'

Sun Jan 27 12:38:09 2019 violapiratejunky [...] gmail.com - Correspondence added

That's what Params::ValidationCompiler does, and I think that makes the most sense. I believe this would also line up with the behavior in the object returned from compile_named_oo. On Fri Jan 25 09:00:40 2019, TOBYINK wrote: Show quoted text

> On 2019-01-23T19:43:53Z, srchulo wrote:

> > https://metacpan.org/pod/Params::ValidationCompiler#named_to_list

> > That looks pretty possible. Would only be a few lines of code, I > think. Might be able to include it before the next stable release. > > What should be done with missing optional parameters (with no > defaults)? Should it just return undef for those?

Sun Jan 27 19:32:30 2019 perl [...] toby.ink - Correspondence added

On 2019-01-27T17:38:09Z, srchulo wrote: Show quoted text

> That's what Params::ValidationCompiler does, and I think that makes > the most sense. I believe this would also line up with the behavior in > the object returned from compile_named_oo.

compile_named_oo will return undef if you call a method corresponding to a missing named parameter, but it does also have `has_foo` methods. With named_to_list, there's effectively no way to distinguish between a value of undef and a missing value. But in many cases, you don't care about that distinction anyway, so I don't think that's a problem provided it's documented.

Sun Jan 27 19:47:12 2019 violapiratejunky [...] gmail.com - Correspondence added

On Sun Jan 27 19:32:30 2019, TOBYINK wrote: Show quoted text

> On 2019-01-27T17:38:09Z, srchulo wrote:

> > That's what Params::ValidationCompiler does, and I think that makes > > the most sense. I believe this would also line up with the behavior > > in > > the object returned from compile_named_oo.

> > compile_named_oo will return undef if you call a method corresponding > to a missing named parameter, but it does also have `has_foo` methods. > > With named_to_list, there's effectively no way to distinguish between > a value of undef and a missing value. But in many cases, you don't > care about that distinction anyway, so I don't think that's a problem > provided it's documented.

Yeah, I agree-- as long as it's documented, I think it's fine. And if someone cares, they can use the oo method. Also, I believe this matches up with the normal compile() with positional parameters? # Positional parameters state $check = compile(Int, Optional[Int], Optional[Int]); my ($foo, $bar, $baz) = $check->(@_); # $bar and $baz are optional

Sun Jan 27 20:01:38 2019 perl [...] toby.ink - Correspondence added

I think with compile(), you should be able to check the length of the returned list. Anyway... https://github.com/tobyink/p5-type-tiny/commit/608b259604cf18ba794239959fdc8f6a2b4e1646 So you should be able to do {named_to_list=>1} to get what you want. You can also do {named_to_list=>\@keys} to indicate what order you want the list to be returned in. (No sanity checks are performed on the key names!)

Sun Jan 27 22:59:12 2019 violapiratejunky [...] gmail.com - Correspondence added

On Sun Jan 27 20:01:38 2019, TOBYINK wrote: Show quoted text

> I think with compile(), you should be able to check the length of the > returned list. > > Anyway... > > https://github.com/tobyink/p5-type- > tiny/commit/608b259604cf18ba794239959fdc8f6a2b4e1646 > > So you should be able to do {named_to_list=>1} to get what you want. > > You can also do {named_to_list=>\@keys} to indicate what order you > want the list to be returned in. (No sanity checks are performed on > the key names!)

That looks great! Thank you :) That looks great! Thank you :)

Mon Jan 28 04:24:07 2019 perl [...] toby.ink - Taken

Mon Jan 28 04:24:28 2019 perl [...] toby.ink - Status changed from 'open' to 'patched'

Tue Jan 29 07:50:00 2019 perl [...] toby.ink - Correspondence added

Release 1.005_002 which is only a trial version, but includes this feature.

Tue Jan 29 07:50:01 2019 perl [...] toby.ink - Status changed from 'patched' to 'resolved'

Tue Jan 29 23:10:00 2019 violapiratejunky [...] gmail.com - Correspondence added

On Tue Jan 29 07:50:00 2019, TOBYINK wrote: Show quoted text

> Release 1.005_002 which is only a trial version, but includes this feature.

Great, thank you! I'm already using it :)

Bug #128337 for Type-Tiny: Support something similar to Params::ValidationCompiler#named_to_list