I see nothing unsafe about either patch, so long as the meta-spec is
specified as version 2. Can you clarify your concerns?

Perhaps reading the section in prerequisite phases in the meta spec might
clarify things a bit:
https://metacpan.org/pod/CPAN::Meta::Spec#PREREQUISITES

(PS. I found the discussion in
https://rt.cpan.org/Ticket/Display.html?id=112568)


-ETHER


On Mon, Feb 29, 2016 at 6:17 PM, Ron Savage <r...@savage.net.au> wrote:

> Hi
>
> Kent Fredric has proposed 2 alternate patches to my most recent
> Makefile.PL for Tree::DAG_Node (V 1.28).
>
> The point is to completely remove the need for an end-user of this - and
> of other - modules to have Test::Pod installed.
>
> It's not clear from these patches, but the upper part of both is under
> TEST_REQUIRES, and he warns me against switching to BUILD_REQUIRES.
>
> a.patch
>
> --- Makefile.PL.3       2016-03-01 14:41:14.147052412 +1300
> +++ Makefile.PL.1       2016-03-01 14:39:20.713057559 +1300
> @@ -38,7 +38,6 @@
>                 'File::Spec'    => 3.40,
>                 'File::Temp'    => 0.19,
>                 'Test::More'    => 1.001014,
> -               'Test::Pod'             => 1.48,
>         },
>         VERSION_FROM => 'lib/Tree/DAG_Node.pm',
>  );
>
>
> b.patch
>
> --- Makefile.PL.3       2016-03-01 14:41:14.147052412 +1300
> +++ Makefile.PL.2       2016-03-01 14:40:09.053055429 +1300
> @@ -38,7 +38,6 @@
>                 'File::Spec'    => 3.40,
>                 'File::Temp'    => 0.19,
>                 'Test::More'    => 1.001014,
> -               'Test::Pod'             => 1.48,
>         },
>         VERSION_FROM => 'lib/Tree/DAG_Node.pm',
>  );
> @@ -67,6 +66,13 @@
>                                 web => '
> https://github.com/ronsavage/Tree-DAG_Node',
>                         },
>                 },
> +               prereqs => {
> +                       develop => {
> +                               requires => {
> +                                       'Test::Pod' => 1.48
> +                               }
> +                       }
> +               }
>         };
>  }
>
> Clearly a.patch is safest because it's minimalist. But is b.patch 100%
> safe? Neither he nor I know. Does anybody?
>
> TIA.
> --
> Ron Savage - savage.net.au
>

Reply via email to