The "right" thing to do is to put it in xt/ without any skips and make
sure the release machinery is scripted to make sure xt tests pass before
release.
David
On Mon, Feb 29, 2016 at 3:14 PM, Olivier Mengué <olivier.men...@gmail.com>
wrote:
> > The common solution is to NOT add "author-only" tests to TEST_REQUIRES,
> but instead to modify your "skip_all" as follows:
> >
> > unless( ( $ENV{'RELEASE_TESTING'} || $ENV{'AUTHOR_TESTING'} ) && eval
> 'use Test::POD; 1;' ) {
> > plan skip_all => 'These tests are for authors only.';
> > exit 0;
> > }
>
> It is the right thing to skip POD tests if !$ENV{RELEASE_TESTING} &&
> !ENV{AUTHOR_TESTING}.
>
> But it is a bad practice to skip the test if the needed module is missing.
> Instead the test should fail if the author lacks the module because if the
> test is skipped, there is the risk of publishing a release that was not
> properly tested and would have invalid POD. What is the point of writing
> tests if they can be silently skipped?
>
> Olivier.
>
> 2015-05-22 16:54 GMT+02:00 David Oswald <daosw...@gmail.com>:
>
>> The common solution is to NOT add "author-only" tests to TEST_REQUIRES,
>> but instead to modify your "skip_all" as follows:
>>
>> unless( ( $ENV{'RELEASE_TESTING'} || $ENV{'AUTHOR_TESTING'} ) && eval
>> 'use Test::POD; 1;' ) {
>> plan skip_all => 'These tests are for authors only.';
>> exit 0;
>> }
>>
>> Or you could split the ENV testing and the "use" testing into different
>> logic paths so that you can give diagnostic messages more appropriate to
>> the failure point. In other words, if the ENV isn't set, you can skip with
>> a message of "Skipping author tests, set ENV{.....} for author testing.",
>> and if the eval fails you can state "Author tests require Test::POD".
>>
>>
>>
>> On Thu, May 21, 2015 at 8:07 PM, Alceu R. de Freitas Jr. <
>> glasswal...@yahoo.com.br> wrote:
>>
>>> Since I'm the author of this module, I think I can have a say about
>>> that. :-)
>>>
>>> Meanwhile I agree to remove Test::Pod from PREREQ_PM, I think that
>>> loosing the opportunity to test the module if test Test::Pod is indeed
>>> available is not interesting.
>>>
>>> I could use a conditional test to execute the test or not as simple as
>>> described in Test::Pod Pod:
>>>
>>> use Test::More;
>>> eval "use Test::Pod 1.00";
>>> plan skip_all => "Test::Pod 1.00 required for testing POD" if $@;
>>> all_pod_files_ok();
>>>
>>> That should be enough for all cases.
>>>
>>> My questions are:
>>>
>>> 1 - Should I declare Test::Pod on TEST_REQUIRES? Documentation is not very
>>> clear on how that works if the module is not available. If I think I should
>>> require a user to download Test::Pod to run the tests? Yes, I think so.
>>> It's a very small price to give a hand to the project (since Test::Pod is
>>> not exactly the same thing on installing a large set of modules
>>> dependencies as happens with Moose, for example), but I'm open to discuss
>>> about it.
>>> 2 - Does it matter? I wrote this module years ago when I was on a network
>>> that wouldn't allow me to send those tests results, so I'm not sure if this
>>> module still works with more recent versions of Microsoft Outlook. Maybe
>>> Barbie could have say about that (since it depends mostly on Mail::Outlook)?
>>>
>>> PS: just checked
>>> http://cpansearch.perl.org/src/ARFREITAS/Test-Reporter-Transport-Outlook-0.01/Makefile.PL
>>> and this module really needs an update.
>>>
>>> Cheers,
>>>
>>> Alceu
>>>
>>> ------------------------------
>>> *De:* Karen Etheridge via RT <
>>> bug-test-reporter-transport-outl...@rt.cpan.org>
>>> *Para:*
>>> *Enviadas:* Quinta-feira, 21 de Maio de 2015 15:28
>>> *Assunto:* [rt.cpan.org #104607] incorrect case
>>>
>>> Queue: Test-Reporter-Transport-Outlook
>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=104607 >
>>>
>>> On 2015-05-21 09:43:17, CHORNY wrote:
>>> > depends on Test::POD. Correct is Test::Pod.
>>>
>>> However, the distribution should not have a dependency on this module at
>>> all. Pod tests should not run for normal user installs - either move the
>>> test to xt/, or guard it with:
>>>
>>> plan skip_all => 'These tests are for authors only!' unless
>>> $ENV{AUTHOR_TESTING} or $ENV{RELEASE_TESTING};
>>>
>>>
>>>
>>
>>
>> --
>>
>> David Oswald
>> daosw...@gmail.com
>>
>
>
--
David Golden <x...@xdg.me> Twitter/IRC/Github: @xdg
