David Cantrell wrote:
On Fri, Jul 31, 2009 at 10:51:57AM -0700, Jonathan Swartz wrote:
Is there a standard for signifying internal-only tests, and for make test to figure out when they should run?

The normal way is to have them skip unless some magic environment
variable is set.

Perl::Critic used to do this and then we'd get bug reports from people who used 
the same environment variables to run their author tests.  Also, Mr. Lester had 
the tests run if a .svn directory existed.  Of course, someone then did a 
checkout and complained when those tests failed.

So, you now have to actively request that the P::C author tests be run.  All 
the author tests are in xt/author.  There is a custom Module::Build authortest 
target.

The basic logic of that is

   $self->test_files( qw< t xt/author > );
   $self->recursive_test_files(1);
   $self->depends_on('test');

http://perlcritic.tigris.org/source/browse/*checkout*/perlcritic/trunk/distributions/Perl-Critic/inc/Perl/Critic/Module/Build.pm

Reply via email to