Ovid wrote: > Hi all, > > This is probaby a weird request and it's something I could use, but not > everyone will. Further, it might require that Test::Harness be changed and I > don't know if this is likely or reasonable. > > Basically, when I run a single test, I do something like this: > > prove t/test_name.t > > However, if I'm using Test::Class, I do something like this: > > prove t/test_class_tests.t > > Typically, Test::Class tests have a wrapper script which loads the classes > and runs the tests. However, I have a different wrapper script > (http://use.perl.org/~Ovid/journal/30510) which allows me to do this:
Actually, the way I use Test::Class is to have one script per class. That way I can just run one if I want to. > perl t/test_class_tests.t path/to/test/class.pm > > Instead of loading all classes, it merely loads the one class I want to test. > This is handy when I'm working on a single class and I don't want to run all > 4,000 friggin' tests. > > What I would prefer to do is something like this: > > prove t/test_class_tests.t path/to/test/class.pm > > I want to do that so I can use Test::Harness, but the path is considered an > argument to prove and not my Test::Class driver wrapper script. > > From what I can tell, Test::Harness doesn't support anything like this > directly and naturally prove doesn't either. Can anyone make any > recommendations on a good approach here? My ideas are terrible things like > setting environment variables and other things I would simply prefer to avoid. Test::Harness doesn't need to support this. I'm pretty sure that you can just make your modules runnable from the command line. http://www.perl.com/pub/a/2006/07/13/lightning-articles.html?page=4 -- Michael Peters Developer Plus Three, LP
