=head1 Testing Your Distribution

Dist::Zilla saves you from wasting a lot of time on things like boilerplate
documentation and bumping version numbers.  You can take that free time and
channel it into writing better tests, right?  Right!  Of course, with tests
written, you'll want to be able to run them.

=head2 Using C<prove> to run tests

No matter what configuration you're using, testing a distribution will usually
be possible with the very simple shell command:

  $ prove -lr t

This uses the L<prove> program to run the test files under F<./t> using the
libraries in F<./lib>.  Dist::Zilla won't be involved, so it can run very
quickly.  There are a few things to keep in mind, though:

=for :list
* If your C<$VERSION> variables come from Dist::Zilla, they won't be present.
* If Dist::Zilla constructs files needed for your tests, they won't be there.
* If your dist uses XS, this pretty much won't work.

=head2 Using C<dzil test> to run tests

If your dist needs to be built for your tests to run, or if you just want to
make sure your tests still run with the built version of your dist, you can run
another simple command:

  $ dzil test

This will build your distribution and run its tests in the build directory.  It
should be exactly the same as what an installing user would be trying to do.
If all the tests pass, the build directory will be deleted afterward.  If tests
fail, it will be left in place so you can inspect the wreckage.

=head2 Running your tests before releasing

You should run your tests B<immediately> before you release your code, to be
sure that the thing you're about to upload is working.  You can tell
Dist::Zilla to run your tests as part of the release process, aborting the
release if they fail, but continuing on if they're all successful.

=for cyoa
? how-build-works ? learn about how Dist::Zilla builds your dist
? release         ? learn how to release your newly-tested dist