On Tue, Apr 02, 2013 at 06:53:28PM +0200, Torsten Bögershausen wrote:

> > I think the check for duplicate-numbers is the only one that does not
> > make sense. 
> []
> Not sure about that, I send a suggestion of a patch in a minute.
> Highlights:
> 1) - rename the contrib test cases and assigns real TC numbers
> 2) - Forward the numbers into the main "test Makefile"

I'm not sure if this is a good idea or not. It puts the
contrib/remote-helpers into the same "number namespace" as the rest of
the test scripts, and enforces uniqueness with test-lint-duplicates,
when "make test" is run from contrib/remote-helpers. But people working
on the main test scripts would not get any such check, and would happily
break contrib/remote-helpers by adding duplicate test numbers.

It makes sense to me to either:

  1. Have the contrib/remote-helpers test live in their own test
     namespace completely, with their own numbers and test-results, and
     pull in relevant bits from the main test harness. We do this
     already with contrib/subtree.  I suggested this when the tests
     first appeared, but there was some argument, and I don't remember
     the details.

  2. Just integrate contrib test scripts into the main repository, but
     leave them off by default. For example, add:

       if test -z "$GIT_TEST_REMOTE_HELPERS"; then
              skip_all="Remote helper tests disabled (define 
GIT_TEST_REMOTE_HELPERS)"
              test_done
       fi

     to the top of the scripts, and then set GIT_TEST_REMOTE_HELPERS
     in contrib/remote-helpers/Makefile before chaining to the test
     Makefile.

-Peff
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to