Bob Friesenhahn <bfrie...@simple.dallas.tx.us> writes: > The numbers are useful. TAP is documented to support reporting the number > of tests at the end of the test run rather than at the beginning.
Yes, that of course works as well. With the basic C TAP library, you can use plan_lazy() instead of plan() and the library will then figure out the number of tests for you. > Using your TAP test driver, and under the umbrella of Automake, how does > one re-execute just one test in order to see the details of how it > failed? The sample tests in your test harness expect that SOURCE and > BUILD have already been set in the environment by Automake so just > running via the suggested 'runtests -o' results in a failed test. SOURCE and BUILD are only not set when you're building the C TAP Harness source out of its distribution tarball, since that creates a generic runtests binary. When incorporating the test framework into another package, you would do (from README): check_PROGRAMS = tests/runtests tests_runtests_CPPFLAGS = -DSOURCE='"$(abs_top_srcdir)/tests"' \ -DBUILD='"$(abs_top_builddir)/tests"' which will take care of defining SOURCE and BUILD, and then tests/runtests -o <test> will work as expected. -- Russ Allbery (r...@stanford.edu) <http://www.eyrie.org/~eagle/>