Enable the collection of test of a directory tree

[flieger19]

No description provided.

[mvandervoord]

I'm not clear on what this update is building? The main test runner means what?

[flieger19]

AFAIK the generate_test_runner.rb script can only generate a runner for a single source file with tests. But not for several source files with tests a single common runner.
Accordingly, I have extended the script so that it can also iterate through a directory tree and scan all matching source files there for tests. These are then collected in a common test runner (main runner).

[tango-taylor]

This would be super useful, I was looking for this exact functionality. I have some tests in C and some in C++ and it seems the easiest way to run them both would be to link one test runner against them.

[tango-taylor]

@flieger19 I gave this a try, I think it's a great start. To be complete I think it should have the following:

• The usage/help text needs to be updated to let the user know it takes directories as input instead of just files.
• Add some unit tests. This gives usage examples "for free" and builds confidence. The last thing we want is the runner generator to drop tests. Looks like test_generate_test_runner.rb would be a good place to start, I'm sure @mvandervoord would know more (I'm not a dev here).
• To keep compatibility with how the script handles files right now, it'd be better if it placed all the tests into one runner file. e.g. I tried it with test.c and test_cpp.cpp and it gave me three output files: test_Runner.c, test_cpp_Runner.cpp, testAll_Runner.c. It would be better if it gave me just testAll_Runner.c (or make the name of the output file mandatory so it doesn't have to make up a name). This might also make it easier to integrate with test_generate_test_runner.rb for unit tests.