https://svn.boost.org/trac10/ticket/13082 <p> I propose to standardise a command-line option to identify the executable as a certain kind of test: I will propose this to boost.test and google.test as well. My personal usecase is the identification of tests for <a class="missing wiki">BoostTestUi</a> (the name is somewhat misleading, it is a test-runner for google-, boost-, catch- and nunit- tests, see github/BoostTestUi.) </p> <p> Currently this requires re-compilation of the test with a special header, I would like to prevent the need for that. </p> <p> --test-runner-identification framework: boost.test version: x.x </p> <p> --test-runner-identification framework: google.test version: x.x </p> <p> --test-runner-identification framework: catch.test version: x.x </p> <p> This will allow test-runners to identify the test-framework used to build the test. That way the test-runners can know what command-line parameters are available to do things like: </p> <ul><li>list all the tests </li><li>run specific tests </li></ul><p> finally, to let the user attach a debugger to the test this kind of code could added in the main(): (for catch that may already exist, I did not check) </p> <p> notice that this option can have any name, we need only 1 standardised option --test-runner-identification, because once we know what we're dealing with, we can just use the respective command-line options. </p> <blockquote> <p> if (arg == "--gui-wait") { </p> <blockquote> <p> std::cout &lt;&lt; "#waiting" &lt;&lt; std::endl; std::getchar(); </p> </blockquote> <p> } </p> </blockquote> en-us /htdocs/site/boost.png https://svn.boost.org/trac10/ticket/13082 Trac 1.4.3 Raffi Enficiaud Mon, 19 Jun 2017 08:56:21 GMT