org.dsource.descent.unittests.flute

• A "line" is any number of ASCII characters (that may contain CR, LF or a CRLF pair, although they will only do so if the host program uses them, for example in the text for an assert() statement) followed by a system-specific line terminator.

• r test signature - (An r, followed by a space, followed by the signature or name of a test). Will run the specified test and print the results to stdout. See "test signature specification" for the specification of what the test signatures will look like. See "test result specification" for a specification of what the results will look like. See "Test Naming" for the specification of how test names are handled.
• l - (An l alone on a line). Prints a list of all the tests in the project. One test specification will appear per line. Named tests will appear as their fully qualified test name. Unnamed tests will appear with their signature. The order in which tests are printed is the alphabetical order of their signatures, and lexical order for unittests within the same scope (that is, names hav nothing to do with the order).
• a - (An a alone on a line). Will execute all the tests in the application. For each test, it will write a line containing "Running: " (without the quotes), then the signature of the test being run on a line. Then it will write the results of a test. After running all the tests, a line containing "SUMMARY: " will be written, then all of the following lines: "PASSED: #/#", "FAILED: #/#", "ERROR: #/#", each preceded by three spaces, where the first number sign in a line is replaced by the number of tests that met that condition, and the second in each line is replaced by the total number of executed tests. Any tests that caused an internal error will not be reported in any of the three categories, nor will they be included in the total. There will be a blank line between each test.
• x - (An x alone on a line). Will exit the program.

 module foo.bar;

 unittest { /+ Test 1 +/ }

 class Baz {
     unittest { /+ Test 2 +/ }
 }

 unittest { /+ Test 3 +/ }

• Test 1 is foo.bar.0
• Test 2 is foo.bar.Baz.0
• Test 3 is foo.bar.1

 module bacon.eggs;
 import org.dsource.descent.unittests.naming;

 class Sausage {
     unittest {
         mixin(test_name("spam"));
         // ...
     }
 }

• bacon.eggs.Sausage.0
• bacon.eggs.Sausage.spam
• :spam

• The test could succeed, in which case a line containing "PASSED" will be printed on a line.
• The test could fail an assertion. A line containing "FAILED" will be printed, followed by the stack trace of the exception.
• The test could throw an exception. A line containing "ERROR" will be printed, followed by the stack trace of the exception. The main rationale for treating test failures differently than exceptions is to allow automated tools to track failures vs. error conditions. However, the tool can't differentiate between assertions failed in the tests and assertions failed in the main program body, so a "FAILED" message can mean either.
• An internal error could occur with the test runner (for example, the test is not found). In this case, a human-readable message that does not begin with "PASSED", "FAILED" or "ERROR" will be printed on a single line. If the test is not found, the message will be "Test test signature not found", where "test signature" will be replaced with the signature of the test. Other error messages may appear and are unspecified.

• No Unicode/internationalization support (planned)
• Untested in low memory situations
• Only tested with D 1.x (future versions will support D2)
• No test suites/categorization (will possibly be in a future version, but the Descent front-end should support this when it gets released.)
• Requires Flectioned (not likely to change)

 module foo.bar;

 unittest { /+ Test 1 +/ }

 void baz() {
     class Quux {
         unittest { /+ Test 2 +/ }
     }
 }

 unittest { /+ Test 3 +/ }