Unity: Add macros for test skipping.

[davidsaada]

Description

The TEST_SKIP and TEST_SKIP_UNLESS macros (w/ and w/o messages) allow the test to skip the test case execution from the point the macro was called (without failing the test).
This is useful in case the conditions (such as missing H/W, memory limitations etc.) don't allow the test to run properly.

Pull request type

[x] Fix
[ ] Refactor
[ ] New target
[ ] Feature
[ ] Breaking change

[davidsaada]

@dannybenor @bulislaw @jamesbeyond

[adbridge]

OK this fix slightly subverts the way Utest/Unity are supposed to work. You may want to read the Utest readme, in particular:

Test Case Attributes

You can modify test case behavior by returning control_t modifiers:

    CaseNext: never repeats and immediately moves to next test case
    CaseNoRepeat: never repeats.
    CaseRepeatAll: repeats test case with setup and teardown handlers.
    CaseRepeatHandler: repeats test case without set and teardown handlers.
    CaseNoTimeout: immediately moves to next test case.
    CaseAwait: waits indefinitely for callback validation (use with caution).
    CaseTimeout(uint32_t ms): waits for callback validation for ms milliseconds, times out after that (fails with REASON_TIMEOUT).
    CaseRepeatAllOnTimeout(uint32_t ms): waits for callback validation for ms milliseconds, repeats test case with setup and teardown handlers on time out.
    CaseRepeatHandlerOnTimeout(uint32_t ms): waits for callback validation for ms milliseconds, repeats test case without setup and teardown handlers on time out.

From what I can make out your new Unity macro just does an immediate 'return' to skip executing the rest of a test case, however if the test case has been designed to return with one of the above custom control_t types then have you investigated what the knock on effect would be ? ie would it just break Utest?

[adbridge]

Why isn't this just -
if (!(condition) {UNITY_TEST_SKIP((line), (message))} ?
Note also other brackets around parameters

[davidsaada]

Right on both. Will change. Regarding the condition, it was just a copy/paste from the UNITY_TEST_ASSERT macro, but that's not an excuse.

[davidsaada]

Pushed a fix now.

[davidsaada]

From what I can make out your new Unity macro just does an immediate 'return' to skip executing the rest of a test case, however if the test case has been designed to return with one of the above custom control_t types then have you investigated what the knock on effect would be ? ie would it just break Utest?

The vast majority of the tests return nothing. Only tests that return the control_t type are in storage_abstraction, which is deprecated, and in the Utest tests themselves of course. Using the macro in these tests would produce an error/warning for not returning a return value. In all other tests (practically all tests), using this macro would simply end the function.
I can add macros with a return value, but again as I see it, it's not used anywhere in the code.
This macro is really intended as a convenient one liner for gracefully ending a test case, which is something which I believe is needed. So instead of having:

if (!normal_state) {
  printf(__LINE__)
  printf(message);
  return;
}

Just call TEST_SKIP_UNLESS(normal_state, message)
But anyway, will add a comment for clarity sake.

[OPpuolitaival]

Looks good

[0xc0170]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 2601
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7448/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2245
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7448/

[mbed-ci]

Test : SUCCESS

Build number : 2354
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7448/2354

[fkjagodzinski]

This might come in handy. 👍