TCP/TLS Socket tests will skip if TCP is not supported

[michalpasztamobica]

Description

I generally avoid using macros, but I couldn't come up with a better way in this case.
Two reasons why macros are used to wrap the open calls:

1. I wanted to minimise the "boilerplate" which gets copied and pasted into every test.
2. The SKIP_TEST_MESSAGE macro contains a "return;" which is intended to return from a current test code. Therefore we cannot use it inside any wrapper function, as then it would merely return from that function, not from the test as such.

The architectures which could make use of this (for example UBLOX_C030_N211) are not available in any RAAS server, but I agreed with @AriParkkila that he will make sure this executes correctly.

Pull request type

[ ] Fix
[ ] Refactor
[ ] Target update
[ ] Functionality change
[ ] Docs update
[x] Test update
[ ] Breaking change

Reviewers

@SeppoTakalo
@KariHaapalehto
@VeijoPesonen
@mtomczykmobica
@tymoteuszblochmobica
@AriParkkila

[SeppoTakalo]

Uh.. You completely lost me with those macros..

I was hoping a way more simpler solution, like:

bool is_tcp_supported() {
    static bool supported;
    static bool tested = false;
    if (!tested) {
        TCPSocket socket;
        supported = socket.open(NetworkInterface::get_default_instance()) == NSAPI_ERROR_OK;
    }
    return supported;
}

void TCPSOCKET_ECHOTEST()
{
    if (!is_tcp_supported()) {
        TEST_SKIP_MESSAGE("TCP not supported");
    }
    ...
}

Add extra 3 lines per TCP test case.. should not be too much.

[ciarmcom]

@michalpasztamobica, thank you for your changes.
@tymoteuszblochmobica @AriParkkila @VeijoPesonen @SeppoTakalo @KariHaapalehto @mtomczykmobica @ARMmbed/mbed-os-ipcore @ARMmbed/mbed-os-test @ARMmbed/mbed-os-maintainers please review.

[michalpasztamobica]

I did not consider this option to avoid influencing test results by creating, opening and closing an extra socket before every (or at least one) test routine.
I was trying to code it in situ .
But if you don't mind the one extra socket operation, we can go ahead with your solution.

In this case however, I would put it on top of #9959 and add it to the test case setup function, rather than copy it into the code of every test function...

[0xc0170]

Review astyle travis failing job

[michalpasztamobica]

@0xc0170 I plan to rewrite the whole code, according to Seppo's suggestion, but I need to get #9959 merged, so I can use the greentea test case setup function...

[michalpasztamobica]

@SeppoTakalo , unfortunately I could not put the SKIP macro inside the testcase's setup function, because the macro returns void and the wrapper needs to return a value.

But I did the second best thing possible - I added the function as you suggested and wrapped it with a simple macro, which got added to every test case.

To test the changes I locally broke LWIP::socket_open() function to return NSAPI_ERROR_UNSUPPORTED when TCP protocol is in use. All TCP and TLS tests printed the SKIP message and showed up as passing in the summary.

Reviewers, please review :)

[cmonr]

@michalpasztamobica Could you check the astyle job results?

[cmonr]

Also, it looks like the preceeding PR has been merged.

[michalpasztamobica]

astyle fixed and the preceding PR is indeed merged (although it turned out I cannot make use of it in the end).
@SeppoTakalo , I'll be grateful for your approval.

[cmonr]

CI started

[mbed-ci]

Test run: SUCCESS

Summary: 6 of 6 test jobs passed
Build number : 1
Build artifacts

[cmonr]

@SeppoTakalo @michalpasztamobica Should we wait on others to chime in?

[mtomczykmobica]

Looka good.

[michalpasztamobica]

@cmonr , I think there was enough time, I'd suggest to merge in.

[cmonr]

Thank you!

[cmonr]

CI started

[mbed-ci]

Test run: SUCCESS

Summary: 6 of 6 test jobs passed
Build number : 2
Build artifacts