Unittests for nsapi_dns

[michalpasztamobica]

Description (required)

Cover the nsapi_dns.h/cpp API with unittests.

Summary of change (What the change is for and why)

nsapi_dns is the last public API not covered with unittests (unlike all other public netsocket APIs).

We are not aiming at 100% coverage, as there are some "convenience" function which basically cast between different types of interfaces. Also - we have good greentea tests in use and there is little point to duplicate them here.
Instead, these tests focus on:

1. Basic scenarios (sync/async), to provide reliable SW base for other tests.
2. Supporting the new getaddrinfo interface (see PR Add Getaddrinfo interface for multiple DNS adresses #11653). Once it gets merged, the scenarios can use the new interface immediately.
3. Testing scenarios which are hard to reproduce in greentea environment (as they require faking the response of the server) - for example multiple retries/attempts. This also applies to point (2).

During the development of these tests some small adjustments had to be made and are submitted in a separate PR #11735.

Documentation (Details of any document updates required)

No need for additional documentation.

Pull request type (required)

[x] Patch update (Bug fix / Target update / Docs update / Test update / Refactor)
[] Feature update (New feature / Functionality change / New API)
[] Major update (Breaking change E.g. Return code change / API behaviour change)

---

Test results (required)

[] No Tests required for this change (E.g docs only update)
[] Covered by existing mbed-os tests (Greentea or Unittest)
[x] Tests / results supplied as part of this PR

---

Reviewers (optional)

@AnttiKauppila
@tymoteuszblochmobica

[ciarmcom]

@michalpasztamobica, thank you for your changes.
@AnttiKauppila @tymoteuszblochmobica @ARMmbed/mbed-os-ipcore @ARMmbed/mbed-os-maintainers please review.

[michalpasztamobica]

We need #11735 to go in first, otherwise tests will not compile.
EDIT: rearranged the sentence.

[michalpasztamobica]

I rebased the code on top of the preceding PR. I checked that it built and ran fine for me locally. I think this is good to go, @0xc0170 .

[0xc0170]

CI started

[mbed-ci]

Test run: FAILED

Summary: 1 of 1 test jobs failed
Build number : 1
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_unittests

[michalpasztamobica]

This is valgrind failure, sorry I forgot to run it. I will investigate and fix.

[michalpasztamobica]

Valgrind's report was valid. It turned out we used delete to delete char arrays allocated with new[].
We should use delete[] instead.
Fix is provided, tests should pass fine now.

[michalpasztamobica]

@0xc0170 , would you restart the CI, please?

[0xc0170]

CI started

[0xc0170]

Early results - unittest failed , CI will report soon

[mbed-ci]

Test run: FAILED

Summary: 1 of 4 test jobs failed
Build number : 3
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_unittests

[michalpasztamobica]

I managed to reproduce after a rebase, so some other commits must have changed the compilation requirements.
Then the tests also went on top of some DNS-related changes, so I had to slightly update the tests.
Everything compiling and passing now (until we need to rebase again :D)

@0xc0170, can we try the CI again, please?

[0xc0170]

Ci restarted

[mbed-ci]

Test run: FAILED

Summary: 1 of 4 test jobs failed
Build number : 4
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_unittests

[michalpasztamobica]

Jenkins runs valgrind with --leak-check=full, hence it found 2 more memory leaks which I missed.
I just pushed a fix to them.
@0xc0170 , would you please trigger the CI?

[0xc0170]

CI restarted

[mbed-ci]

Test run: FAILED

Summary: 1 of 4 test jobs failed
Build number : 5
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_build-GCC_ARM

[michalpasztamobica]

@0xc0170 , this PR sure has no luck with CI... but this time the failure seems unrelated to the code:

[2019-11-06T11:24:37.714Z] [mbed] Adding library "deprecated/BLE_LEDBlinker/shields/TARGET_CORDIO_BLUENRG" from "https://github.com/ARMmbed/cordio-ble-x-nucleo-idb0xa1" at rev #5e49580a0303
[2019-11-06T11:24:37.714Z] [EXAMPLES]> ERROR    mbed-cli deploy command failed for 'mbed-os-example-ble'

Would you take a quick look?

[0xc0170]

There was internal java failure, restarted

[mbed-ci]

Test run: FAILED

Summary: 1 of 4 test jobs failed
Build number : 6
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_build-GCC_ARM

[michalpasztamobica]

Still an internal error, but seems like a different one?

[2019-11-06T12:42:29.926Z] [mbed] Working path "/builds/ws/mbed-os-ci_build-GCC_ARM@5/examples/mbed-os-example-mbed-crypto" (program)
[2019-11-06T12:42:29.926Z] [EXAMPLES]> INFO     In folder 'mbed-os-example-nfc'
[2019-11-06T12:42:29.926Z] [EXAMPLES]> INFO     Executing command 'mbed-cli deploy'...
[2019-11-06T12:42:34.019Z] [mbed] WARNING: Python 3 is not yet fully supported: Python errors may occur when compiling, testing and exporting
[2019-11-06T12:42:34.025Z] ---
[2019-11-06T12:42:34.025Z] [mbed] ERROR: OS Error: File exists

[0xc0170]

I could find again java error, restarted

[mbed-ci]

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 7
Build artifacts

[adbridge]

This is sitting on top of #11735 currently targeting 6.0