cynara-test depends on random initialization of global instances about cynara

[pohly]

m_wildcardValue was just the tip of the iceberg. Several other global instances had similar issues. Some of that code is also used in cynarad, so this issue might go beyond just testing.

For the fixes that were good enough for me at the moment, see:
https://github.com/pohly/meta-intel-iot-security/blob/cynara-tests/meta-security-framework/recipes-security/cynara/cynara/PolicyKeyFeature-avoid-complex-global-constants.patch
https://github.com/pohly/meta-intel-iot-security/blob/cynara-tests/meta-security-framework/recipes-security/cynara/cynara/globals-avoid-copying-other-globals.patch

However, you might want to do a more thorough code review and avoid this kind of issue completely. I only fixed the cases that really caused segfaults.

[JacquesBurac]

Thanks for finding these issues and for the patches. Yes, we need to get rid of direct std::string -> std::string dependencies. First patch looks good to me.
The second patch if I understand static variable initialization correctly makes it less likely to fail, because there are less variables that we depend on. However, we still rely on variable being referenced to be initialized first (references can point to variable that have not yet been initialized).
I think that all std::string constants should depend on:
a) C string literals/variables
or
b) wrapper functions like in the first patch

[pohly]

The second patch if I understand static variable initialization correctly makes it less likely to fail,
because there are less variables that we depend on. However, we still rely on variable being
referenced to be initialized first (references can point to variable that have not yet been initialized).

Correct. It's just mitigating the problem without solving it properly.

I agree with the proposed solution, but won't have time to implement it, I'm afraid.

[JacquesBurac]

Ok, I understand. Right now I'm working on some urgent task, but I should finish soon and I'll be able to take care of it.

[JacquesBurac]

I've pushed your first patch to gerrit:
https://review.tizen.org/gerrit/#/c/41353

And also made some more changes:
https://review.tizen.org/gerrit/#/c/41354