Add minimal c++ example and check

[aidenfoxivey]

Adding a simple C++ test to CI.

Fixes #1770.

• Does this PR change the input/output behaviour of a cryptographic algorithm (i.e., does it change known answer test values)? (If so, a version bump will be required from x.y.z to x.(y+1).0.)
  Doesn't change any cryptographic algorithms.

• Does this PR change the list of algorithms available -- either adding, removing, or renaming? Does this PR otherwise change an API? (If so, PRs in fully supported downstream projects dependent on these, i.e., oqs-provider will also need to be ready for review and merge by the time this is merged.)
  Doesn't change any cryptographic algorithms.

[vsoftco]

@aidenfoxivey We have a header-only C++ wrapper that provides all the required C++ RAII for liboqs, https://github.com/open-quantum-safe/liboqs-cpp

[aidenfoxivey]

@vsoftco Would it be fair to rework this with the C++ bindings and then continue working on the PR?

[aidenfoxivey]

Absolutely fine to close the PR too, but I feel like it should probably be addressed in the issue first, just to make the procedure clear.

[Martyrshot]

@vsoftco The purpose of this is to catch syntax errors in C++. For example when the stateful-sigs PR was in progress we almost merged something that compiled for C but failed to link against a C++ target.

Since we have #ifdef CPLUSPLUS blocks in our headers I think this check needs to be done with our standard C headers so we catch issues before merging things into liboqs main rather than waiting for the C++ wrappers to fail

[aidenfoxivey]

@Martyrshot Do we want to compile and run the resulting example file? (and/or use valgrind to check it for leaks?) in my existing implementation I'm just compiling and linking it.

[Martyrshot]

I think valgrind for this is overkill, but running it is a good idea. It could catch weird runtime linking errors. Sorry I should have thought about this sooner. Would it be much work for you to add this test to the ninja run_tests rule? Then it would always be run rather then having a specific ci job.

If it's too much work, I can take a look into doing this later in the week.

[aidenfoxivey]

Oh no, don't worry about it! I can absolutely sort it out later tonight or tomorrow.

[aidenfoxivey]

@Martyrshot I thought about it a bit and, if I understand you correctly, this requires modifying the CMakeLists.txt to add C++ as a LANGUAGE. Are you cool with that?

[Martyrshot]

Would we be able to isolate that change to just this test?

[aidenfoxivey]

The ninja run-tests is generated based on CMakeLists.txt right? I think in that case, you pretty much have to change the CMake infrastructure to allow it to build whenever you call ninja run-tests. The reason I didn't have to earlier was because I was just hardcoding a call to g++ in the Github action.

[Martyrshot]

That's fair. Then in that case I think we should avoid modifying CMakeLists.txt. Let's relocate the .cpp file you created into the .github folder somewhere so that it's not mistaken as a core liboqs test. Then have the c++ action run the test and enforce that it doesn't error.

[aidenfoxivey]

Should have this sorted by end of next week. Apologies about the delay - midterms :)

[baentsch]

midterms :)

Good luck!

[Martyrshot]

Should have this sorted by end of next week. Apologies about the delay - midterms :)

No worries! Good luck!