Explicit cast from void* to corresponding type

[iakov]

Fix compilation without -fpermissive
Originally was PR signal11/hidapi#430

[Qbicz]

In C, void* should be promoted to any pointer type without problem. Do you really compile this code with C++ compiler?

[Youw]

In C, void* should be promoted to any pointer type without problem

At certain compilation flags, even C compiler can start giving warning for the implicit pointer casts, which becomes a real problem with -Werror or smth.
For some projects that's a big dial, and they might use builds systems of their own.

[LudovicRousseau]

Please provide a copy of a C (not C++) compiler warning for this case.
See also https://en.wikipedia.org/wiki/C_dynamic_memory_allocation#Type_safety

For my own code I always include the warning or error message reported by the compiler in the commit message. So that it is clear what problem is solved when reviewing the PR or the commit 10 years later.

[Youw]

My first thought was about -Wconversion which, apparently isn't the case for malloc (i.e. casting from void*) and C compiler.

There is -Wc++-compat which is quite verbose regarding compatibility with C++, and it gives me a second thought.

@iakov why do you need have the ability to compile this source using C++ compiler?

[Qbicz]

Can you provide the warning text that is being fixed in commit message?

[iakov]

Well, yes, my fault. We use C++ as a main language for https://github.com/trikset/trik-studio , and */hid.c files were included into .cpp, so it is a C++ related issue. Nevertheless, if this does not blow up code, and looks correct from pure C point of view, I hope these edits are useful for those like me (who includes .c files into C++ code).

[Youw]

I'm not get use to follow C best practices, so I wouldn't make a last call on this.

But I don't consider including .c/.cpp files in other .c/.cpp files as a good practice unless absolutely necessary.

Although, I don't see anything strictly wrong with this changes.