Fixing C warnings

[Amaras]

Currently, no warnings are activated during compilation for C.
Because of this, the code might be inconsistent in certain places.
I aim to fix or remove some errors in the C code, depending on if we consistently violate warnings or not.

Currently, only the following algorithms build without warnings:

• approximate_counting
• computus
• forward_euler_method
• graham_scan
• jarvis_march
• monte_carlo_integration
• thomas_algorithm
• tree_traversal
• verlet_integration

The main errors I see in the rest of them are sign-compare, unused-parameter and unused-variable, along with a implicit-function-declaration and two return-type.
I did not implement any changes to the code yet, but I'll do that in the following days (hence why it's a draft PR).

[Amaras]

It seems that, actually, I have done immediately exactly what I have said I would not do.

Anyway, I did 1 commit per algorithm fix, so that reverting is easier.

I can tell there is probably something wrong on the Gaussian elimination algorithm because I could completely eliminate the rows parameter in one function... This kind of fix was to brutally remove unused variables and parameters and adjust callers.

The rest of the fixes are simply fiddling around between size_t and int until the compiler stops throwing signed-compare errors.

[ntindle]

Does this mean that Gaussian is wrong as a whole?

[Amaras]

Does this mean that Gaussian is wrong as a whole?

I have not checked the correctness of the code yet, so I can't answer properly. It's just that, intuitively, there's something wrong if you're able to not pass one of the dimensions of the matrix.

EDIT: Also, I did not run the executables. Seeing how it's coded, the Gaussian elimination is probably going to segfault if we step outside the assumptions that were made.

[Amaras]

Okay, I think I know the basic assumption of the Gauss-Jordan algorithm's implementation. We assume that we have an n*n matrix, and we augment it with the vector, which means that we have an n*(n+1) matrix. As such, we technically don't need the rows parameter, since it's just cols - 1

[stormofice]

Hi, thanks for the fixes!

Overall the changes look good to me, however there is a small typo in the gaussian elimination code.

Apart from that, it is ready to be merged.

[stormofice]

This needs to be i >= 0, otherwise the loop body will never be executed