Problem converting c to cpp with matrices

[Ghoti]

Hi folks,

Hi folks,

i use the matrix functions that were part of the examples in the SDK:

Code: Select all

void matrix_multiply(float* result, float* a, float* b)
	{
		unsigned int i,j,k;
		float temp[16];

		for &#40;i = 0; i < 4; ++i&#41;
		&#123;
			for &#40;j = 0; j < 4; ++j&#41;
			&#123;
				float t = 0.0f;
				for &#40;k = 0; k < 4; ++k&#41;
					t += a&#91;&#40;k << 2&#41;+j&#93; * b&#91;&#40;i << 2&#41;+k&#93;;
				temp&#91;&#40;i << 2&#41;+j&#93; = t;
			&#125;
		&#125;

		memcpy&#40;result,temp,sizeof&#40;float&#41;*16&#41;;
	&#125;

when i use this in a C source file i use it like this:

Code: Select all

matrix_identity((float*)&world);


		
			matrix_identity((float*)&tmpworld);
			
			matrix_multiply((float*)&world, &world, &tmpworld);

this works fine.

since its part of a 3d game i want it to be in cpp cause of the object oriented way of programming.
well i copied the code but i got an error with that exact code:

Code: Select all

bj.cpp:165: error: cannot convert 'ScePspFMatrix4*' to 'float*' for argument '2' to 'void matrix_multiply(float*, float*, float*)'
O

so i tried in doing (float*) infront of the &world, and &tmpworld. This stopped the error but my code is not running anymore.

my question: Why this difference in code? why does my old code won't work anymore?

[rapso]

how du u declare world and tmpworld? probably this are already pointers?

[cooleyes]

Hi folks,

Hi folks,

i use the matrix functions that were part of the examples in the SDK:

Code: Select all

void matrix_multiply(float* result, float* a, float* b)
	{
		unsigned int i,j,k;
		float temp[16];

		for &#40;i = 0; i < 4; ++i&#41;
		&#123;
			for &#40;j = 0; j < 4; ++j&#41;
			&#123;
				float t = 0.0f;
				for &#40;k = 0; k < 4; ++k&#41;
					t += a&#91;&#40;k << 2&#41;+j&#93; * b&#91;&#40;i << 2&#41;+k&#93;;
				temp&#91;&#40;i << 2&#41;+j&#93; = t;
			&#125;
		&#125;

		memcpy&#40;result,temp,sizeof&#40;float&#41;*16&#41;;
	&#125;

when i use this in a C source file i use it like this:

Code: Select all

matrix_identity((float*)&world);


		
			matrix_identity((float*)&tmpworld);
			
			matrix_multiply((float*)&world, &world, &tmpworld);

this works fine.

since its part of a 3d game i want it to be in cpp cause of the object oriented way of programming.
well i copied the code but i got an error with that exact code:

Code: Select all

bj.cpp:165: error: cannot convert 'ScePspFMatrix4*' to 'float*' for argument '2' to 'void matrix_multiply(float*, float*, float*)'
O

so i tried in doing (float*) infront of the &world, and &tmpworld. This stopped the error but my code is not running anymore.

my question: Why this difference in code? why does my old code won't work anymore?

you must use force typecast
just like

Code: Select all

matrix_identity((float*)&tmpworld); 
matrix_multiply((float*)&world, (float*)&world, (float*)&tmpworld);

[Ghoti]

@rapso: I declare it in both situations this way:

Code: Select all

			ScePspFMatrix4	world;
			ScePspFMatrix4	tmpworld;

the only difference is that i this in cpp for in a class header and in the C source i do it just at the top of the source file

@cooleyes: yes i have done that, but why that difference, why do i have to do that in cpp and not in c?

[chp]

Because C++ is not as relaxed when it comes to type checking as ordinary C is, and thats a good thing.

If you intend to always access ScePspFMatrix4-objects using that method, rewrite it to take a pointer of that type, or write a inlined overload that does the conversion for you.

[Ghoti]

ah okay thanks :)