Well, I had to reduce rotation step to 4° and translation speed to 0.1 because it was too fast

Actually  I successed making a basic engine work (flat raycasting)
I'll post a screenshot in an hour or so, and I'll start adding textures

To make it clear, my goal is to write a fully reusable raycasting engine, for those who want to write Wolfenstein 3D- like games (not Doom-like though, I didn't plan to support floors, stairs or things like that. Maybe variable-height walls.).

Which was that rayX and rayY went outbounds at the really start of the execution (which is in either way a maths problem, since this should never happen with the map I used). I'll redo the engine with floats.

Jacobly helped me on IRC so I was able to get rid of the crash, but maths are still wrong.

<btw, thanks for saying that I'm coding like shit Levak>

So here's the new code :


Once again, tell me if you spot any math error. I'm having a hard time with scaling integers to have a fixed point.

I'd looove to play the Binding of Isaac on-calc good luck with it, hope you'll finish it.

Hallaw,

I'm stuck with a weird error when programming with Ndless. The emulated Ti-Nspire CX (not sure if CAS) crashes when executing :
int* D3_sortFaces(Vertex *vertexTable, Vertex camera, int **faces, int numberOfFaces, int *verticesPerFace)
{
int i, j, flag, intermediate;
char *order;
int *zAverage;

order = malloc(numberOfFaces * sizeof(char));
if(order == NULL)
{
printf("error with order");
return NULL;
}

zAverage = malloc(numberOfFaces * sizeof(int));
if(zAverage == NULL)
{
free(order);
printf("error with zAverage");
return NULL;
}

for(i = 0; i < numberOfFaces; i++)
{
order[i] = i;
zAverage[i] = 0;
for(j = 0; j < verticesPerFace[i]; j++)
{
printf("%d,%d\n",i,j);
zAverage[i] += vertexTable[faces[i][j]].z - camera.z; // crashes here
}
zAverage[i] /= verticesPerFace[i];
}
printf("zAverage passed");
// etcGiven :
char faces[6][4] = { {0, 1, 2, 3}, {4, 5, 6, 7}, {0, 4, 7, 3}, {3, 2, 6, 7}, {0, 1, 5, 4}, {1, 5, 6, 2 } };
char *order;
int vpf[6] = { 4, 4, 4, 4, 4, 4 };


// ...
// rotated is a 8 elements-large table of bunches of 3 values
order = D3_sortFaces(rotated, (Vertex){0, 0, -150}, faces, 6, vpf);
I don't know if you understand well, but I don't want to release a complete source for now

What this does is that it displays "0,0", \n then "Reset" and → crash.

If you have any idea of what's wrong, please tell me

I'm sure you can do it games are meant to be beaten eh ?

Bad news ...

I can't write a proper clipped triangle. I don't know why, but this just doesn't want to work. I re-read the code many many times and still can't see the error. If you guys wanna help me with this, here is the complete AxeJh3D code at pastebin.

Actually, clipping works perfectly with y < 0, y > 63, x > 95 but not with x < 0. I also attached a small Axe program to test the routine, if you want to see what's wrong or to fix it.

I need to explain some things related to the way I draw the triangle. First, I constantly use 3 variables : startX1 and startX2, two 16.8 numbers, which refers to the left and right pixels of the line to draw (because I draw the triangle with horizontal lines), startY, which is the Y coordinate of the line. I also constantly add two 8.8 variables to startX1 and startX2, namely dx1 and dx2, which are the slope of each side (dx1 is updated during the process to the slope of the 3rd edge of the triangle). I did it that way because with only 8.8 startX*, the triangle went too fast from 127 to -128, making the triangle being drawn wrongly. So by doing that, my goal was to fix that error, since it's harder to go from 32767 to -32768.

EDIT : with help from Runer112 on IRC, we made it that the triangle is properly clipped when x < 0 ; but the 127→-128 thing still appears, so the problem is not solved yet.