While loop not working with bool statements in C (beginner!) - c

Question

I'm learning the basics right now and am making a simple game using a basic graphics library provided by my uni.
There is a projectile thrown (the path of which is drawn)
an obstacle (a wall) - which the projectile can't pass through so must go over
and a target (must also appear solid so line stops being drawn when it is hit)
Right now my projectile is being thrown right through the wall
The issue is with the while loop
Any conditions I add (after (y_displacement < FLOOR_HEIGHT)) have had no effect
(This projectile on its own stops drawing when y_displacement => FLOOR_HEIGHT (Y-Axis is inverted), but any addition of bool statements or attempts at using bool (after #including too) to stop the projectile line being drawn don't make any changes.
TRIED BOOL FOR WALL (doesn't change anything):
bool hit_wall = false;
while ((y_displacement < FLOOR_HEIGHT) && (hit_wall == false))
{
time = (x_displacement - X_HAND) / x_velocity; //speed = distance/time
y_displacement = (Y_HAND - y_velocity * time) - (GRAVITY * pow(time, 2)/2);
GFX_DrawLineTo(x_displacement, y_displacement, 3);
x_displacement += 1;
if ((x_displacement == (X_MAX/2.5)) && ((y_displacement > (Y_MAX/2))))
{
hit_wall = true;
}
}
}
If I can manage to sort this out then I should be able to do the same for my target..
Is there something wrong with what I'm doing?
BACKGROUND:
The full function is this:
void throwBall(int(x_position), int(y_position))
{
GFX_SetColour(YELLOW);
int x_mouse;
int y_mouse;
int x_distance;
int y_distance;
double angle;
float initial_velocity;
float x_velocity;
float y_velocity;
float time;
GFX_MoveTo(X_HAND, Y_HAND);
GFX_GetMouseCoordinates(&x_mouse, &y_mouse);
x_distance = x_mouse - X_HAND;
y_distance = y_mouse - Y_HAND;
angle = getAngle(x_distance, y_distance);
initial_velocity = sqrt(pow(x_distance, 2) + pow(y_distance, 2));
//these have been divided by 5 as the magnitude was too large for the window
y_velocity = (initial_velocity * sin (angle) - GRAVITY * time)/(X_MAX/256);
x_velocity = (initial_velocity * cos (angle))/(X_MAX/256);
float y_displacement;
float x_displacement;
x_displacement = X_HAND;
y_displacement = Y_HAND;
bool hit_wall = false;
while ((y_displacement < FLOOR_HEIGHT) && (hit_wall == false))
{
time = (x_displacement - X_HAND) / x_velocity; //speed = distance/time
y_displacement = (Y_HAND - y_velocity * time) - (GRAVITY * pow(time, 2)/2);
GFX_DrawLineTo(x_displacement, y_displacement, 3);
x_displacement += 1;
if ((x_displacement == (X_MAX/2.5)) && ((y_displacement > (Y_MAX/2))))
{
hit_wall = true;
}
}
}

Answer 1

if ((x_displacement == (X_MAX/2.5)) && ((y_displacement > (Y_MAX/2))))
{
hit_wall = true;
}
you use flaot values, and due to some reasons (values are aprroximatelly) you almost never get in this if.
you should use >, >=, < , <= poerators, but it also will not help you in 100% of cases.
best choice will be to use epsilon (Observational error, Neighbourhood)
like
float eps = 0,0000001f;
if ( abs(x_displacement - (X_MAX/2.5) ) < eps
&& y_displacement > ( (Y_MAX/2) - eps ) )
{
hit_wall = true;
}

Answer 2

expanding on previous answers regarding floating point equivalence and round off error, another way to do it is multiply your float or double value by some factor of 10 to move the decimal point to the right, then cast it to an int or long it. Then you can do a boolean comparison which will be on integers which will work reliably. I mention this because it can sometimes be easier to code and also read easier than doing the epilson method (which you would also have to do an absolute value on) described in the other answer .
The big thing to recognize is the factor of 10 by which you multiply with or set epsilon to.
Do not go crazy and multiply by something like 1e9 or set epsilon = 1e-9 when you don't need to because that can re-introduce the same problem.
For example if your variables for distance are in meters then you need to recognize you do not need resolution greater than ???. Let's use 1/100 of a millimeter as an example... as far as you are concerned 1.230012345mm == 1.230456789mm. So then multiply by 1e5 then cast to int or do epsilon=1e-5;. Same principle applies for units of inches, you typically do not need resolution greater than 0.0001" for anything so multiply by 1e4 or epsilon = 1e-4.
/* example */
# define D2I 1e2 /* dbl to int factor, 0.01mm resolution */
double x_displacement; /* units of millimeters for this example*/
double y_displacement; /* units of millimeters for this example*/
/* whatever previous calculations result in...
x_displacement having value of 3.00000249687738566
y_displacement having value of 120.00000085639820699
out past 6 or so decimal places a float or double will have arbitrary numbers, so don't multiply by some huge 1e12 number
*/
if ( ( (int)(x_displacement * D2I) == (int)((X_MAX / 2.5) * D2I ) ) &&
( (int)(y_displacement * D2I ) > (int)((Y_MAX / 2 ) * D2I ) )
)
{
hit_wall = true;
}