ATan2

I'm using the atan2 function to work out the angle to get to a waypoint (in 2d)and then angle the sprite in the correct direction.

Because atan2 doesn't work in 360 - just 180 I'd thought just add 180 to the result - but that doesn't work.

I'm sure it's something very simple, but I can't see the solution atm.

I've worked around the problem by rotating the sprite to 180 degree's and loading using loadanimimage as usual.

Here's my code...

angle# = ATan2(y_dist#,x_dist#)
If angle# < 0.0 angle# = angle# + 360.0

The blitz rotation system in 2d is a bit funny. 0 deg is the sprite/image moving right. Also, the angle returned is between -180 and 180 i'm sure. So..

angle# = ATan2(y_dist#,x_dist#) 
If angle# < 0.0 angle# = angle# + 360.0

try

angle# = ATan2(y_dist#,x_dist#) 
angle# = angle# + 180.0

And don't rotate the sprite/image to 180. :o)

Yes, I noticed this with Blitz3D when I was doing a 3D compass for games - it doesn't work on 360 degrees - you end up with what "Ross C" said at the start of his post.

EDIT:
A text version (of my rough compass) looks like this:

compass#=EntityYaw(campiv)
Print compass
If compass>-45 And compass<45 Then Locate 440,10: Print "NORTH"
If compass<-45 And compass>-135 Then Locate 440,10: Print "EAST"
If compass<-135 Or compass>135 Then Locate 440,10: Print "SOUTH"
If compass>45 And compass<135 Then Locate 440,10: Print "WEST"

Ross, no I tried that before and my character moves away from the waypoint!

ATan2() returns the theta component from standard polar coordinates. In polar coordinates, 0 degrees is facing positive X, i.e. to the right. From 0 to 90 degrees will bring you to positive Y, that would normally be up, but since the Y coordinate is flipped on the screen, that'll point your object downward.. From 0 to -90 will point you into -Y. 90 to 180 is -X, +Y, -90 to -180 is -X, -Y.
So:

     -90
      |
      |
-180  |     0
------O------
180   |
      |
      |
      90

It's much easier to use if you start using -179 to 180 all the time instead of 0 to 359.

Thanks for the info - In that case I'll stick with my workaround as this seems the simplest way around the problem.