+- QB64 Phoenix Edition (https://staging.qb64phoenix.com)
+-- Forum: QB64 Rising (https://staging.qb64phoenix.com/forumdisplay.php?fid=1)
+--- Forum: Code and Stuff (https://staging.qb64phoenix.com/forumdisplay.php?fid=3)
+---- Forum: Help Me! (https://staging.qb64phoenix.com/forumdisplay.php?fid=10)
+---- Thread: 2D Physics Engine Help (/showthread.php?tid=1749)
Pages:
1
2
2D Physics Engine Help - TerryRitchie - 06-12-2023
Over the past few months I've made a few attempts at creating a 2D physics library for QB64 but have failed miserably. My first few attempts were writing something from scratch. I quickly realized that while I have a fair grasp on basic trig and vector math I have nowhere near the knowledge to implement such things as angular momentum, raytracing, and 2D collision physics. Even after trying to tutor myself on the subject I seem to still be just as confused (if not more).
I then decided to follow a few video tutorials I found on Youtube related to creating a 2D physics engine. Of course these are all either meant for C++, Java, or JavaScript using OOP. I figured the process of converting the Java OOP code to functions and procedures would be fairly straight forward ... not so much. (Below I provide a link to the video series I am following along with the code I created so far).
My next thought was to incorporate the Box2D physics engine (the engine that Rovio used to create Angry Birds) into QB64. The Library is written in C++ and I figured by using DECLARE LIBRARY I could get this done. However, my C++ knowledge is lacking as well. Trying to figure out where pointers are used versus variables, their types, and when I need _OFFSETs is just confusing this old thick headed brain of mine. Here is a link to the Box2D physics engine:
https://box2d.org/about/
I feel QB64 needs a 2D physics engine to help attract more users. I know my games would be vastly improved if I had access to something that could create Angry Birds style game physics.
Is anyone with more knowledge on either subject, importing Box2D into QB64, or tutoring and helping me build an engine, willing to help? Box2D also has a light version, Box2D_Lite, that may be a good start if porting the engine is an option.
Below is a link to the video series I was following. I got to the end of Lesson 9 and am totally confused on what the presenter did when introducing the FOR loop.
The video series:
https://www.youtube.com/watch?v=vcgtwY39FT0&list=PLtrSb4XxIVbpZpV65kk73OoUcIrBzoSiO&index=2
And the code I hacked together so far trying to follow along and convert OOP to QB64 on the fly.
Code: (Select All)
'https://www.youtube.com/watch?v=XG6yOtEpRSw&list=PLtrSb4XxIVbpZpV65kk73OoUcIrBzoSiO&index=2
OPTION _EXPLICIT
CONST MIN_VALUE = -2.802597E-45
TYPE Type_Vector2
x AS SINGLE
y AS SINGLE
END TYPE
TYPE Type_Ray2D
origin AS Type_Vector2
direction AS Type_Vector2
END TYPE
TYPE Type_RaycastResult
ppoint AS Type_Vector2
normal AS Type_Vector2
t AS SINGLE
hit AS INTEGER
END TYPE
TYPE Type_Line2D
from AS Type_Vector2
too AS Type_Vector2
colour AS _UNSIGNED LONG
lifetime AS INTEGER
END TYPE
TYPE Type_Rigidbody2D
position AS Type_Vector2
rotation AS SINGLE
END TYPE
TYPE Type_Box2D ' rotated bounding box
size AS Type_Vector2
halfSize AS Type_Vector2
rigidbody AS Type_Rigidbody2D
END TYPE
TYPE Type_AABB ' axis aligned bounding box (not rotated)
size AS Type_Vector2
halfSize AS Type_Vector2
rigidbody AS Type_Rigidbody2D
END TYPE
TYPE Type_Circle
Radius AS SINGLE
rigidbody AS Type_Rigidbody2D
END TYPE
DIM __AABB AS Type_AABB
DIM __Circle AS Type_Circle
DIM __rigidbody2D AS Type_Rigidbody2D
DIM Vertices(4) AS Type_Vector2
SUB Line2D.setFromToo (__line2D AS Type_Line2D, from AS Type_Vector2, too AS Type_Vector2)
__line2D.from = from
__line2D.too = too
END SUB
SUB Line2D (__line2D AS Type_Line2D, from AS Type_Vector2, too AS Type_Vector2, colour AS _UNSIGNED LONG, lifetime AS INTEGER)
__line2D.from = from
__line2D.too = too
__line2D.colour = colour
__line2D.lifetime = lifetime
END SUB
FUNCTION Line2D.beginFrame (__line2d AS Type_Line2D)
__line2d.lifetime = __line2d.lifetime - 1
Line2D.beginFrame = __line2d.lifetime
END FUNCTION
SUB Line2D.getFrom (__line2d AS Type_Line2D, from AS Type_Vector2)
from = __line2d.from
END SUB
SUB Line2D.getToo (__line2d AS Type_Line2D, too AS Type_Vector2)
too = __line2d.too
END SUB
SUB Line2D.getStart (__line2d AS Type_Line2D, start AS Type_Vector2)
start = __line2d.from
END SUB
SUB Line2D.getEnd (__line2d AS Type_Line2D, endd AS Type_Vector2)
endd = __line2d.too
END SUB
FUNCTION Line2D.getColour (__line2d AS Type_Line2D)
Line2D.getColour = __line2d.colour
END FUNCTION
FUNCTION Line2D.lengthSquared (__line2d AS Type_Line2D)
DIM from AS Type_Vector2
DIM too AS Type_Vector2
DIM length AS Type_Vector2
Line2D.getFrom __line2d, from
Line2D.getToo __line2d, too
length.x = too.x - from.x
length.y = too.y - from.y
Line2D.lengthSquared = lengthSquared(length)
END FUNCTION
'********************************
'* RIGIDBODY * <-----------------------------------------------------------------------
'********************************
'--------------------------------
'---- IntersectionDetector2D ----
'--------------------------------
FUNCTION PointOnLine (TestPoint AS Type_Vector2, __line2D AS Type_Line2D)
' based on the Slope Intercept Form of the equation of a straight line
'
'
' | S = Line Start = (0,1)
' | (8,5) E = Line End = (8,5)
' 5+ __ù Need to get values of this formula: y = m * x + b
' | _- Solve for m:
' | _-- - dy = Ey - Sy = 5 - 1 = 4
' 4+ __- - dx = Ex - Sx = 8 - 0 = 8
' | _- dy 4 1
' | P _-- - m = ---- = --- = --- (m solved)
' 3+ ù __- dx 8 2
' | (2,3) _- Solve for b:
' | _-- - b = y - mx (plug in x (0) and y (1) from line start)
' 2+ __- 1
' | _- - b = 1 - --- * 0 = 1 - 0 = 1 (b solved)
' |_-- 2
' 1ù Plug in values from Px along with solved m and b to compare y result with Py
'(0,1) 1
' | - y = --- * x + 1 = y = .5 * 2 + 1 = y = 2 (FALSE) 2 is not equal to Py (3)
' +---+---+---+---+---+---+---+---+--- 2
' 0 1 2 3 4 5 6 7 8
'
DIM lineStart AS Type_Vector2 ' start vector of line (x,y)
DIM lineEnd AS Type_Vector2 ' end vector of line (x,y)
DIM dx AS SINGLE ' run (delta in the x direction)
DIM dy AS SINGLE ' rise (delta in the y direction)
DIM m AS SINGLE ' slope (rise over run)
DIM b AS SINGLE ' the y intercept
PointOnLine = 0 ' assume point not on line
'----------
lineStart = __line2D.from
lineEnd = __line2D.too
'Line2D.getStart __line2D, lineStart ' get line start vector (x,y)
'Line2D.getEnd __line2D, lineEnd ' get line end vector (x,y)
'----------
dy = lineEnd.y - lineStart.y ' calculate rise
dx = lineEnd.x - lineStart.x ' calculate run
IF dx = 0 THEN ' vertical line? (avoid divide by 0)
IF TestPoint.x = lineStart.x THEN ' yes, do x values match?
PointOnLine = -1 ' yes, must be on the line
EXIT FUNCTION ' leave
END IF
END IF
m = dy / dx ' calculate slope
b = lineStart.y - (m * lineStart.x) ' calculate y intercept
IF TestPoint.y = m * TestPoint.x + b THEN PointOnLine = -1 ' point on line if y = mx + b
END FUNCTION
FUNCTION PointInCircle (TestPoint AS Type_Vector2, __circle AS Type_Circle)
' Check for point within circle.
'
'
' ********* - Simply use Pythagoras to solve.
' **** ****
' ***\ *** - Calculate x and y sides from point to center
' ** \ ** - if x side * x side + y side * y side <= radius * radius then point within circle
' * \ x2 * - (this method negates having to use square root)
' * Radius\ +--------------*---__ù p2
' * \ y2| _*-- p1: x1 = p1.x - center.x
' * \ | __-- * y1 = p1.y - center.y
' * \ | __-- * x1 * x1 + y1 * y1 <= radius * radius (TRUE - point within circle)
' * \| __-- L2 *
' * Center x,y ù-- * p2: x2 = p2.x - center.x
' * |\ * y2 = p2.y - center.y
' * | \ L1 * x2 * x2 + y2 * y2 <= radius * radius (FALSE - point NOT within circle)
' * y1| \ *
' * | \ *
' * +----ù *
' * x1 p1 *
' ** **
' *** ***
' **** ****
' *********
DIM circleCenter AS Type_Vector2 ' center coordinate of circle (x,y)
DIM centerToPoint AS Type_Vector2 ' x,y lengths
DIM radius AS SINGLE ' radius of circle
PointInCircle = 0 ' assume point not within circle
'----------
circleCenter = __circle.rigidbody.position
'Circle.getCenter __circle, circleCenter ' get center coordinate of circle (x,y)
'----------
'----------
radius = __circle.Radius
'radius = Circle.getRadius(__circle) ' get radius of circle
'----------
centerToPoint.x = TestPoint.x - circleCenter.x ' calculate x distance from point to center of circle
centerToPoint.y = TestPoint.y - circleCenter.y ' calculate y distance from point to center of circle
IF lengthSquared(centerToPoint) <= radius * radius THEN PointInCircle = -1 ' return true if length <= radus of circle
END FUNCTION
FUNCTION PointInAABB (TestPoint AS Type_Vector2, box AS Type_AABB)
' Check for a point inside standard rectangle (AABB axis aligned bounding box)
'
' . . Four simple checks needed to see if point is within a non rotated rectangle
' . .
' . . p1: p1.x <= max.x (TRUE) AND
' ..... +-------------------------------+ ..... min.x <= p1.x (TRUE) AND
' |min(x,y) | p1.y <= max.y (TRUE) AND
' | | min.y <= p1.y (TRUE) = All TRUE means within rectangle
' | |
' | p1 | p2: p2.x <= max.x (FALSE) AND
' | . | min.x <= p2.x (TRUE) AND
' | | p2 p2.y <= max.y (TRUE) AND
' | | . min.y <= p2.y (TRUE) = Any FALSE means NOT within rectangle
' | |
' | |
' | |
' | max(x,y)|
' ..... +-------------------------------+ .....
' . .
' . .
' . .
DIM min AS Type_Vector2 ' upper left rectangular coordinate (x,y)
DIM max AS Type_Vector2 ' lower right rectangular coordinate (x,y)
AABB.getMin box, min ' get upper left coordinate
AABB.getMax box, max ' get lower right coordinate
PointInAABB = 0 ' assume point not within AABB
IF TestPoint.x <= max.x THEN ' perform the four checks
IF min.x <= TestPoint.x THEN
IF TestPoint.y <= max.y THEN
IF min.y <= TestPoint.y THEN
PointInAABB = -1 ' if all true report point within
END IF
END IF
END IF
END IF
END FUNCTION
FUNCTION PointInBox2D (TestPoint AS Type_Vector2, box AS Type_Box2D)
'
' Test for a point in a rotated 2D box by rotating the point into the box's local space
'
' _-\
' _- \ P
' _- \ _ù rotated position
' _- \ _-
' _- \_-
' _- _-\ - Rotate point P around origin C the same degree as rotated box
' - Rotated _- \ - Point P is now in the local boxe's space
' \ AABB _- \ - From here it's just a simple AABB min/max check
' \ Cù \
' \ \ _-
' \ \ _-
' \ \ _- |
' \ _\ /
' \ _- \ /
' \ _- \ _-
' \- \ _-
' ù original position
' P
DIM pointLocalBoxSpace AS Type_Vector2
DIM min AS Type_Vector2
DIM max AS Type_Vector2
PointInBox2D = 0 ' assume point not within
Box2D.getMin box, min ' get upper left coordinate
Box2D.getMax box, max ' get lower right coordinate
pointLocalBoxSpace = TestPoint ' copy test point
'+------------------------------------------------+
'| Translate the point into the box's local space |
'+------------------------------------------------+
Rotate pointLocalBoxSpace, box.rigidbody.rotation, box.rigidbody.position ' rotate point into box's local space
'+-----------------------------------------+
'| Perform standard point within AABB test |
'+-----------------------------------------+
IF pointLocalBoxSpace.x <= max.x THEN ' perform the four AABB checks
IF min.x <= pointLocalBoxSpace.x THEN
IF pointLocalBoxSpace.y <= max.y THEN
IF min.y <= pointLocalBoxSpace.y THEN
PointInBox2D = -1 ' if all true report point within
END IF
END IF
END IF
END IF
END FUNCTION
FUNCTION lineAndCircle (__line2D AS Type_Line2D, __circle AS Type_Circle)
'
' Use projection to determine if a line is intersecting a circle
'
' ********* Determine if line B is intersecting circle:
' **** **** - Line B end points are outside circle (check points within circle)
' *** *** - If either end point within circle then line B is intersecting (return TRUE)
' ** ** - Get line segment x and y lengths and store in ab.x and ab.y
' * * - Get vector x and y lengths from center of circle to start of line segment
' * * - Store in centerToLineStart.x and centerToLineStart.y
' * * - Perform dot product of vectors to get a percentage of line segment
' * * - centerToLineStart ù ab
' * Center * t = ------------------------ t = 0 to 1 (percentage of A to B)
' * x,y * ab ù ab
' * __+ * - Add (ab * t) to Start to get point C (closest point to center)
' * __--- | * - check for point C within circle
' * ___--- | * - (this method negates the need to use square root)
' A ___-- | *
' __--- * | *
' ___--- * *
' Start ù------------------------------ù-----------------------------ù End
' ** C **
' B *** closest point ***
' **** ****
' | ********* |
' |---------------------------- ab --------------------------|
' | |
DIM LineStart AS Type_Vector2 ' start line vector (x,y)
DIM LineEnd AS Type_Vector2 ' end line vector (x,y)
DIM ab AS Type_Vector2 ' line segment (ex-sx,ey-sy)
DIM circleCenter AS Type_Vector2 ' center of circle (x,y)
DIM centerToLineStart AS Type_Vector2 ' start of line to center (cx-sx,cy-sy)
DIM t AS SINGLE ' percentage of the line (0 to 1)
DIM closestPoint AS Type_Vector2 ' closest point on line to center
lineAndCircle = 0 ' assume no intersection
'----------
LineStart = __line2D.from
LineEnd = __line2D.too
'Line2D.getStart __line2D, LineStart ' get start vector of line (x,y)
'Line2D.getEnd __line2D, LineEnd ' get end vector of line (x,y)
'----------
IF PointInCircle(LineStart, __circle) OR PointInCircle(LineEnd, __circle) THEN ' is either line end point within circle?
lineAndCircle = -1 ' yes, then line must be intersecting circle
EXIT FUNCTION ' leave
END IF
ab.x = LineEnd.x - LineStart.x ' calculate line segment length
ab.y = LineEnd.y - LineStart.y
'+--------------------------------------------------------+
'| Project point (circle position) onto ab (line segment) |
'| result = parameterized position t |
'+--------------------------------------------------------+
'----------
circleCenter = __circle.rigidbody.position
'Circle.getCenter __circle, circleCenter ' get center of circle
'----------
centerToLineStart.x = circleCenter.x - LineStart.x ' calculate length from center to start of line segment
centerToLineStart.y = circleCenter.y - LineStart.y
t = Dot(centerToLineStart, ab) / Dot(ab, ab) ' perform dot product on vectors to get percentage
IF t < 0 OR t > 1 THEN EXIT FUNCTION ' leave if not between the line segment, no intersection
'+--------------------------------------------+
'| Find the closest point to the line segment |
'+--------------------------------------------+
closestPoint.x = LineStart.x + ab.x * t ' calculate closest line point to center of circle
closestPoint.y = LineStart.y + ab.y * t
lineAndCircle = PointInCircle(closestPoint, __circle) ' return result of closest point within circle
END FUNCTION
FUNCTION lineAndAABB (__line2D AS Type_Line2D, box AS Type_AABB)
'Raycasting
DIM lineStart AS Type_Vector2
DIM lineEnd AS Type_Vector2
DIM unitVector AS Type_Vector2
DIM min AS Type_Vector2
DIM max AS Type_Vector2
DIM tmin AS SINGLE
DIM tmax AS SINGLE
DIM t AS SINGLE
lineAndAABB = 0
Line2D.getStart __line2D, lineStart
Line2D.getEnd __line2D, lineEnd
IF PointInAABB(lineStart, box) OR PointInAABB(lineEnd, box) THEN
lineAndAABB = -1
EXIT FUNCTION
END IF
unitVector.x = lineEnd.x - lineStart.x
unitVector.y = lineEnd.y - lineEnd.y
Normalize unitVector
IF unitVector.x <> 0 THEN unitVector.x = 1 / unitVector.x
IF unitVector.y <> 0 THEN unitVector.y = 1 / unitVector.y
AABB.getMin box, min
min.x = min.x - lineStart.x * unitVector.x
min.y = min.y - lineStart.y * unitVector.y
AABB.getMax box, max
max.x = max.x - lineStart.x * unitVector.x
max.y = max.y - lineStart.y * unitVector.y
tmin = MathMax(MathMin(min.x, max.x), MathMin(min.y, max.y))
tmax = MathMin(MathMax(min.x, max.x), MathMax(min.y, max.y))
IF tmax < 0 OR tmin > tmax THEN EXIT FUNCTION
IF tmin < 0 THEN t = tmax ELSE t = tmin
IF t > 0 AND t * t < Line2D.lengthSquared(__line2D) THEN lineAndAABB = -1
END FUNCTION
FUNCTION lineAndBox2D (__line2d AS Type_Line2D, box AS Type_Box2D)
'Rotate the line into the box's local space
DIM theta AS SINGLE
DIM center AS Type_Vector2
DIM localStart AS Type_Vector2
DIM localEnd AS Type_Vector2
DIM localLine AS Type_Line2D
DIM min AS Type_Vector2
DIM max AS Type_Vector2
DIM __aabb AS Type_AABB
theta = -box.rigidbody.rotation
center = box.rigidbody.position
Line2D.getStart __line2d, localStart
Line2D.getEnd __line2d, localEnd
Rotate localStart, theta, center
Rotate localEnd, theta, center
'Line2D localLine, localStart, localEnd, _RGB32(255, 255, 255), 1 (instead of 2 lines below)
localLine.from = localStart
localLine.too = localEnd
Box2D.getMin box, min
Box2D.getMax box, max
AABB __aabb, min, max
lineAndBox2D = lineAndAABB(localLine, __aabb)
END FUNCTION
' +----------+
' | Raycasts |
' +----------+
FUNCTION RaycastCircle (__circle AS Type_Circle, ray AS Type_Ray2D, result AS Type_RaycastResult)
DIM originToCircle AS Type_Vector2
DIM center AS Type_Vector2
DIM origin AS Type_Vector2
DIM radius AS SINGLE
DIM radiusSquared AS SINGLE
DIM originToCircleLengthSquared AS SINGLE
DIM direction AS Type_Vector2
DIM a AS SINGLE
DIM bSq AS SINGLE
DIM f AS SINGLE
DIM t AS SINGLE
DIM ppoint AS Type_Vector2
DIM normal AS Type_Vector2
RaycastCircle = 0
RaycastResult.reset result
Circle.getCenter __circle, center
Ray2D.getOrigin ray, origin
originToCircle.x = center.x - origin.x
originToCircle.y = center.y - origin.y
radius = Circle.getRadius(__circle)
radiusSquared = radius * radius
originToCircleLengthSquared = lengthSquared(originToCircle)
' Project the vector from the ray origin onto the direction of the ray
Ray2D.getDirection ray, direction
a = Dot(originToCircle, direction)
bSq = originToCircleLengthSquared - (a * a)
IF radiusSquared - bSq < 0 THEN EXIT FUNCTION
f = SQR(radiusSquared - bSq)
t = 0
IF originToCircleLengthSquared < radiusSquared THEN
t = a + f ' ray starts inside the circle
ELSE
t = a - f ' ray starts outside the circle
END IF
IF result.ppoint.x + result.ppoint.y = 0 THEN
ppoint.x = origin.x + direction.x * t
ppoint.y = origin.y + direction.y * t
normal.x = ppoint.x - center.x
normal.y = ppoint.y - center.y
Normalize normal
result.ppoint = ppoint
result.normal = normal
result.t = t
result.hit = -1
END IF
RaycastCircle = -1
END FUNCTION
FUNCTION RaycastAABB (box AS Type_AABB, __Ray2D AS Type_Ray2D, result AS Type_RaycastResult)
'DIM lineStart AS Type_Vector2
'DIM lineEnd AS Type_Vector2
DIM unitVector AS Type_Vector2
DIM min AS Type_Vector2
DIM max AS Type_Vector2
DIM tmin AS SINGLE
DIM tmax AS SINGLE
DIM t AS SINGLE
DIM hit AS INTEGER
DIM ppoint AS Type_Vector2
DIM normal AS Type_Vector2
RaycastAABB = 0
RaycastResult.reset result
unitVector.x = __Ray2D.direction.x ' lineEnd.x - lineStart.x
unitVector.y = __Ray2D.direction.y 'lineEnd.y - lineEnd.y
Normalize unitVector
IF unitVector.x <> 0 THEN unitVector.x = 1 / unitVector.x
IF unitVector.y <> 0 THEN unitVector.y = 1 / unitVector.y
AABB.getMin box, min
min.x = min.x - __Ray2D.origin.x 'lineStart.x * unitVector.x
min.y = min.y - __Ray2D.origin.y 'lineStart.y * unitVector.y
AABB.getMax box, max
max.x = max.x - __Ray2D.origin.x 'lineStart.x * unitVector.x
max.y = max.y - __Ray2D.origin.y 'lineStart.y * unitVector.y
tmin = MathMax(MathMin(min.x, max.x), MathMin(min.y, max.y))
tmax = MathMin(MathMax(min.x, max.x), MathMax(min.y, max.y))
IF tmax < 0 OR tmin > tmax THEN EXIT FUNCTION
IF tmin < 0 THEN t = tmax ELSE t = tmin
IF t > 0 THEN hit = -1
IF NOT hit THEN EXIT FUNCTION
IF result.ppoint.x = 0 AND result.ppoint.y = 0 THEN
ppoint.x = __Ray2D.origin.x + __Ray2D.direction.x * t
ppoint.y = __Ray2D.origin.y + __Ray2D.direction.y * t
normal.x = __Ray2D.origin.x - ppoint.x
normal.y = __Ray2D.origin.y - ppoint.y
Normalize normal
result.ppoint = ppoint
result.normal = normal
result.t = t
result.hit = -1
END IF
RaycastAABB = -1
END FUNCTION
FUNCTION RaycastBox2D (box AS Type_Box2D, __Ray2D AS Type_Ray2D, result AS Type_RaycastResult)
DIM xAxis AS Type_Vector2
DIM yAxis AS Type_Vector2
DIM zerozero AS Type_Vector2
DIM p AS Type_Vector2
DIM f AS Type_Vector2
DIM e AS Type_Vector2
DIM size AS Type_Vector2
RaycastBox2D = 0
RaycastResult.reset result
Box2D.halfSize box, size
xAxis.x = 1
xAxis.y = 0
yAxis.x = 0
yAxis.y = 1
Rotate xAxis, -box.rigidbody.rotation, zerozero
Rotate yAxis, -box.rigidbody.rotation, zerozero
p.x = box.rigidbody.position.x - __Ray2D.origin.x
p.y = box.rigidbody.position.y - __Ray2D.origin.y
' Project the direction of the ray onto each axis of the box
f.x = Dot(xAxis, __Ray2D.direction)
f.y = Dot(yAxis, __Ray2D.direction)
' Next, project p onto every axis of the box
e.x = Dot(xAxis, p)
e.y = Dot(yAxis, p)
RaycastBox2D = -1
END FUNCTION
'--------------------------------
'--------- RIGIDBODY2D ----------
'--------------------------------
SUB RigidBody2D.getPosition (__rigidbody2D AS Type_Rigidbody2D, position AS Type_Vector2)
position.x = __rigidbody2D.position.x
position.y = __rigidbody2D.position.y
END SUB
SUB RigidBody2D.setPosition (__rigidbody2D AS Type_Rigidbody2D, position AS Type_Vector2)
__rigidbody2D.position.x = position.x
__rigidbody2D.position.y = position.y
END SUB
FUNCTION RigidBody2D.getRotation (__rigidbody2D AS Type_Rigidbody2D)
RigidBody2D.getRotation = __rigidbody2D.rotation
END FUNCTION
SUB RigidBody2D.setRotation (__rigidbody2D AS Type_Rigidbody2D, rotation AS SINGLE)
__rigidbody2D.rotation = rotation
END SUB
'********************************
'* PHYSICS2D PRIMATIVES * <-----------------------------------------------------------------------
'********************************
'--------------------------------
'------------ AABB --------------
'--------------------------------
SUB AABB (__AABB AS Type_AABB, min AS Type_Vector2, max AS Type_Vector2)
__AABB.size.x = max.x - min.x ' set size of object
__AABB.size.y = max.y - min.y
__AABB.halfSize.x = __AABB.size.x * .5
__AABB.halfSize.y = __AABB.size.y * .5
END SUB
SUB AABB.halfSize (__AABB AS Type_AABB, halfSize AS Type_Vector2)
halfSize.x = __AABB.size.x * .5
halfSize.y = __AABB.size.y * .5
END SUB
SUB AABB.getMin (__AABB AS Type_AABB, min AS Type_Vector2)
DIM halfSize AS Type_Vector2
AABB.halfSize __AABB, halfSize
min.x = __AABB.rigidbody.position.x - halfSize.x
min.y = __AABB.rigidbody.position.y - halfSize.y
END SUB
SUB AABB.getMax (__AABB AS Type_AABB, max AS Type_Vector2)
DIM halfSize AS Type_Vector2
AABB.halfSize __AABB, halfSize
max.x = __AABB.rigidbody.position.x + halfSize.x
max.y = __AABB.rigidbody.position.y + halfSize.y
END SUB
'--------------------------------
'----------- Box2D --------------
'--------------------------------
SUB Box2D (__box2D AS Type_Box2D, min AS Type_Vector2, max AS Type_Vector2)
__box2D.size.x = max.x - min.x ' set size of object
__box2D.size.y = max.y - min.y
__box2D.halfSize.x = __box2D.size.x * .5
__box2D.halfSize.y = __box2D.size.y * .5
END SUB
SUB Box2D.halfSize (__box2D AS Type_Box2D, halfSize AS Type_Vector2)
halfSize.x = __box2D.size.x * .5
halfSize.y = __box2D.size.y * .5
END SUB
SUB Box2D.getMin (__box2D AS Type_Box2D, min AS Type_Vector2)
DIM halfSize AS Type_Vector2
Box2D.halfSize __box2D, halfSize
min.x = __box2D.rigidbody.position.x - halfSize.x
min.y = __box2D.rigidbody.position.y - halfSize.y
END SUB
SUB Box2D.getMax (__box2D AS Type_Box2D, max AS Type_Vector2)
DIM halfSize AS Type_Vector2
Box2D.halfSize __box2D, halfSize
max.x = __box2D.rigidbody.position.x + halfSize.x
max.y = __box2D.rigidbody.position.y + halfSize.y
END SUB
SUB Box2D.getVertices (__box2d AS Type_Box2D, Vertices() AS Type_Vector2)
DIM min AS Type_Vector2
DIM max AS Type_Vector2
DIM vert AS Type_Vector2
DIM vCount AS INTEGER
Box2D.getMin __box2d, min
Box2D.getMax __box2d, max
Vertices(1).x = min.x
Vertices(1).y = min.y
Vertices(2).x = min.x
Vertices(2).y = max.y
Vertices(3).x = max.x
Vertices(3).y = min.y
Vertices(4).x = max.x
Vertices(4).y = max.y
IF __box2d.rigidbody.rotation <> 0 THEN
vCount = 0
DO
vert = Vertices(vCount)
Rotate vert, __box2d.rigidbody.rotation, __box2d.rigidbody.position
LOOP UNTIL vCount = 4
END IF
END SUB
'--------------------------------
'----------- Circle -------------
'--------------------------------
FUNCTION Circle.getRadius (__circle AS Type_Circle)
Circle.getRadius = __circle.Radius
END FUNCTION
SUB Circle.setRadius (__circle AS Type_Circle, radius AS SINGLE)
__circle.Radius = radius
END SUB
SUB Circle.getCenter (__circle AS Type_Circle, center AS Type_Vector2)
center = __circle.rigidbody.position
END SUB
'--------------------------------
'--------- Collider2D -----------
'--------------------------------
'--------------------------------
'------------ Ray2D -------------
'--------------------------------
SUB Ray2D (__Ray2D AS Type_Ray2D, origin AS Type_Vector2, direction AS Type_Vector2)
__Ray2D.origin = origin
__Ray2D.direction = direction
Normalize __Ray2D.direction
END SUB
SUB Ray2D.getOrigin (__ray2D AS Type_Ray2D, origin AS Type_Vector2)
origin = __ray2D.origin
END SUB
SUB Ray2D.getDirection (__ray2D AS Type_Ray2D, direction AS Type_Vector2)
direction = __ray2D.direction
END SUB
'--------------------------------
'--------- RaycastResult --------
'--------------------------------
SUB RaycastResult (__RaycastResult AS Type_RaycastResult)
__RaycastResult.ppoint.x = 0
__RaycastResult.ppoint.y = 0
__RaycastResult.normal.x = 0
__RaycastResult.normal.y = 0
__RaycastResult.t = -1
__RaycastResult.hit = 0
END SUB
SUB RaycastResult.init (__RaycastResult AS Type_RaycastResult, ppoint AS Type_Vector2, normal AS Type_Vector2, t AS SINGLE, hit AS INTEGER)
__RaycastResult.ppoint = ppoint
__RaycastResult.normal = normal
__RaycastResult.t = t
__RaycastResult.hit = hit
END SUB
SUB RaycastResult.reset (result AS Type_RaycastResult)
IF result.ppoint.x OR result.ppoint.y THEN
result.ppoint.x = 0
result.ppoint.y = 0
result.normal.x = 0
result.normal.y = 0
result.t = -1
result.hit = 0
END IF
END SUB
SUB AddVectors (V1 AS Type_Vector2, V2 AS Type_Vector2, Vout AS Type_Vector2)
' - -
' Formula: V1 + V2 = (V1.x, v1.y) + (V2.x, V2.y) = (V1.x + V2.x, V1.y + V2.y)
' V1 - input : Vector 1
' V2 - input : Vector 2
' Vout - output: the new vector
Vout.x = V1.x + V2.x ' x value of vector 2 gets added to x value of vector 1
Vout.y = V1.y + V2.y ' y value of vector 2 gets added to y value of vector 1
END SUB
SUB SubtractVectors (V1 AS Type_Vector2, V2 AS Type_Vector2, Vout AS Type_Vector2)
' - -
' Formula: V1 + V2 = (V1.x, v1.y) - (V2.x, V2.y) = (V1.x - V2.x, V1.y - V2.y)
' V1 - input : Vector 1
' V2 - input : Vector 2
' Vout - output: the new vector
Vout.x = V1.x - V2.x ' x value of vector 2 gets subtracted from x value of vector 1
Vout.y = V1.y - V2.y ' y value of vector 2 gets subtracted from y value of vector 1
END SUB
SUB ScalarMultiplyVector (V AS Type_Vector2, Scalar AS SINGLE, Vout AS Type_Vector2)
' "Scaling the vector"
'
' Formula: V * Scalar = (Vx, Vy) * Scalar = (Vx * Scalar, Vy * Scalar)
' V - input : Vector
' Scalar - input : scalar multiplication value
' Vout - output: the new vector
Vout.x = V.x * Scalar ' x value of vector gets multiplied by scalar
Vout.y = V.y * Scalar ' y value of vector gets multiplied by scalar
END SUB
FUNCTION Dot (V1 AS Type_Vector2, V2 AS Type_Vector2)
' Dot product of vectors
' - -
' Formula: V1 ù V2 = (V1.x, V1.y) ù (V2.x, V2.y) = (V1.x * v2.x) + (V1.y * V2.y)
Dot = V1.x * V2.x + V1.y * V2.y ' multiply vector x values then add multiplied vector y values
END FUNCTION
FUNCTION CrossProductVectors (V1 AS Type_Vector2, V2 AS Type_Vector2)
' Also known as a "Wedge Product" or "Perp Product" for 2D vectors
' - - ³ x y ³
' Formula: V1 * V2 = (V1.x, V1.y) * (V2.x, V2.y) = ³V1.x V1.y³ = (V1.x * V2.y) - (V1.y * V2.x)
' ³V2.x V2.y³
CrossProductVectors = V1.x * V2.y - V1.y * V2.x
END FUNCTION
FUNCTION VectorLength (V AS Type_Vector2)
' _______________________
' Formula: º V º = û V.x * V.x + V.y * V.y
VectorLength = _HYPOT(V.x, V.y)
END FUNCTION
SUB Normalize (v AS Type_Vector2)
' Also known as a unit vector
' _______________________
' Formula: V / º V º = (V1.x, V1.y) / û V.x * V.x + V.y * v.y
DIM VecLength AS SINGLE
VecLength = _HYPOT(v.x, v.y) ' length of vector
v.x = v.x / VecLength ' normalized x length
v.y = v.y / VecLength ' normalized y length
END SUB
SUB Rotate (vec AS Type_Vector2, angleDeg AS SINGLE, origin AS Type_Vector2)
' Rotate a point around an origin using linear transformations.
'
' Rotating from (x,y) to (x',y') |
' | (x',y') : L = R cosé | All of this shows how to get to this
' | ù : A = x' | -----------
' | /.\ : B = L cosè = R cosè cosé = x cosé | |
' | / .è\ : (note - * opposite angles are equal) | |
' | / . \ : C = R siné | +----+
' | / . \ : D = C sinè = R sinè siné = y siné | |
' | / . \ : Y = R sinè | |
' | / . \C : X = r cosè |
' | / . \ __ | -----------------
' | / . \ L stops : x' = B - |AB| = X cosé - Y siné |
' | / . \ here
' | / . \ | All of this just to show how to get from x to x' using (X cosé - Y siné)
' | / . \ | Use the same linear transformation methods to get y' using (X siné + Y cosé)
' | R/ . \ |
' | / .¿ D \ Change the origin point of all rotations to (0,0) by subtracting the current
' | / .------------âù_--ù (x,y) origin point from the current vector length. Add it back when rotation is
' | / . __--. . completed.
' | / . * __-- . .
' | / . __-- . .
' | / L __-- . .
' | / __-- . . .Y
' | / __-- * . . .
' | / __-- . . .
' | / é __-- . . .
' |/__-- è â. . .
' ù-----------------------ù-------------ù---ù------------
' A B
' |------------------- X -------------------|
DIM x AS SINGLE
DIM y AS SINGLE
DIM __cos AS SINGLE
DIM __sin AS SINGLE
DIM xPrime AS SINGLE
DIM yPrime AS SINGLE
x = vec.x - origin.x ' move rotation vector origin to 0
y = vec.y - origin.y
__cos = COS(_D2R(angleDeg)) ' get cosine and sine of angle
__sin = SIN(_D2R(angleDeg))
xPrime = (x * __cos) - (y * __sin) ' calculate rotated location of vector
yPrime = (x * __sin) + (y * __cos)
xPrime = xPrime + origin.x ' move back to original origin
yPrime = yPrime + origin.y
vec.x = xPrime ' pass back rotated vector
vec.y = yPrime
END SUB
FUNCTION compareXYEpsilon (x AS SINGLE, y AS SINGLE, epsilon AS SINGLE)
compareXYEpsilon = 0
IF ABS(x - y) <= epsilon * MathMax(1, MathMax(ABS(x), ABS(y))) THEN compareXYEpsilon = -1
END FUNCTION
FUNCTION compareVecEpsilon (vec1 AS Type_Vector2, vec2 AS Type_Vector2, epsilon AS SINGLE)
compareVecEpsilon = 0
IF compareXYEpsilon(vec1.x, vec2.x, epsilon) AND compareXYEpsilon(vec1.y, vec2.y, epsilon) THEN compareVecEpsilon = -1
END FUNCTION
FUNCTION compareXY (x AS SINGLE, y AS SINGLE)
compareXY = 0
IF ABS(x - y) <= MIN_VALUE * MathMax(1, MathMax(ABS(x), ABS(y))) THEN compareXY = -1
END FUNCTION
FUNCTION compareVec (vec1 AS Type_Vector2, vec2 AS Type_Vector2)
compareVec = 0
IF compareXY(vec1.x, vec2.x) AND compareXY(vec1.y, vec2.y) THEN compareVec = -1
END FUNCTION
FUNCTION lengthSquared (length AS Type_Vector2)
lengthSquared = length.x * length.x + length.y + length.y
END FUNCTION
FUNCTION MathMax (num1 AS SINGLE, num2 AS SINGLE)
IF num1 >= num2 THEN MathMax = num1 ELSE MathMax = num2
END FUNCTION
FUNCTION MathMin (num1 AS SINGLE, num2 AS SINGLE)
IF num1 <= num2 THEN MathMin = num1 ELSE MathMin = num2
END FUNCTION
RE: 2D Physics Engine Help - a740g - 06-12-2023
I think mechatronic3000 ported Randy Gaul's Cute Physics 2D to QB64 a long time ago: 2D Physics Engine (alephc.xyz)
The updated version of that is here: QBPool/QBPOOL/fzxNGN/physics at main · mechatronic3000/QBPool · GitHub
Box2D will require a considerable effort to port. A QB64-PE wrapper may be possible. I'll need to check.
Update: Randy's latest C++ library is here: RandyGaul/ImpulseEngine: Simple, open source, 2D impulse based physics engine for educational use. (github.com). It's called Impulse Engine and not Cute Physics like I noted above.
RE: 2D Physics Engine Help - madscijr - 06-12-2023
(06-12-2023, 03:16 PM)TerryRitchie Wrote: I figured the process of converting the Java OOP code to functions and procedures would be fairly straight forward ... not so much.Yikes! I would recommend sticking with procedural code, if you can find it. Regular C is procedural, and would surely be easier to translate to procedural QB64PE.
(06-12-2023, 03:16 PM)TerryRitchie Wrote: My next thought was to incorporate the Box2D physics engine (the engine that Rovio used to create Angry Birds) into QB64. The Library is written in C++ and I figured by using DECLARE LIBRARY I could get this done.Spriggsy is the guy you'll want to talk to about using external libraries or APIs within QB64/PE. However, I think dependencies on libraries outside of the native OS would complicate things...
https://box2d.org/about/
Good Luck!
RE: 2D Physics Engine Help - Ultraman - 06-12-2023
It all depends on how easy it is to incorporate and how easy it is to map out. I don't know anything about physics engines so I wouldn't even know if I'd be doing it right even if I managed to get it to compile.
RE: 2D Physics Engine Help - a740g - 06-12-2023
(06-12-2023, 04:11 PM)madscijr Wrote:Yeah. Box2D is C++. So, it might be problematic to make a binding for QB64-PE.(06-12-2023, 03:16 PM)TerryRitchie Wrote: I figured the process of converting the Java OOP code to functions and procedures would be fairly straight forward ... not so much.Yikes! I would recommend sticking with procedural code, if you can find it. Regular C is procedural, and would surely be easier to translate to procedural QB64PE.
There is Chipmunk2D that is a plain ANSI C 2D physics library.
Then there is also the tiny single header Physac that I am planning to write a QB64-PE binding for as part of my larger raylib binding library.
RE: 2D Physics Engine Help - justsomeguy - 06-12-2023
I'm mechatronic3000.
I have been working on my physics engine (its 100% written in QB64) to make it more generalized and refactoring large portions of it to be easier to use. Its far from being feature complete.
- It currently has only two primitives, rectangles and circles. Concave objects are not supported yet.
- It has joints and you chain them together.
I've attached a simple example program.
Forgive my poor example, It was the only one I had to test the latest version of the refactor.
I'm using OpenGL in this example, but it can easily use native QB64 commands if you wish.
If you guys want I can post more about it.
![[Image: screenshot.jpg]](https://i.ibb.co/fH4G9bG/screenshot.jpg)
RE: 2D Physics Engine Help - justsomeguy - 06-12-2023
(06-12-2023, 09:34 PM)justsomeguy Wrote: I'm mechatronic3000.Sorry, I have yet to update my GITHUB. I will as soon as I get it in better shape.
I have been working on my physics engine (its 100% written in QB64) to make it more generalized and refactoring large portions of it to be easier to use. Its far from being feature complete.
- It currently has only two primitives, rectangles and circles. Concave objects are not supported yet.
- It has joints and you chain them together.
I've attached a simple example program.
Forgive my poor example, It was the only one I had to test the latest version of the refactor.
I'm using OpenGL in this example, but it can easily use native QB64 commands if you wish.
If you guys want I can post more about it.
RE: 2D Physics Engine Help - Kernelpanic - 06-12-2023
Quote:@Terry - The Library is written in C++ and I figured by using DECLARE LIBRARY I could get this done. However, my C++ knowledge is lacking as well. Trying to figure out where pointers are used versus variables, . . .@Terry - just in case you want to try something like this, pointers are created like this in C/C++:
int *zgr, **zzgr; (**zzgr is a pointer to a pointer. Yes, you can do that in C, ad nauseam.)
Here is an old example program that flips a number and a character array. - With zzgr = &zgr the pointer-to-pointer is assigned the address of zgr. The pointer then moves backwards through the field (zgr--), and **zzgr displays it. Hope that's halfway understandable.
Code: (Select All)
#include <stdio.h>
#include <stdlib.h>
#define MAX 9
int main(void)
{
int feld[] = { 1,2,3,4,5,6,7,8,9 };
char char_feld[] = { 'a','b','c','d','e','f','g','h','z' };
int i, *zgr, **zzgr;
char *char_zgr, **cchar_zgr;
for ( i = 0; i < MAX; i++ )
{
zgr = &feld[i];
printf("%d ", *zgr);
}
printf("\t");
for ( i = 0; i < MAX; i++ )
{
char_zgr = &char_feld[i];
printf("%c ", *char_zgr);
}
printf("\n\n");
//Mit zzgr = &zgr wird dem Zeiger-auf-Zeiger die Adresse
//von zgr zugewiesen. Der Zeiger zgr geht dann rückwärtz
//das Feld durch (zgr--), und **zzgr läßt es anzeigen.
for ( i = 0; i < MAX; i++ )
{
zzgr = &zgr;
printf("%d ",**zzgr);
zgr--;
}
printf("\t");
for ( i = 0; i < MAX; i++ )
{
cchar_zgr = &char_zgr;
printf("%c ", **cchar_zgr);
char_zgr--;
}
return(0);
}If you want to compile the program yourself, it says in the screenshot.
![[Image: zgr-auf-zgr-2023-06-13.jpg]](https://i.ibb.co/26FNv62/zgr-auf-zgr-2023-06-13.jpg)
RE: 2D Physics Engine Help - Sprezzo - 06-13-2023
i vouch for justsomeguy. he's definitely your man terry
RE: 2D Physics Engine Help - justsomeguy - 06-15-2023
I was looking over your code and WOW you put a lot of effort in your comments with ASCII art and everything. Good job! I wish I had the patience to do that. I will try to improve my source with more commenting.
If you still want to more or less roll your own physics engine, another resource I rolled across is https://github.com/matthias-research/pages/tree/master/tenMinutePhysics
His code is written in fair portable javascript. He has a Youtube channel as well, and can help explain the source.
Let me know if I can help you.