colgreen / box2dx Goto Github PK

What steps will reproduce the problem?
1. Add two shapes to a body
2. Delete the first one using DestroyShape
3. Delete the second one using DestroyShape

What is the expected output? What do you see instead?
Expected: Both shapes to be deleted.
Saw: First delete clobbers the shape list.

What version of the product are you using? On what operating system?
SVN Revision 56 (6/11/2008).  Development Platform: Windows XP.  Target 
Platforms: XNA on PC and XBox.

Please provide any additional information below.
Looks like this has been ported correctly but the original code has a bug.  
This has been fixed in the latest flash version.  Not sure if it is fixed 
in the C++ version.

Here is the code patched with the Flash Fix:

            Shape ppS = null; // Fix pointer-pointer stuff
            bool found = false;

            while (node != null)
            {
                if (node == shape)
                {
                    if (ppS != null)
                    {
                        ppS._next = shape._next;
                    }
                    else
                    {
                        _shapeList = shape._next;
                    }
                    //node = s.m_next;
                    found = true;
                    break;
                }

                ppS = node;
                node = node._next;
            }

What steps will reproduce the problem?
1. Calling b2Body::GetContactList()

What is the expected output? What do you see instead?
It's called "GetConactList", should read, "GetContactList"


What version of the product are you using? 
Latest from SVN as of Nov27 09.

What steps will reproduce the problem?
1. Use the Vec2 != Operator on a pair of Vec2 where x is the same but y is 
different.

What is the expected output? What do you see instead?
The != operator should return true but instead returns false.

What version of the product are you using? On what operating system?
SVN Revision 56 (6/11/2008).  Development Platform: Windows XP.  Target 
Platforms: XNA on PC and XBox.

Please provide any additional information below.

Current Implementation:
public static bool operator !=(Vec2 a, Vec2 b)
{
   return a.X != b.X && a.Y != b.Y;
}

Correct Implementation:
public static bool operator !=(Vec2 a, Vec2 b)
{
   return a.X != b.X || a.Y != b.Y;
}

What steps will reproduce the problem?
1. Run any non-trivial TestBed scenario (E.g. Pyramid) under the CLR 
Profiler (See http://msdn.microsoft.com/en-us/library/ms979205.aspx for 
details on running the CLR Profiler).
2. Examine the Timeline for an interval during which the test is running 
normally.

What is the expected output? What do you see instead?
I would expect to see a steep curve that gradually levels out indicating 
that Box2DX is reusing objects through pooling.

I see a sawtooth pattern indicating an allocate/collect cycle is occuring.  
Examining the types of objects allocated shows that Box2DX objects, such as 
Vec2 [], ManifoldPoint, Position [] etc., are being allocated and released 
with each step.  In a simple scenario in my game Box2DX is allocating 
around 200k per second which will trigger a garbage collection and 
consequently a small drop in framerate every 5 seconds or so.

What version of the product are you using? On what operating system?
SVN Revision 56 (6/11/2008).  Development Platform: Windows XP.  Target 
Platforms: XNA on PC and XBox.

Please provide any additional information below.
The problem is that when the garbage collector runs for XNA games on the 
xbox there is a perceivable jerk in the framerate.  In order for Box2DX to 
work well on this platform there needs to be at least a compile option to 
use pooled memory rather than garbage collected memory.

Great work on box2dx so far.  I understand if this issue isn't important 
for you and you don't want to complicate the code.  I'll probably be making 
the changes for my own project anyway.  I'd be willing to help with this or 
other parts of box2dx if you need.

- Dale

Allow us to define a different friction value for x and y. For instance, a
piece of paper falling drops slowly, but still moves quickly parallel to
the ground.

Custom types Vector2D and Color collide with XNA.

Enclosed is an update that uses Vector2D and Color from XNA, if XNA is defined 
in the project.

What steps will reproduce the problem?
1. Compile library
2. Compile Hello World Example
3. Run with mono --debug (to enable assertions)
4. Fails on Collision.TimeOfImpact:53

Here is a stack trace:

at Box2DX.Collision.Collision.TimeOfImpact
(Box2DX.Collision.Shape,Box2DX.Common.Sweep,Box2DX.Collision.Shape,Box2DX.Common
.Sweep)
[0x0000e] in
/home/touchtech/cth/box2dx-read-only/src/Collision/Collision.TimeOfImpact.cs:53
  at Box2DX.Dynamics.World.SolveTOI (Box2DX.Dynamics.TimeStep) [0x001de] in
/home/touchtech/cth/box2dx-read-only/src/Dynamics/World.cs:946
  at Box2DX.Dynamics.World.Step (single,int,int) [0x000ac] in
/home/touchtech/cth/box2dx-read-only/src/Dynamics/World.cs:592
  at HelloWorld.Program.Main (string[]) [0x00116] in
/home/touchtech/cth/box2dx-read-only/Examples/HelloWorld/Program.cs:105
  at (wrapper runtime-invoke)
HelloWorld.Program.runtime_invoke_void_string[]
(object,intptr,intptr,intptr) <0xffffffff>

Using latest svn (r56)

---

When not using debug mode it works in the hello world example. In our
program however it only works for a while until it after some random short
time stops working and crashes in World.cs around line 900+ in this loop:

for (Contact c = _contactList; c != null; c = c._next)


It crashes with a null pointer exception on c in different places in this
loop, which is kind of odd.

Our program works fine on both 32 bit linux and windows.

I don't know where to put that, but maybe it will save some time somebody.

For now last working revision is 59. Everything Above is WORK in progres.
http://code.google.com/p/box2dx/source/list

1. make a sensor with zero mass.
2. make another object with any mass.
3. notice no collision.

Collision when sensor touching another object. When mass of sensor is > 0
there is feedback.

What version of the product are you using? On what operating system?

C++ Box2D 2.0.1, Ubuntu Linux 9.10

Thanks.

Many of the constructors do something like this:

_mass = 0;
_rotation = 0;
_position = new Vec(0,0);
etc.

this is just waiting cycles because value types are already init'ed to the 
default value (0 for numeric types, Vec(0,0) for Vec's).

you can delete all of this code.

[deleted issue]

What steps will reproduce the problem?
-Call Math.Random() or Math.Random(float lo, float hi) in close succession.

What is the expected output? What do you see instead?
-Expect a random result.
-Same result if called in close succession.

What version of the product are you using? On what operating system?
-http://box2dx.googlecode.com/svn/trunk R36
-Vista X64

Please provide any additional information below.
-Making "Random rnd = new Random();" a static member of Math class would
resolve the issue.

1. Checkout solution from trunk
2. Open in Visual Studio 9
3. Set TestBed as startup project
4. Start Debugging (F5)

System.DllNotFoundException was unhandled
  Message="Unable to load DLL 'freetype6.dll': The specified module could
not be found. (Exception from HRESULT: 0x8007007E)"
  Source="Tao.FreeType"
  TypeName=""
  StackTrace:
       at Tao.FreeType.FT.FT_Init_FreeType(IntPtr& alibrary)
       at ISE.FTFont.ftInit() in
D:\Archives\Projects\box2dx\Examples\TestBed\OpenGL Render\FTclass.cs:line 168
       at ISE.FTFont..ctor(String resourcefilename, Int32& Success) in
D:\Archives\Projects\box2dx\Examples\TestBed\OpenGL Render\FTclass.cs:line 195
       at TestBed.OpenGLDebugDraw.InitTextRenderer(SimpleOpenGlControl
openGlCtrl) in D:\Archives\Projects\box2dx\Examples\TestBed\Render.cs:line 184
       at TestBed.MainForm..ctor() in
D:\Archives\Projects\box2dx\Examples\TestBed\MainForm.cs:line 64
       at TestBed.Program.Main() in
D:\Archives\Projects\box2dx\Examples\TestBed\Program.cs:line 40
  InnerException: 

Hello ihar3r, i'm posting this here because I didn't find your email
anywhere. I was wondering, is there any chance that you keep this proyect
up to date with the original box2d, which currently is in v2.1 and has some
neat features. I'm very interested in the EdgeShape stuff, I might be able
to help you adding these new features. Anyway, if possible contact me
(juancampa at gmail)

TIA,
Juan G.

mass

What steps will reproduce the problem?
1. Download and expand the .zip on windows 7
2. Run the testbed.exe

What is the expected output? What do you see instead?
I expected something, I saw nothing.

What version of the product are you using? On what operating system?
Version 2.0.1.3-r175 on Win 7

Please provide any additional information below.
Compatibility mode and administrative privileges doesn't help. 

What steps will reproduce the problem?
1. Install the Silverlight SDK and Create a VS project using the 
"Silverlight Class Library" template.
2. Add the box2dx files to the project.
3. Compile, it won't work.

It would be great to keep the fast unsafe methods for desktop apps, but 
provide safe versions for Silverlight.

arrays in .NET have a Length member that can be used instead.

What steps will reproduce the problem?
1. Check out the latest revision (r56) from the repository
2. Open the project in Visual Studio
3. Try to compile in "Release" mode



What is the expected output? What do you see instead?

The library does not compile because of the error "Unsafe code may only
appear if compiling with /unsafe"

What steps will reproduce the problem?
1. Create a body with multiple shapes
2. Destroy a shape in the middle of the internal shape list.

What is the expected output? What do you see instead?
I expect the remaining shapes on the body to still exist within the shape 
list.

I have fixed this bug by storing the previous shape.

What steps will reproduce the problem?
1. Register a ContactListener
2. Create two dynamic bodies and one static body
3. Cause the two dynamic bodies to collide
4. Cause one of the dynamic bodies to collide with the static body
5. Note that Add() events are only generated between dynamic and static
bodies, but Remove() events are generated for all types of collisions.

What is the expected output? What do you see instead?
I would expect Add() events to be raised for all collision types.

What version of the product are you using? On what operating system?
2.0.0alpha on Vista

What steps will reproduce the problem?

1. I'm not entirely sure; it occurs in r59 (the last "stable" revision of 
the code), but not in r72. It happens spontaneously, without visible 
provocation.


What is the expected output? What do you see instead?

Island.Add tries to add a Body to _bodies beyond the bounds of the array. 
More bizarre than that, the Assertion at the beginning of the method 
doesn't trigger, even though the condition is Clearly false (3 is not less 
than 2...).


What version of the product are you using? On what operating system?

I'm using r59 on Windows XP SP2.


Please provide any additional information below.

None that I know of.