<?xml version="1.0"?><?xml-stylesheet type="text/xsl" href="http://www.codeplex.com/rss.xsl"?><rss version="2.0"><channel><title>xnadevru Work Item Rss Feed</title><link>http://www.codeplex.com/WorkItem/List.aspx?ProjectName=xnadevru</link><description>xnadevru Work Item Rss Description</description><item><title>Created Issue: Wayyy behind in Framework versions...</title><link>http://xnadevru.codeplex.com/WorkItem/View.aspx?WorkItemId=24467</link><description>Your project files are all for the .NET Framework 1.0. No one that has a modern computer can use that have .NET 2.0, 3.0, or now even 3.5... Please update it so people can use it.&lt;br /&gt;</description><author>amerninja13</author><pubDate>Thu, 27 Aug 2009 21:34:52 GMT</pubDate><guid isPermaLink="false">Created Issue: Wayyy behind in Framework versions... 20090827093452P</guid></item><item><title>Commented Issue: Ray Casting with Triangle Mesh Shape</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=20964</link><description>It looks like CollisionWorld.CS lines 106&amp;#58;&lt;br /&gt;&lt;br /&gt;The if statement is lined up incorrectly, so you never enter the triangle mesh section.&lt;br /&gt;&lt;br /&gt;Then in TriangleMeshShape.cs&lt;br /&gt;You pass aabbMax twice, even though it should be&lt;br /&gt;aabbMin, aabbMax&lt;br /&gt;Comments: ** Comment from web user: nightst4r ** &lt;p&gt;Thats line 109 in TriangleMeshShape.cs&lt;/p&gt;</description><author>nightst4r</author><pubDate>Tue, 03 Feb 2009 01:45:04 GMT</pubDate><guid isPermaLink="false">Commented Issue: Ray Casting with Triangle Mesh Shape 20090203014504A</guid></item><item><title>Created Issue: Ray Casting with Triangle Mesh Shape</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=20964</link><description>It looks like CollisionWorld.CS lines 106&amp;#58;&lt;br /&gt;&lt;br /&gt;The if statement is lined up incorrectly, so you never enter the triangle mesh section.&lt;br /&gt;&lt;br /&gt;Then in TriangleMeshShape.cs&lt;br /&gt;You pass aabbMax twice, even though it should be&lt;br /&gt;aabbMin, aabbMax&lt;br /&gt;</description><author>nightst4r</author><pubDate>Tue, 03 Feb 2009 01:44:23 GMT</pubDate><guid isPermaLink="false">Created Issue: Ray Casting with Triangle Mesh Shape 20090203014423A</guid></item><item><title>Commented Issue: A minor bug in BulletX Source (2.5 Alpha)</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11068</link><description>In BulletX Source &amp;#40;2.5 Alpha&amp;#41; in TriangleMeshShape.cs at line 109&amp;#58;&lt;br /&gt;LocalProcessAllTriangles&amp;#40;callback, aabbMax, aabbMax&amp;#41;&amp;#59;&lt;br /&gt;should be&amp;#58;&lt;br /&gt;LocalProcessAllTriangles&amp;#40;callback, aabbMin, aabbMax&amp;#41;&amp;#59;&lt;br /&gt;&lt;br /&gt;It didn&amp;#39;t get the tringle mesh shape collision work, but I think it can help.&lt;br /&gt;Keep it up&amp;#33;&lt;br /&gt;Comments: ** Comment from web user: Corngood ** &lt;p&gt;It was missing some contacts between a box and a tri-mesh, so I made another small change.&lt;/p&gt;&lt;p&gt;On ConvexTriangleCallback.cs&amp;#40;71&amp;#41; I reversed the order of the multiply, so the line looks like&amp;#58;&lt;/p&gt;&lt;p&gt;Matrix convexInTriangleSpace &amp;#61; _convexBody.WorldTransform &amp;#42; MathHelper.InvertMatrix&amp;#40;_triBody.WorldTransform&amp;#41;&amp;#59;&lt;/p&gt;&lt;p&gt;I didn&amp;#39;t look into the problem too deeply, but I think there are some general issues with the matrix representation used in bullet, and how it differs from Microsoft.DirectX.Matrix.&lt;/p&gt;</description><author>Corngood</author><pubDate>Mon, 22 Dec 2008 02:16:55 GMT</pubDate><guid isPermaLink="false">Commented Issue: A minor bug in BulletX Source (2.5 Alpha) 20081222021655A</guid></item><item><title>CREATED ISSUE: Matrix.Inverse</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=16065</link><description>Is there a reason why BulletX uses MathHelper.InverseMatrix instead of the built-in Matrix.Inverse method&amp;#63;&lt;br /&gt;</description><author>DrDeth</author><pubDate>Fri, 11 Apr 2008 05:05:06 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: Matrix.Inverse 20080411050506A</guid></item><item><title>CREATED ISSUE: XCB feature request: maintain contentproj references in xcbproj files</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=15970</link><description>XCB should be capable of load not only a assest &amp;#40;FX, FBX, X, etc...&amp;#41; but it should also parse and load a Content.contentproj file.  The xcbproj file should maintain a refernce to this contentproj file rather than a refernce to the assests themselves if the assets were loaded in this manner.&lt;br /&gt;</description><author>missileControl</author><pubDate>Mon, 07 Apr 2008 15:04:47 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: XCB feature request: maintain contentproj references in xcbproj files 20080407030447P</guid></item><item><title>CREATED ISSUE: How can I rotate a BoxSphape so that a SphereShape can roll down from it?</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=14183</link><description>Hi,&lt;br /&gt;&lt;br /&gt;Here&amp;#39;s how I update RigidBody&amp;#39;s world transform matrix&amp;#58;&lt;br /&gt;&lt;br /&gt;Matrix rotationTransform &amp;#61; Matrix.CreateFromQuaternion&amp;#40;e.Orientation&amp;#41;&amp;#59;&lt;br /&gt;Matrix translateTransform &amp;#61; Matrix.CreateTranslation&amp;#40;e.Position&amp;#41;&amp;#59;&lt;br /&gt;&lt;br /&gt;&amp;#47;&amp;#47; for some reason BulletX doesnt take into consideration scale transform&lt;br /&gt;&amp;#47;&amp;#47; empirically found that setting CollisionShape.LocalScaling works&lt;br /&gt;body.CollisionShape.LocalScaling &amp;#61; e.Scale&amp;#59; &lt;br /&gt;body.WorldTransform &amp;#61; rotationTransform &amp;#42; translateTransform&amp;#59;&lt;br /&gt;&lt;br /&gt;It works for the scale and position, but doesn&amp;#39;t work for orientation. I would like to pitch the box by 45 degrees, so that falling spheres can roll down from it. Is it possible&amp;#63;&lt;br /&gt;</description><author>arilou</author><pubDate>Sun, 02 Dec 2007 15:14:40 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: How can I rotate a BoxSphape so that a SphereShape can roll down from it? 20071202031440P</guid></item><item><title>COMMENTED ISSUE: WorldPosition's scale being reset</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=12037</link><description>Hello.  As my model grows in real time, I want to update the scaling of the rigidbody.  So I set the rigid body&amp;#39;s world position to have a scale of, in this example, 10.0 for all axis.  However, after I call StepSimulation, the translation is still correct, but the scale is reset to 1.0f.  Is this the intended behavior of BulletX or a bug&amp;#63;  I am hoping this is the correct C&amp;#35; forum.&lt;br /&gt;Comments: ** Comment from web user: arilou ** &lt;p&gt;Instead of setting the scale component of the transformation matrix, apply it to body.CollisionShape property. It works for me that way...&lt;/p&gt;</description><author>arilou</author><pubDate>Sun, 02 Dec 2007 15:09:30 GMT</pubDate><guid isPermaLink="false">COMMENTED ISSUE: WorldPosition's scale being reset 20071202030930P</guid></item><item><title>COMMENTED ISSUE: Solution to giant speed improvement</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=12214</link><description>Ohhh MY GOD&amp;#33;&amp;#33;&lt;br /&gt;Replace everywhere, where possible all method calls with structs, with &amp;#34;ref struct&amp;#34;, instead of struct.&lt;br /&gt;Where are you guiys from, from the Java world&amp;#63; From Delphi&amp;#63; No other way to explain this. It goes through the whole code. You make massive calls to methods where you give over 5, in words&amp;#58; _five_ structs&amp;#33; Without &amp;#34;ref&amp;#34;&amp;#33; These structs will be copied and cloned in every call you make, every time&amp;#33; They behave like &amp;#34;C&amp;#43;&amp;#43;&amp;#34; structs&amp;#33; But I am almost shure, in the original &amp;#34;C&amp;#43;&amp;#43;&amp;#34; Bullet Code were the methods declared this way&amp;#58; &amp;#34;Matrix &amp;#38;m&amp;#34; &amp;#40;&amp;#38; &amp;#61;&amp;#61; ref&amp;#41;. Do this everywhere in your code where it was done in the C&amp;#43;&amp;#43; original, and you should see a giant speed improvement&amp;#33;&lt;br /&gt;Of course it makes no sense to use &amp;#34;ref&amp;#34; when you need a copy of the struct for the called method anyway. So you have to look at each method call with structs &amp;#40;Vector2, Vector3, Matrix...&amp;#41; and consider if you commit it by &amp;#34;ref&amp;#34; or by copy.&lt;br /&gt;&lt;br /&gt;&amp;#40;The project I looked at was &amp;#34;Bullet XNA&amp;#34;&amp;#41;&lt;br /&gt;&lt;br /&gt;I wrote a simple test with only 4 Matrices committed as parameters to a method, and the speed difference to the better &amp;#34;ref&amp;#34;-solution is around 15&amp;#37;&amp;#33;&lt;br /&gt;This is also the way the MS XNA Team recommends&amp;#33;&lt;br/&gt;Comments: ** Comment from web user: XnaCoder ** &lt;p&gt;I did it myself. While it is a speed improvement and should be included into the code, it was not the big speed improvement I hoped for. The engine is too busy with other important stuff, where no Matrices are committed as parameters.&lt;/p&gt;</description><author>XnaCoder</author><pubDate>Thu, 16 Aug 2007 12:05:01 GMT</pubDate><guid isPermaLink="false">COMMENTED ISSUE: Solution to giant speed improvement 20070816120501P</guid></item><item><title>CREATED ISSUE: Solution to giant speed improvement</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=12214</link><description>Ohhh MY GOD&amp;#33;&amp;#33;&lt;br /&gt;Replace everywhere, where possible all method calls with structs, with &amp;#34;ref struct&amp;#34;, instead of struct.&lt;br /&gt;Where are you guiys from, from the Java world&amp;#63; From Delphi&amp;#63; No other way to explain this. It goes through the whole code. You make massive calls to methods where you give over 5, in words&amp;#58; _five_ structs&amp;#33; Without &amp;#34;ref&amp;#34;&amp;#33; These structs will be copied and cloned in every call you make, every time&amp;#33; They behave like &amp;#34;C&amp;#43;&amp;#43;&amp;#34; structs&amp;#33; But I am almost shure, in the original &amp;#34;C&amp;#43;&amp;#43;&amp;#34; Bullet Code were the methods declared this way&amp;#58; &amp;#34;Matrix &amp;#38;m&amp;#34; &amp;#40;&amp;#38; &amp;#61;&amp;#61; ref&amp;#41;. Do this everywhere in your code where it was done in the C&amp;#43;&amp;#43; original, and you should see a giant speed improvement&amp;#33;&lt;br /&gt;Of course it makes no sense to use &amp;#34;ref&amp;#34; when you need a copy of the struct for the called method anyway. So you have to look at each method call with structs &amp;#40;Vector2, Vector3, Matrix...&amp;#41; and consider if you commit it by &amp;#34;ref&amp;#34; or by copy.&lt;br /&gt;&lt;br /&gt;&amp;#40;The project I looked at was &amp;#34;Bullet XNA&amp;#34;&amp;#41;&lt;br /&gt;&lt;br /&gt;I wrote a simple test with only 4 Matrices committed as parameters to a method, and the speed difference to the better &amp;#34;ref&amp;#34;-solution is around 15&amp;#37;&amp;#33;&lt;br /&gt;This is also the way the MS XNA Team recommends&amp;#33;&lt;br/&gt;</description><author>XnaCoder</author><pubDate>Tue, 14 Aug 2007 17:08:58 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: Solution to giant speed improvement 20070814050858P</guid></item><item><title>CREATED ISSUE: Active object getting stuck as IslandSleeping</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=12063</link><description>My main active rigid body becomes unmovable if I let it sit there for about 20 seconds.  After stepping into the code, I have found that its activation state becomes IslandSleeping, and never leaves that state.  Explicitly setting it back to Active allows me to move it around again, but that worries me.  I am not too sure what in all is going on under the covers, but I see it being set to IslandSleeping a lot.  I guess that is a way to help cull out objects to test&amp;#63;  I am applying force to it for movement, and I would guess that &amp;#39;should&amp;#39; cause it to wake up if it is sleeping, but it is not in this case.&lt;br/&gt;</description><author>KleMiX</author><pubDate>Fri, 03 Aug 2007 19:56:04 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: Active object getting stuck as IslandSleeping 20070803075604P</guid></item><item><title>CREATED ISSUE: WorldPosition's scale being reset</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=12037</link><description>Hello.  As my model grows in real time, I want to update the scaling of the rigidbody.  So I set the rigid body&amp;#39;s world position to have a scale of, in this example, 10.0 for all axis.  However, after I call StepSimulation, the translation is still correct, but the scale is reset to 1.0f.  Is this the intended behavior of BulletX or a bug&amp;#63;  I am hoping this is the correct C&amp;#35; forum.&lt;br/&gt;</description><author>KleMiX</author><pubDate>Thu, 02 Aug 2007 19:46:01 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: WorldPosition's scale being reset 20070802074601P</guid></item><item><title>CREATED ISSUE: RemoveOverlappingPairsContainingProxy, bug</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11535</link><description>In C# version, OverlappingPairChache.cs : RemoveOverlappingPairsContainingProxy(), you are iterating through the _overlappingPairs collection in reverse order.  However, when you remove one, you then increment the 'i' counter.  This is not needed when going in reverse order.  If you were removing items going in order, then you would need to decrement the counter after removing one so you don't skip an element.  By doing what you are doing, you will be testing an element twice if the one before it is removed, and cause an exception in other cases.&lt;br/&gt;&lt;br/&gt;When there is only one item in the _overlappingPairs array, this causes an exception.  You can see this by adding two dynamic RigidBodies at the same location, and then removing the second one.&lt;br/&gt;</description><author>Zimbarbo</author><pubDate>Thu, 05 Jul 2007 00:32:27 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: RemoveOverlappingPairsContainingProxy, bug 20070705123227A</guid></item><item><title>COMMENTED FEATURE: XPA and Bullet components must be serializable</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=4305</link><description>XPA and Bullet components must be serializable. 
It will be easier to load all physics data in simulation. Comments: Associated with changeset 23924.</description><author>KleMiX</author><pubDate>Wed, 20 Jun 2007 16:43:51 GMT</pubDate><guid isPermaLink="false">COMMENTED FEATURE: XPA and Bullet components must be serializable 20070620044351P</guid></item><item><title>CREATED ISSUE: Grabbing objects in demos doesn't work</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11155</link><description>In Ccd demo of Release 2.5 Alpha of the BulletX grabbing objects with moddle mouse doesn't work always. I think there is a problem with _world.RayTest(). </description><author>KamiSama</author><pubDate>Sun, 17 Jun 2007 02:46:53 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: Grabbing objects in demos doesn't work 20070617024653A</guid></item><item><title>CLOSED ISSUE: A minor bug in BulletX Source (2.5 Alpha)</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11068</link><description>In BulletX Source (2.5 Alpha) in TriangleMeshShape.cs at line 109:
LocalProcessAllTriangles(callback, aabbMax, aabbMax);
should be:
LocalProcessAllTriangles(callback, aabbMin, aabbMax);

It didn't get the tringle mesh shape collision work, but I think it can help.
Keep it up! </description><author>KamiSama</author><pubDate>Thu, 14 Jun 2007 08:21:33 GMT</pubDate><guid isPermaLink="false">CLOSED ISSUE: A minor bug in BulletX Source (2.5 Alpha) 20070614082133A</guid></item><item><title>COMMENTED ISSUE: A minor bug in BulletX Source (2.5 Alpha)</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11068</link><description>In BulletX Source (2.5 Alpha) in TriangleMeshShape.cs at line 109:
LocalProcessAllTriangles(callback, aabbMax, aabbMax);
should be:
LocalProcessAllTriangles(callback, aabbMin, aabbMax);

It didn't get the tringle mesh shape collision work, but I think it can help.
Keep it up! Comments: ** Comment from web user: KamiSama ** &lt;p&gt;After reversing the triangle windings I could get some results. but it&amp;#39;s not good. It works sometimes. Maybe it&amp;#39;s a scale problem&amp;#63; I&amp;#39;m working on a scene exported from Max and it&amp;#39;s units are centimeters &amp;#40;100 times bigger&amp;#41;.&lt;/p&gt;</description><author>KamiSama</author><pubDate>Wed, 13 Jun 2007 23:45:10 GMT</pubDate><guid isPermaLink="false">COMMENTED ISSUE: A minor bug in BulletX Source (2.5 Alpha) 20070613114510P</guid></item><item><title>CREATED ISSUE: A minor bug in BulletX Source (2.5 Alpha)</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=11068</link><description>In BulletX Source (2.5 Alpha) in TriangleMeshShape.cs at line 109:
LocalProcessAllTriangles(callback, aabbMax, aabbMax);
should be:
LocalProcessAllTriangles(callback, aabbMin, aabbMax);

It didn't get the tringle mesh shape collision work, but I think it can help.
Keep it up! </description><author>KamiSama</author><pubDate>Wed, 13 Jun 2007 23:10:02 GMT</pubDate><guid isPermaLink="false">CREATED ISSUE: A minor bug in BulletX Source (2.5 Alpha) 20070613111002P</guid></item><item><title>COMMENTED FEATURE: Use XNA types instead of custom Bullet types</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=10811</link><description>I see that the custom Vector3 is now deprecated in favor of the XNA version, but it would also make sense to use the BoundingBox and TimeSpan structs in place of the many Vector3 aabbMin/aabbMax and float timeStep references.

I've made the changes in my copy of the source, anyone have an opinion? Comments: ** Comment from web user: KleMiX ** &lt;p&gt;I&amp;#39;ll check in it soon, with support of gamepads&lt;/p&gt;</description><author>KleMiX</author><pubDate>Fri, 08 Jun 2007 05:58:12 GMT</pubDate><guid isPermaLink="false">COMMENTED FEATURE: Use XNA types instead of custom Bullet types 20070608055812A</guid></item><item><title>COMMENTED FEATURE: Use XNA types instead of custom Bullet types</title><link>http://www.codeplex.com/xnadevru/WorkItem/View.aspx?WorkItemId=10811</link><description>I see that the custom Vector3 is now deprecated in favor of the XNA version, but it would also make sense to use the BoundingBox and TimeSpan structs in place of the many Vector3 aabbMin/aabbMax and float timeStep references.

I've made the changes in my copy of the source, anyone have an opinion? Comments: ** Comment from web user: Johnnylightbulb ** &lt;p&gt;Yes - I tested against the demo in the 21894 change set - I&amp;#39;ll wait until the alpha gets checked in before doing anything&lt;/p&gt;</description><author>Johnnylightbulb</author><pubDate>Thu, 07 Jun 2007 14:35:33 GMT</pubDate><guid isPermaLink="false">COMMENTED FEATURE: Use XNA types instead of custom Bullet types 20070607023533P</guid></item></channel></rss>