Ruslan Muhlinin

Yes, it's a great idea but it must be good reason to break backward compatibility. There are projects based on qml-box2d and that could affect them too. Btw, you can...

May be in some particular case it makes sense. But it would be great to just move the `World` instead of that, I think.

Yes, `World` is just a phisical environment without visual representation but it still has base coordinates for bodies. Now it always (0,0). That's what I meant when I said "to...

Ok, guys, can you try this example, please: ``` javascript import QtQuick 2.4 import Box2D 2.0 import QtQuick.Window 2.2 Window { width: 800 height: 600 visible: true World {} Item...

No, that just extends the problem. I've updated my example to exhibit the problem. ``` javascript import QtQuick 2.4 import Box2D 2.0 import QtQuick.Window 2.2 Window { width: 800 height:...

Ok, I've updated my comments.

Sounds good, in terms of performance improvements. Also I think it have to be done in some way to avoid recreation in certain cases. For example: ``` javascript Body {...

I get these warnings too while compiling the source with `MinGW` in `Release` mode. Than does not happen with `MinGW`/`Debug` mode, `clang` or `gcc` in Linux. And yes, as @bjorn...

What about b2Contact::GetManifold ()? We have the instance of b2Contact class inside BeginContact/EndContact/PreSolve/PostSolve so maybe it would be convenient to add 3-4 variables of b2Manifold to Box2DContact class? I mean...

Yes, you're right @bjorn, I 've read it inattentively. btw, why we emit `beginContact`/`endContact` inside `Box2DWorld::step()`? Wouldn't it be right to emit them directly, from corresponding functions of `ContactListener` as...