Small OgreBites bug?

[lingfors]

I think I just found a small bug in OgreBites/SdkTrays:

Code: Select all

/*-----------------------------------------------------------------------------
| Converts a 3D scene position to a 2D screen coordinate (in pixels).
-----------------------------------------------------------------------------*/
static Ogre::Vector2 sceneToScreen(const Ogre::Camera* cam, const Ogre::Vector3& pt)
{
	Ogre::Vector3 result = cam->getProjectionMatrix() * cam->getViewMatrix() * pt;
	return Ogre::Vector2((result.x + 1) / 2, -(result.y + 1) / 2);
}

should be

Code: Select all

/*-----------------------------------------------------------------------------
| Converts a 3D scene position to a 2D screen coordinate (in pixels).
-----------------------------------------------------------------------------*/
static Ogre::Vector2 sceneToScreen(const Ogre::Camera* cam, const Ogre::Vector3& pt)
{
	Ogre::Vector3 result = cam->getProjectionMatrix() * cam->getViewMatrix() * pt;
	return Ogre::Vector2((result.x + 1) / 2, (-result.y + 1) / 2);
}

[Mind Calamity]

Aren't those two code blocks you just posted exactly the same ?

[syedhs]

I think lingfors forget to multiply the return value by the screen dimension - which is what the patch is all about.

Edit: Or maybe the code simply returns the 0.00 to 1.00 which works okay, but it is the comment that should be changed. I don't read actual source code though.

[spacegaier]

Then please add it to the patch tracker. Otherwise it will get lot here...

[CABAListic]

The difference is in the placement of the - sign before result.y.

[Mind Calamity]

The difference is in the placement of the - sign before result.y.

Oh, I missed that! The lengths of the return lines were exactly the same so I didn't even bother to read it.

[lingfors]

The difference is in the placement of the - sign before result.y.

Oh, I missed that! The lengths of the return lines were exactly the same so I didn't even bother to read it.

The devil is in the detail.

Added to the bug tracker.

Although, as has been pointed out, the comment isn't correct either...