[SOLVED] Mac os x retina problems

[osgoz]

Hi

We have been traying to do an aplication for Mac OSX.
All was fine until we test it in a retina macbook pro.
The irrlicht device was a quarter of the screen.
We managed to expand the device to the real size (2560x1600) changing the source code in CIrrDeviceMacOSX.mm
But after that the coords that irrlicht returns to test colisions for select objects are wrong.
Does anybody have this issues, or are only that we are doing something wrong?
And when we test it on an intel graphics macbook pro retina 13" , the meshes are displayed with the material and the wireframe too.

Apologices for my english, it is not my mother language

Regards.

[Nadro]

I don't have access to any retina displays on OSX, maybe retina displays are handle under OSX in similar way like in iOS, where you have to mul all window cordinates by scaleFactor, but I'm not sure.

[osgoz]

Thank you for your reply.

Yes I think that too.
Indeed, Nadro, we are using your version of GL-ES irrlicht for tests on iOS and we need to scale all coords by the retina scale factor.
But this is not the only problem. In a macbook pro retina 15" with a discrete graphics card (nvidia) irrlicht renders ok.
But in a macbook pro retina 13" with only an integrated intel HD graphics card, irrlicht render the meshes with the textures and the framewire visible too.
Please see the attached images for the integrated graphics card.
https://www.dropbox.com/s/enca2f6ubyzsk ... r.jpg?dl=0

Any idea?

Regards

[osgoz]

I answer my self.

On a retina display there is no need for antialiasing, so "good" graphics cards (discrete cards) support it on retina display, but integrated ones does not.
So setting antialias to 0 when initializing the device its suffice to correct my problems.

Nadro, as you said, simply we have to mul all window cordinates by scaleFactor and it is working now.
Thank you for you time.

Regards.