No mention of XGL?

I’m surprised that no one’s mentioned XGL yet.

Is an equivalent to XGL for Haiku planned, or will the OS just make use of its POSIX compatibility layer to port XGL over?

raynevandunem wrote:
I'm surprised that no one's mentioned XGL yet.

Is an equivalent to XGL for Haiku planned, or will the OS just make use of its POSIX compatibility layer to port XGL over?

Neither, I guess.

First, I think you know that Haiku doesn’t use the X Window System and, considering Haiku aims to be an OS with a clean design, a kludge with an X-Server running in parallel to Haiku’s App Server is (fortunately) very unlikely to happen, unless it’s being done by a third party.

Second, is XGL being released under the GPL? That would mean that Haiku would probably get into conflict with that license.

Third, AFAIK there’s no agreed-upon plan for the desktop environment beyond the first release - at least I seem to remember some debate on the issue of eye-candy, UI skinning etc. on the mailing list recently. And if I got the gist of it right, you probably won’t get pixel shaders on icons and particle systems when opening menus. :wink:

I think Rudolf was experimenting with drawing 2d graphics using his 3d engine for nVidia cards - I think remember some comments about it on his blog at one point.

In any case, as ChrisK alluded to, I suspect you won’t see a lot of shading, transparency, or skinning support in Haiku any time soon if several of the core developers can prevent it.

It’s sort of a hotly-debated topic both here on the forums, as well as the mailing list.

Moving to a 3D accelerated (via OpenGL) graphical compositing is not planned for Haiku R1. But it was, is and will be highly debated before we decide how to do it for a future release.
Most probably, it won’t be a plain XGL port but a native one. Nathan Whitehorn do some limited tests in the past. François Revol as well, IIRC.

Anyway, for R1, nope, for post R1, most probably. But before we need those 3d accelerated OpenGL renderers add-ons to be written…

phoudoin wrote:
Moving to a 3D accelerated (via OpenGL) graphical compositing is not planned for Haiku R1. But it was, is and will be highly debated before we decide how to do it for a future release. Most probably, it won't be a plain XGL port but a native one. Nathan Whitehorn do some limited tests in the past. François Revol as well, IIRC.

Anyway, for R1, nope, for post R1, most probably. But before we need those 3d accelerated OpenGL renderers add-ons to be written…

Understandable, as the main purpose of R1 is simply to be as compatible with BeOS R5 as possible, which means that you’ll be using GCC v2.5 (turn of the century?) rather than v4.1 (as of May 2006). Obviously, I don’t think such an ancient version could handle something as complex as a native MIT-licensed graphics layer to rival Quartz 2D Extreme, XGL, or DWM/Aero, right?

Maybe Glass Elevator? It’s supposed to break all backwards compatibility with R1 (thus compelling those who were developing apps for BeOS to develop for Haiku on its own merit); so is a break with R1’s graphics or API (which will behoove the inclusion of such a graphics layer) intended as well?

Or is a smoother transition intended, one in which the core system will be updated, but the graphics and API will remain the same until the next major release (R3)?