Re: GNUstep on DirectFB

[Lisbon Acid]

Hello,

It has been a long time since I have been around these parts. I've been working hard as a physical chemist researcher the past months, and as such have been out of the loop regarding the progress of GNUstep and its graphical backends.

A project I have been working on lately, however, and what I'm writing about specifically is a Linux framebuffer based client/server windowing system written in ObjC with the GNUstep foundation libraries. The drawing and compositing of surfaces is done with OpenGL (right now, just offscreen software rendered mesa). The class structure is designed for the GNUstep GUI, and is done in a modular way as to allow for future changes in the display driver, context allocation, opengl library, etc.

I'm curious about the thoughts of those who work on this project regularly. I'm very fond of the GNUstep project, and my aim is to provide an open, accelerated, lightweight graphics backend tailored to the GNUstep framework, rather than adapted to it.

On Mon, Dec 8, 2008 at 8:45 AM, Riccardo Mottola <address@hidden> wrote:

Hi,

I can't be sure, but I think it might be some weird endianness problem, although you are working on intel only, guessing from your screenshots.

In my case the porblem is when the application is on a big-endian machine and the display is on a little-endian one. I don't know how exactly the bytes are encoded though and why cairo is doing it wrong. Direct xlib in gnustep works fine.

For the backend itself, once you have something even remotely usable (those visual bugs can be fixed for sure) you should probably sign up a FSF release and contribute it to gnustep. Ask Gregory for that (partecipating to the project itself) and Fred Kiefer (current gui/back maintainer).

Your code should also follow the GNU style. Ask those two persons for help in any case.

Riccardo

On 2008-12-08 00:33:24 +0100 "Lisbon Acid" <address@hidden> wrote:

Awesome, thank you for the info. Do you know why it would happen on some
surfaces and not others? That is the only part that is puzzling me about the
shifting. It may, though, be a problem in my code... not entirely sure. I'll
get to it after these finals are over...

I'm glad there is an interest in this backend. I believe bringing these two
projects together has the potential to catalyze interesting developments...

On Sun, Dec 7, 2008 at 4:36 PM, Riccardo Mottola <address@hidden> wrote: