[Bf-committers] 2.28a release & meeting notes

[John K. Walton]

give me a URL to begin from. and a week to get back to you, i'll
be on vacation.

On Tue, 12 Aug 2003, Dean Giberson wrote:

> If there are specific problems with GHOST on a platform I would like to 
> know what they are. Of course I don't have access to that kind of 
> hardware. If you would be willing to provide some profiling for GHOST 
> then I (and I bet the rest of the GHOST maintainers) would be willing to 
> look into fixing it.
> 
> If you have the chance grab the cvs daily tarball & build the GHOST 
> tests. From there it should be easy to find out why the effect you are 
> seeing is happening.
> 
> As a pure windowing system GHOST is a much cleaner design than Glut. 
> I've used both & I'm happy that GHOST was created. Please give us a 
> chance to fix it before stripping blender of this tool.
> 
> Thanks,
> Dean Giberson (voidptr)
> 
> John K. Walton wrote:
> > 
> > While I agree with everything you said, there are serious performance
> > problems in 3d window with ghost based versions. mips3, or older
> > O2's suffer from annoying lag when rotating (and stuff), as if it's
> > polling at a low rate and missing the mouse action. once the action
> > gets going, it's ok.impact type systems (which were the only ones
> > after early blender days) never had this problem. 
> > 
> > my understanding behind externalizing ghost was to make the display
> > subsystem independant (a good thing). one use for this would be to
> > use GLUT (even if it is ugly, everything is ugly from one perspective
> > or another) when appropriate, and ghost when appropriate or some
> > distributed multidisplay visualization thing. and not just to make
> > an arbitrary API.
> > 
> > to summarize, I'm looking for an opportunity, not applying a restriction
> > or objecting in anyway, I've got enough hardware, I can benefit from
> > anything short of dropping everythign except OSX ;-)
> > 
> > On Tue, 12 Aug 2003, Ton Roosendaal wrote:
> > 
> > 
> >>Hi,
> >>
> >>Apart from a major missing feature (24 bits display) ghost works  
> >>technically much better than Glut. In '98 I had to to do lot of hacks  
> >>in the glut code to make it working, and it was needed to patch Blender  
> >>for it too. This was so much more ugly and awkward than the previous  
> >>used nice irisgl windowing lib.
> >>Glut has (and had) no official ports for BeOS or OSX available (the  
> >>current beGlut and OSX-glut are forks), which made & makes porting  
> >>Blender very cumbersome.
> >>
> >>I've checked your forwarded comments from Zr on adding 24 bits support  
> >>for Ghost. It really requires someone with more Ghost and C++ insight  
> >>and an X11 development envirmont to do it.
> >>
> >>There's a load of unix coders here though... nobody interested in cool  
> >>single buffered 24 bits display for rendering?
> >>
> >>(from zr:)
> >>The quick fix solution would be:
> >>Add single buffer mode argument to C++ ghost constructor
> >>Add same arg to GHOST_CreateWindow
> >>Add same arg to window_open
> >>Add code to renderwin.c to determine what mode to use.
> >>
> >>I can help with renderwin.c, that's what I'm familiar with.
> >>
> >>-Ton-
> >>
> >>
> >>On Tuesday, Aug 12, 2003, at 15:48 Europe/Amsterdam, John K. Walton  
> >>wrote:
> >>
> >>
> >>>On Mon, 11 Aug 2003, Ton Roosendaal wrote:
> >>>
> >>>>2. ghost
> >>>>- the Ghost team intends to go for an 'official launch'... they're
> >>>>finalizing docs and demos for it. Ton will help getting marketing
> >>>>exposure at opengl.org
> >>>>- this also implies we (bf-blender) should get ready to remove ghost
> >>>>from the Blender cvs... and include it as extern lib (such as jpeg).
> >>>>- agreed on is to freeze now bf-blender ghost; meaning that commits to
> >>>>it are not allowed unless it's also done in the official ghost  
> >>>>project.
> >>>>During the next period, hopefully before the 2.29 release, we can
> >>>>discuss & decide on how to incorporate the ghost lib.
> >>>>- Ghost is being ported to BeOS (R5, last official release).
> >>>
> >>>is ghost being externalized, and abstracted to the degree that
> >>>we could put GLUT back in for the hardware that worked MUCH
> >>>better with it?
> >>>
> >>>
> >>>_______________________________________________
> >>>Bf-committers mailing list
> >>>Bf-committers@blender.org
> >>>http://www.blender.org/mailman/listinfo/bf-committers
> >>>
> >>>
> >>
> >>------------------------------------------------------------------------ 
> >>--
> >>Ton Roosendaal  Blender Foundation ton@blender.org  
> >>http://www.blender.org
> >>
> >>_______________________________________________
> >>Bf-committers mailing list
> >>Bf-committers@blender.org
> >>http://www.blender.org/mailman/listinfo/bf-committers
> >>
> > 
> > 
> > _______________________________________________
> > Bf-committers mailing list
> > Bf-committers@blender.org
> > http://www.blender.org/mailman/listinfo/bf-committers
> 
> _______________________________________________
> Bf-committers mailing list
> Bf-committers@blender.org
> http://www.blender.org/mailman/listinfo/bf-committers
>