[Bf-committers] Mac OS X status

[Stefan Gartner]

On Thursday 19 December 2002 21:22, Maarten Gribnau wrote:
> Hi Stefan,
>
> > yesterday I tried compiling blender on Mac OS X 10.2.2 using the
> > original
> > makefiles and succeeded after fooling around a bit:
>
> Good to hear it is working. I am running on the exact same system.
>
> > * in /System/Library/Frameworks/OpenGL.framework I made GL a symlink =
to
> > Headers (so that the compiler finds header files like <GL/gl.h>)
>
> Strange, this was already present on my machine and I did not know it
> was not standard. It could be one of the NaN developers did this
> because I am working on a ex-NaN machine.

Actually, I created the symlink a long time ago, but I thought I mention =
it=20
because it seems to be a problem for several people.

>
> > * I had to uncomment the line defining NAN_PYTHON_BINARY in
> > nan_definitions.mk
> > (what is it used for?)
>
> In the repository it is uncommented. According to CVS it was entered by
> Kent "Committing LarstiQ's patch for NAN_PYTHON_BINARY and
> NAN_MXTEXTTOOLS". I don't know what it is for either.
>
> > * as the README already suggests, I had to run ranlib on the .a files
> > in
> > lib/darwin-6.2-powerpc and obj/darwin-6.2-powerpc. libfrozen.a was a
> > bit odd
> > though, as it seems to be recreated everytime I run make. My
> > (short-term)
> > solution was to edit source/blender/bpython/frozen/Makefile and add
> > "ranlib
> > $(OCGDIR)/blender/bpython/$(DEBUG_DIR)libfrozen.a" to the libfrozen
> > target.
> > Unfortunately the makefile is generated by running the freeze script =
in
> > intern/python/freeze...
> > (On desaster's tree I just edited the Makefiles in intern/ to run
> > ranlib after
> > creating the .a files in lib/darwin-*, which seems to work. Maybe
> > that's a
> > solution to the ranlib problem. For some reason the problems with
> > libfrozen.a
> > didn't show up there, though.)
>
> What is desaster's tree?

Upi Tamminen (desaster) has created an experimental cvs tree at=20
http://nyarlathotep.dragonlight.fi/cgi-bin/viewcvs.cgi/blender

>
> > * I had to create a lib dir in source/ode (otherwise ar whines about
> > lib/libode.a not existing or something like that), but i think this i=
s
> > related to cvs deleting empty directories and not a fault of the buil=
d
> > system
> > itself.
>
> You don't have to build ode anymore. The current Makefiles skip ode and
> link with the ode lib that is in CVS in the lib/darwin-6.1-powerpc
> directory. Maybe you should update you Makefiles?

Yeah, I probably just wanted to do some extra work :)

>
> > * the gameengine doesn't seem to work (I think it's the same on
> > Linux). Does
> > anyone have an idea why?
>
> It should run, both in Blender and the stand-alone game engine. There
> is a major problem in that the old files have their physics (and the
> ray sensor) broken.

I figured that I needed -DUSE_GAMEBLENDER=3D1 when I built it on linux th=
e other=20
day. It works there, so I guess it should work on OS X as well.

>
> > I'll try building it with Project Builder next and see how far I'll
> > get...
>
> I am curious! The version in CVS is the first attempt. Be gentle!
> If you run into the problem that Blender's window is too large, you can
> give the famous command line option -p 32 32 800 600 through an option
> in Project Builder. It is hard to find (at least it took me a while).
> In the targets screen, click on the blender executable.

Thanks for the tip, I didn't have the time to compile it yet, though.

>
> Maarten
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers@blender.org
> http://www.blender.org/mailman/listinfo/bf-committers