[Bf-committers] Re: CVS commit: blender/source nan_definitions.mk

[Hans Lambermont]

John K. Walton wrote:

> yes and when i looked i saw:
>    ifneq ($(OS), irix)
>       export NAN_SDL ?= $(shell sdl-config --prefix)
>       export NAN_SDLLIBS ?= $(shell sdl-config --libs)
>       export NAN_SDLCFLAGS ?= $(shell sdl-config --cflags)
>    endif
> 
> in a section that explicitly says "generic defaults for all platforms
> which should be constant"

But this (bad ifdef list idea in itself) broke it for FreeBSD. FreeBSD
thinks we should use sdl11-config instead :-(

And with this ifneq part in front of the platform dependent section all
other NAN_SDL* definitions in the platform dependent section wouldn't
work anyway. That't why I cleaned it all up.

> dependant. IMHO you don't set defaults, you test to make sure
> every platform set it to something in the end.

With the requirement to be able to predefine NAN_ stuff *and* have at
least more than one platform that differs (we have 4) this is the only
way.

> but i'm a hardware engineer, and we have to get it right the first
> time (yeah, like THAT ever happened :-) . anyway, easy fix! and irix
> is compiling again.

yep :) Sorry for the inconvenience, apparently noone tested my proposed
patch on Irix when I sent the announcement a week ago.

> since nobody ever gave me write cvs permissions, so all i can do

I think you just need to ask, I cannot deal out cvs write permissions
anymore since we moved to the sourceforge setup.

regards,
  Hans
-- 
http://lambermont.webhop.org/   () ascii ribbon campaign - against HTML mail,
                                /\ vCards and proprietary formats