...ok, now we're getting somewhere...after IOhannes' changes, I can now run autoconf and test my changes to configure.in, and I've gotten thru the configure phase...haven't committed it because I haven't figured out where to put the framework linkage flags...or where to change it from "Gem.pd_linux" to "Gem.pd_darwin"...
...but then I hit 'make', and for some reason it tries to compile a "Manips/copy.o"? This file isn't in my cvs, so...? Maybe it went wrong earlier: is it supposed to print out the following before the "echo"?
tiggity:~/puredataDev/Gem/src/gnu tigital$ make for d in ../Manips ../Particles ../Base ../MarkEx ../Pixes ../ Controls ../Nongeos ../Geos ../openGL; do \ echo ""\ echo "----------$d----------"; \ cd $d && make -j4|| exit -1; \ done echo ----------../Manips---------- g++ -c -g -O2 -fPIC -DMACOSX -DYUV -DALTIVEC -fno-builtin -Os -c - faltivec -fno-builtin -falign-loops=16 -funroll-loops -ffast-math - I../ -I/Users/tigital/puredataDev/GemLibs/freetype-2.1.9/include/ -I/ Users/tigital/puredataDev/GemLibs/FTGL/include/ -I/Users/tigital/ puredataDev/pd-0.38-4/src/ accumrotate.cpp -o accumrotate.o
...hmmm...I'm also unsure of where it's picking up the "-O2"? And another question: how do I make a debug (-O0) build?
l8r, jamie
On May 20, 2005, at 10:40 PM, james tittle wrote:
g++ -c -g -O2 -fPIC -DMACOSX -DYUV -DALTIVEC -fno-builtin -Os -c -faltivec -fno-builtin -falign-loops=16 -funroll-loops -ffast-math -I../ -I/Users/tigital/puredataDev/GemLibs/freetype-2.1.9/include/ -I/Users/tigital/puredataDev/GemLibs/FTGL/include/ -I/Users/tigital/puredataDev/pd-0.38-4/src/ accumrotate.cpp -o accumrotate.o
The -DMACOSX is not needed because the compiler will pick up the __APPLE__ by default. Same with the -DALTIVEC since -faltivec defines __VEC__ anyway. The -O2 and -Os conflict and who knows which is used. Some sort of detection for the host CPU or a way to specify the chip would be nice too although there is a -ffast type flag that covers most of that.
I have no idea about the other stuff.