Get latest panda build from cvs.
Copy in thirdparty directory from pre-compiled panda version. (As template.)


sudo apt-get install bison
sudo apt-get install flex
sudo apt-get install libosmesa-dev
sudo apt-get install libssl-dev

export PATH=/home/lethe/Desktop/panda3d-1.3.2/built/bin:$PATH
export LD_LIBRARY_PATH=/home/lethe/Desktop/panda3d-1.3.2/built/lib:$LD_LIBRARY_PATH


makepanda/makepanda.py --everything --lzma --version=1.3.2 --no-dx8 --no-dx9 --no-maya6 --no-maya65 --no-maya7 --no-max6 --no-max8 --no-max7 --no-ffmpeg --no-fftw --no-vrpn --no-fmodex --installer


Need directory 'samples' in base directory.


cppExpression.h:95 to
return _u._integer;

cppExpression.h:119 to
return (double)_u._integer;


makepanda.py:1224 to
cmd = cmd + " -lpthread -ldl" # Hacky, but works.


pystub.cxx:
Created Py_InitModule4_64 in the two locations straight after Py_InitModule4, identical.


(Below are all wrong approach for bitMask, but correct approach would require deep understanding.)

bitMask.h:170 to
typedef BitMask32 BitMaskNative;

bitMask.h:143 & 145 - comment out

bitMask.I:878
Rename to arbitary name, edit below function to use it. Ambiguity resolution.

bitMask.I:897
Comment out 32 bit version of function, simple way of resolving ambiguities.


config_nativenet.cxx:61 - surround with #ifdef HAVE_OPENSSL / #endif - thats not a 64 bit bug!


stl_compares.I:94 & 95, change to pass in the entire pointer.


Download Cg library from Nvidia website, extract the .so files into the thirdparty directory.
Download Fmod library, from fmod.org, go over thirdparty directory, again. I just renamed files to match names already in directory, rather than finding where its speced in the compile scripts. Copied over include files, not sure if needed.

Checkout netscape portable runtime library [nspr] from mozilla.
Go through directrory structure, edit LIBRARY_NAME in assorted makefile.in files - in 3 places for each of the 3 files.
Follow the provided build instructions from mozilla, you need all 3 given options!


dcPacker.cxx:
Edit int to Py_ssize_t on line 765.