The Panda3D app distribution for macOS allows one to select platform tags exclusively for the x86_64 architecture (Intel Macs):
https://docs.panda3d.org/1.10/python/distribution/building-binaries#platform-tags
What about the arm64 architecture? The documentation doesn’t mention this. Is there no possibility? Or is it not documented?
Interestingly, at the IDE level, Panda3D generally works with Python natively running on arm64.
No, this is not yet fully supported. It can build but the code signing restrictions are stricter for arm64 so the produced build doesn’t run.
We currently have someone looking into it.
Track this issue for updates:
I found that the current Panda3d will work on arm64 macbooks using rosetta2 to run the x86 compiled libraries. You can force the macbook to use rosetta2 (x86 emulation) with
arch -x86_64 <command>
I found it was a good idea to install a virtual environment for Python and install x86 versions of key Python modules there (e.g. numpy). You can set-up a python path in the venv to choose the Panda3D x86 installed libraries. On my mac they were put in /Libraries/Developer/Panda3d
arch -x86_64 python3 -m pip install --user virtualenv
arch -x86_64 python3 -m venv zshenv
source zshenv/bin/activate
arch -x86_64 python3 -m pip install numpy
arch -x86_64 python3 -m pip install matplotlib
echo "/Library/Developer/Panda3D" > zshenv/lib/python3.12/site-packages/panda3d.pth
For here I can just run without the arch -x86_64, e.g.
python panda3dtest.py
For repeat usage, it was convenient to just add
arch -x86_64 zsh -i
To the end of the activate script so I could just do
source zshenv/bin/activate
This was particularly useful in VSCode since you can automate starting the venv but it wants to just run activate.
I am probably taking a performance hit from the rosetta2 x86 emulation but it was fine for simple examples.