Levi Burner
Levi Burner
I agree on both points. Linking to GLFW from the bindings isn't desirable to begin with, however it gets even worse when you consider that the Python `glfw` package brings...
Hi, I have updated the bindings so that the Python bindings no longer link to `glfw` and the C++ physics loop is reused. The bindings are very different from the...
What timing! I think it all works out in the end. The latest bindings do not need to `import glfw` nor know anything about `glfw` (they don't even need the...
I agree. I did not mean to imply that the dispatch table wasn't needed, it definitely is for the reason you said. I just mean that the 2nd version of...
Ok the dynamic dispatch table works, the latest commit is sets the handle and mixes Python and C++ glfw calls. Very cool! :) However, I had to remove the `dlclose`...
Sorry just found the issue, it's something very basic, as usual :). The `Glfw()` prefixes were missing in the physics loop. In this PR I move code from `main.cc` to...
Should `libsimulate` remain a static library? Previous versions of this PR built `libsimulate` as shared so that GLFW/lodepng symbols were present in `liblibsimulate.so` which was necessary because CMake won't copy...
That will work! Thanks for the quick response.
Hi, This should be ready for a detailed review. I've been able to test Ubuntu and MacOS but not Windows. For some reason Windows is taking excessively long to compile...
I've tested on Windows. `main.py`, `double_pendulum.py`, and `fixation.py` all work.