Automate upstream llama.cpp sync

I'm trying to build with a newer upstream version of llama.cpp. Syncing the changes from upstream `llama.h` into `llama_cpp.py` is quite a toil. Has anyone (successfully) attempted to use LLM assistance to make it less painful?