Tutorials for debugging native code.
Complete the setup before the session. One root Pixi workspace provides separate environments for the two exercises:
- The NumPy reference-counting tutorial uses
debug Python, SIGUSR1, and an LLDB hardware watchpoint to follow a one-line
bug across the extension/interpreter boundary. Activate it with
pixi shell --locked. - The Samply profiling tutorial uses free-threaded
Python to explore thread scaling and compare native profiles with Python
sampling profiles. Activate it with
pixi shell --locked -e profiling.
Run exit before switching environments. Commands in bash blocks run in
your terminal; commands prefixed with (lldb) run at the debugger prompt.