Mark Shannon
Mark Shannon
How can you tell the impact of allowing the extra backward edge in the trace? I suspect that it is beneficial, but some data would be good. Any toy examples...
I patched main with just the changes to optimizer_bytecodes.c, but not the change to JUMP_BACKWARD handling, and I get this pair of traces (running for a bit longer to warm...
You'll need to update the test, which is a bit awkward as there are two executors to check if you're looking for the `_YIELD_VALUE`. We really need a proper test...
Android build failure looks unrelated
Can we have a test for this?
@Fidget-Spinner We could add a "jit_depth" field to the thread state, initialized to 0. * On entering `_PyEval_EvalFrameDefault`: if jit_depth is non-zero, increase it by one. * On leaving: if...
Let's not worry about a reproducer, but ensure that tracing is active only once in any thread. Then we don't need to worry about "inner" or "outer" traces, as there...
Not once across all threads. Once *per* thread.
How do you detect when to stop tracing? If an exception is raised and tracing is active, but higher up the stack, how do you know *not* to stop tracing?
Maybe, but the current approach seems hard to follow and fragile. I'm fairly sure we don't always stop tracing cleanly when exceptions are raised. I think we need a clear...