Dennis Sweeney

While 0.4% on individual benchmarks is probably not noteworthy, it's nice to see on the geometric mean of an entire suite. Very few changes will ever move the pyperformance geometric...

I opened https://github.com/psf/pyperf/pull/118. Is there still interest?

@Marco-Sulla the sporadic results might just be due to string hash randomization: the lowest 3 bits of random strings sometimes collide, so the dict lookup has to take an extra...

I opened https://github.com/nedbat/coveragepy/pull/1394, which I think should help at least a little. Another idea, probably not particular to 3.11: Would it make sense to cache the latest result of `PyDict_GetItem(self->should_trace_cache,...

Can you describe how to reproduce this issue?

> op=((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((0x0, None), 83), None), 55), None), 13), None), 42), None), 70), None), 7), None), 8), None), 51), None), None), ()), (...)), (...)), None), None), 15), None), 64), None), 94),...

A benchmark: ```python from time import perf_counter from pathlib import Path import sympy.integrals.rubi.rules.sine as sine text = Path(sine.__file__).read_text("utf-8") t0 = perf_counter() for _ in range(10): compile(text, "sine", "exec") t1 =...

I could measure no compilation performance difference from adding `fast_scan_many_locals`, but it does restore some of the functionality. I ran this: ``` import sympy.integrals.rubi.rules.sine as sine from collections import Counter...

@iritkatriel or @markshannon, are there any objections to adding `basicblock->b_unsafe_locals_mask`? If not, I think this is ready.

I'm getting the following FutureWarning on a certain regular expression. I think it just needs "[]" to become "\\[\\]". 0:05:36 load avg: 0.00 [ 53/427] test_check_c_globals ...\\Tools\\c-analyzer\\c_common\\tables.py:236: FutureWarning: Possible nested...