Ned Batchelder

@markshannon @pablogsal I am also seeing a problem with coverage.py being 2x slower on 3.11 than on 3.10: ``` Median for bug1339.py, python3.10, cov=none: 0.180s Median for bug1339.py, python3.10, cov=6.4.1:...

Here are the same numbers in a more comprehensible form: | Coverage | file | 3.10 | 3.11 | 3.11 vs 3.10 |----------|--------------------------------|-----------------|-----------------|-------- | none | bug1339.py | 0.180 s...

Here is how I ran the tests. You can adjust the details if you want. - Get lab/benchmark.py from this repo: https://github.com/nedbat/coveragepy/blob/50e8d938315dacce9c5d3de1a25c771cdfe943ec/lab/benchmark.py - Clone https://github.com/plasma-umass/slipcover . I put it at...

@pablogsal I'm not sure how to subscribe you, but I pinged you, maybe that's what you meant: https://github.com/python/cpython/issues/93516

> Would it make sense to cache the latest result of PyDict_GetItem(self->should_trace_cache, filename)? That call only happens during a CALL event, so I'm not sure it would pay off, but...

I tried this out in this pull request: https://github.com/nedbat/coveragepy/pull/1401 It seemed to only improve things by about .5%

I'm not opposed to this, but could use help changing the tooling.

*Original comment by* **Abdeali Kothari (Bitbucket: [AbdealiJK](https://bitbucket.org/AbdealiJK), GitHub: [AbdealiJK](https://github.com/AbdealiJK))** ---------------------------------------- Any thoughts on this Ned ? I'm not sure if I'm doing something wrong for the subprocess thing. Or is...

*Original comment by* **Abdeali Kothari (Bitbucket: [AbdealiJK](https://bitbucket.org/AbdealiJK), GitHub: [AbdealiJK](https://github.com/AbdealiJK))** ---------------------------------------- I am running it with my root anaconda - so, I think this is the right one. (Considering it was...

@AbdealiJK Thanks for doing all this! One thing that looks wrong to me: the COVERAGE_PROCESS_START environment variable needs to refer to the location of the .covergerc file to use: ```...