Ned Batchelder

This is now released as part of [coverage 7.6.10](https://pypi.org/project/coverage/7.6.10).

I'd like to reproduce this, but I don't use poetry so I'm not sure how to get everything in order. Can you give the exact commands you used, starting with...

I think the three lines are the docstrings, but I'm not sure why they are treated differently.

There is no OS difference in the settrace implementation. Likely the difference is do to OS differences in textual repo. If you don't think that's the case, let's look at...

> But could you confirm: > len(missing_lines) + len(executed_lines) + len(excluded_lines) == total_stmts > > Is invariant for every file under coverage? As I mentioned, docstrings complicate matters. I'm also...

You have logging in your coverage management code. Can you show the output you are seeing? Also, are there new subprocesses spawned to handle requests?

You should read this: https://coverage.readthedocs.io/en/latest/subprocess.html coverage 5.3 is unsupported and Python 2.7 is unsupported, but that might fix your problem.

Thanks for the reproducible report. It seems that torch creates and executes a file in your temp directory. Some other ways to avoid the message: ``` coverage run --source=. test_foo.py...

> However, it's still worth documenting this weird behavior. I'm not sure what you want added to the documentation. This behavior is due to an intricacy of PyTorch.

I see what you mean. The message could be longer: "Couldn't find source for code observed at ...." or something? I'm not coming up with a one-line message that explains...