[report] omit overwrites [run] omit

[nedbat]

Originally reported by Max Nordlund (Bitbucket: maxnordlund, GitHub: maxnordlund)

---

First let me say thanks for providing such an amazing tool, really makes a huge difference for code health.

I'm using pytest-cov, but it doesn't parse the .coveragerc so I believe it's a coverage bug. Basically if you provide an omit option in [run] and [report] the last in the file wins.

This is because both store their value in _omit, as you can see here. I'm not sure how to fix it though, as I just started to read through the source.

---

• Bitbucket: https://bitbucket.org/ned/coveragepy/issue/622

[nedbat]

Thanks for using coverage.py. Can you provide more detail? If reporting is going to omit a file, then why is it a problem for run to omit it?

[nedbat]

Original comment by Max Nordlund (Bitbucket: maxnordlund, GitHub: maxnordlund)

---

I want to exclude some files completely, not contributing to the global
coverage score, but there are some I would like to avoid in the report even if
they contribute to the total.

It sounds kind of weird, but it's to reduce noise until coverage goes up
enough. It's a legacy project and ~50% is covered so for now I want to filter
out a bit more aggressively. But at the same time don't get the false
impression that coverage is going up. Something like how :skip-covered
reports how many files it skipped, and improves signal-to-noise.

Now, it would be sort of OK for run to omit what report is omitting, but it
currently it overwrites the run omit with the report omit, which means it let's
through a bunch of files that shouldn't be covered.

Does this make it clearer?

[nedbat]

Fixed in 1031ead7ca4c (bb)

[nedbat]

Don't confuse run-include with report-include (and also omit). Fixes #621 and #622.

→ <<cset 1031ead7ca4c (bb)>>

[nedbat]

This was included in coverage.py v4.5, released today.

[nedbat]

Original comment by Max Nordlund (Bitbucket: maxnordlund, GitHub: maxnordlund)

---

Thank you