Mypy linting errors should not be displayed if they are located in other Python files

Please note that this issue is based on a previous issue for the same problem: #9496 I’ve tried each version of the Python extension that was released after this original issue was fixed, and I see the error in each version.

Environment data

• VS Code version: 1.41.1
• Extension version (available under the Extensions sidebar): 2020.2.63072
• OS and version: Arch Linux
• Python version (& distribution if applicable, e.g. Anaconda): 3.8.1
• Type of virtual environment used (N/A | venv | virtualenv | conda | …): venv
• Relevant/affected Python packages and their versions: mypy
• Jedi or Language Server? (i.e. what is "python.jediEnabled" set to; more info #3977): jedi
• Value of the python.languageServer setting: Microsoft

Expected behaviour

The Python code editor only highlights errors in the file in question.

Actual behaviour

The Python code editor wrongly highlights unrelated errors found in other files that Mypy checked.

Steps to reproduce:

See #9496

I am testing with mixology, which can be cloned from here: https://github.com/sdispater/mixology I view the files mixology/partial_solution.py, mixology/version_solver.py, and mixology/package_source.py. There may be more.

Logs

##########Linting Output - mypy##########
mixology/package.py:17:5: error: Type signature has too few arguments
Found 1 error in 1 file (checked 1 source file)

[AWhetter]

Hi @kimadeline. Thanks for the response. Yes there is an error in the code. However the error is being reported in every file that imports it, not just in the file with the error.

To clarify my original issue, there’s an error in mixology/package.py, as you mentioned, on line 17. When viewing mixology/partial_solution.py, mixology/version_solver.py, or mixology/package_source.py VSCode reports that same error on line 17 of each of those files, even though the error is in mixology/package.py.

To outline the example in #9496, an error is added to midi.py, but VSCode reports that error in redblue.py and circuits.py as well, on the line that has the same line number as the error in midi.py.

[kimadeline]

We mark an issue as a regression if it gets introduced in a new version of the extension, and wasn’t present in earlier releases.

Since this issue was not introduced at the same time as #9496, but rather was present for a long time, it was marked as preexisting.

It sounds like using --follow-imports=silent might fix 1, however we’ll still need to take care of 2 (from the comment above), is that right?

1. When parsing the output of running mypy on a file, we should exclude errors reported on other files:
2. Once step 1 is addressed, run mypy with filepaths relative to the workspace directory

In any case, this is not a high-priority issue at the moment, so you are more than welcome to give it a try.