Skip to content

Flake8 plugins issues #140

Closed
Closed
@ivan94fi

Description

@ivan94fi

Hi, currently there are two options for flake8 functionality:

  • the thid party plugin, pyls-flake8, which is recommended in the readme
  • the internal plugin, in pylsp/plugins/flake8_lint.py, which is not advertised in the documentation

I would like to report some issues with these plugins.

First issue: two functionally equivalent plugins (mostly)

The first issue I would like to point out is that the two plugins are equivalent: the internal one should be preferred, and the suggestion for the third party one removed.

Actually the pip version of the external plugin is buggy as it does not respect the boolean options passed to flake8's executable. I submitted a pull request to the repository to fix this issue and it was merged, but it is not published on pypi yet. Also, some functionality is missing with respect to the inner plugin, and overall it seems a bit more crude.

Second issue: flake8 support not explicitly stated in readme

While the "Configuration" section in the readme explains how to use configuration from flake8, the "Installation" section does not mention flake8 in the "list of optional providers that will be available (if the respective dependencies are found)". I think it should be explicitly stated, as at first it is not clear.

Third issue: flake8 vs pyflakes + pycodestyle + mccabe

As flake8 is a composition of pyflakes, pycodestyle and mccabe (plus various other flake8 plugins, if installed) I think it should be the default to disable the individual components if flake8 is run, as to avoid duplicated runtimes and reports. A simpler solution might be to advertise flake8 in the readme, as per the previous issue, and explicitly state something along the lines of "Flake8: linter to detect various errors (composition of pyflakes, pycodestyle and mccabe: use a combination of the components or flake8 alone to avoid duplicated error reports)."

This is an example of duplicated messages from pylsp, using atom client:

Click to view image

double_reports_pylsp

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions