I'd start with running a linter on CI but not failing on errors. From there, individual issues / PRs can be created and addressed with a somewhat narrow-focus (example: one issue fixes E111 indentation not a multiple of 4, etc.)
Once we get the "good enough" results (up to discussion?), then maybe linter failures could prevent PR merges.
See #11 from https://github.com/cocagne/txdbus/issues/11#issuecomment-306205172 onwards.
I'd start with running a linter on CI but not failing on errors. From there, individual issues / PRs can be created and addressed with a somewhat narrow-focus (example: one issue fixes E111 indentation not a multiple of 4, etc.)
Once we get the "good enough" results (up to discussion?), then maybe linter failures could prevent PR merges.
Thoughts welcome.