Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.52%. Comparing base (fef1c72) to head (22b6fcb).
⚠️ Report is 2 commits behind head on dev.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

I can help add some context of why this PR is useful. In nf-core/smrnaseq we had many issues with warnings when linting, as showed here: https://github.com/nf-core/smrnaseq/actions/runs/20112360084/job/57712837018. This fixes two edge-cases in the subworkflow linter that can generate misleading “Included component not used in main.nf” warnings.

Basically the message would say that almost all components of a subworkflow, e.g. INDEX_RRNA were not included in the main file, while they were there.

The linting would break after finding the word 'emit' (which was written as a comment in that subworkflow). This PR fixes that issue :)

Nice, thanks for the contribution.

Do you have time to also add test case to tests/modules/lint/test_main_nf.py where the old linting failed and your solution works? I can have a look otherwise

Thanks @mashehu!
I added two test cases to tests/subworkflows/test_lint.py. I haven't used Pytest before, so it's probably far from perfect. Can you please review it? Feel free to make any changes.

Hi! As @mashehu mentioned, we agreed that one subworkflow would have only one workflow. I think this should be the standard, and we should update the utils subworkflow in the template instead. However, taking care of comments containing the word "emit" is a good fix, I would instead skip all comments from the parsing, so lines that start with //.

Hi @mirpedrol, thanks for the review!
I have removed the "test_file_with_multiple_workflows" test, since it won't be required after the template changes.

As per skipping all comments from the parsing, that's kind of how it works. The problem was that previously, the checking of whether the line is a comment or not, happened after searching for the keywords (emit, take, etc.) and switching the section. After this fix, the check will happen before, so it will not switch sections based on commented lines.