changed-files: [BUG] Issue with `files_yaml`?

Is there an existing issue for this?

I have searched the existing issues

Does this issue exist in the latest version?

I’m using the latest release

Describe the bug?

It seems that changes in one subset are propagating to other subsets, e.g.:

  base:
    - Source/**
    - Lib/std/**
    - Lib/typemaps/**
    - Lib/xml/**
    - Lib/*
    - Examples/Makefile.in
    - Examples/test-suite/*
    - Examples/test-suite/errors/*
  python:
    - Lib/python/**
    - Examples/python/**
    - Examples/test-suite/python/**

It seems to propagate into _other_changed_files

  .github/outputs/base_all_modified_files.json:[]
  .github/outputs/base_any_changed.json:true
  .github/outputs/base_any_modified.json:true
  .github/outputs/base_other_changed_files.json:[".github/workflows/step_change-detection.yml",".github/workflows/step_get-variants.yml",".github/workflows/step_lang-test.yml",".github/workflows/step_nuget.yml",".github/workflows/step_toolchain-test.yml","Lib/python/file.i"]

Note that base_other_changed_files.json contains Lib/python/file.i, which is not supposed to be there from the glob expression of base

To Reproduce

https://github.com/LecrisUT/swig/actions/runs/6074609826/job/16479045195?pr=1

What OS are you seeing the problem on?

ubuntu-latest or ubuntu-22.04

Expected behavior?

.github/outputs/base_any_changed.json

false

Relevant log output

Truncated log above

Has all relevant logs been included?

I’ve included all relevant logs

Anything else?

No response

Code of Conduct

I agree to follow this project’s Code of Conduct

About this issue

Original URL
State: closed
Created 10 months ago
Comments: 18 (11 by maintainers)

Most upvoted comments

Tested the fix and it now works as expected. Thank you for the work on this

LecrisUT on Sep 4, 2023

But that is exactly what is confusing me. I indeed do NOT want Lib/python/file.i to be matched in base. It is meant to be matched only in python, that is why I am using Lib/* instead of Lib/**.

The issue is that Lib/python/file.i is included in base_other_modified_files (it should not be there because it is not part of the glob pattern), but more importantly base_any_modified should be false (but it is set to true probably because of base_other_modified_files; any_changed, any_deleted and any_modified are all set to true). Otherwise I do not see which variable can I reliably check for the glob that is given in base.

As a temporary workaround, checking base_all_modified_files should pick up all expected changes instead of base_any_changed?

I see this is a bug in that case since they should all be false

jackton1 on Sep 4, 2023