Reproduction Steps

1. Create a unversioned S3 bucket and add a file ( foo.txt )
2. Create a pipeline that uses a s3 resource pointing to foo.txt
3. Switch the S3 bucket to versioned and upload another file ( bar.txt )
4. Modify the existing s3 resource pointing it to bar.txt
5. At this point foo.txt has no version id and bar.txt does

Actual

Running the pipeline fails with the following error:

Expected

Running the pipeline should not fail. Pushing the same configuration to a brand new pipeline works.