Quick comments; will look into more details later. Gradle wrapper jar >
I'm of the opinion that this is fine, but others may disagree. Other projects do it. It is a standard Gradle approach. I suggest following up with Justin, and understanding what his concerns may be. > the existence of PySpark in the resources of the executor/test folder The redistribution of this file is fine. I thought that Justin found that the header is missing. If so, just the header should be added, and that's it. 3. Justin had a strange remark about an MIT licenced file which I think > refers to codegen.py which is a BSD licensed file, I realize I need to add > the original license of the project to the licenses folder, however, > codegen doesn't have one. what can we do in that case? We cannot distribute components with unknown license. If it is MIT/BSD, there's a standard approach we use. If it is public domain, there are typically options as well. I suggest understanding the genesis of software in question.
