Understanding apache-ivy 2.4.0, I found out that the source-tarball code in the published one and the code in the repository are different. Why is the code not the same? Why is this done for the Java project? Perhaps this is some general pattern or practice - tell me.
For a C ++ project, I could come up with a reason. For example, it is possible to pack in a source-tarball code prepared for platform-specific copying, and pre-generate / compile everything that is platform-independent. But why is it necessary for Java?
Links to the repository and the source-tarball apache-ivy 2.4.0 are here: https://ant.apache.org/ivy/download.cgi
Ps. (give two different links and create an apache-ivy tag I don't have enough reputation)