back2source: Refine ScanCode.io d2d pipeline for Java

[pombredanne]

We need to refine the the Java d2d pipeline to support more file types and more languages.

This could include first to support for common JVM languages like Kotlin and Scala

• Later this can be improved as outlined in JVM d2d: Improve d2d pipeline to support generated code and more #1900

[pombredanne]

This PR is a step in that direction:

• Prefetch packages when using resource/filter_by_checksums purldb#538

[pombredanne]

The first step would be to collect a few example with these:

• kotlin: .kt files: See back2source: Add d2d mapping for Koltin #1806
• scala: .scala files: See back2source: Add d2d mapping for Scala mapping #1805
• For JAXB and other less common cases, see: JVM d2d: Improve d2d pipeline to support generated code and more #1900

[TG1999]

java_package_re = re.compile(r"^\spackage\s+([\w.]+)\s;?")

Try this for Kotlin, Scala and Java

[pombredanne]

There are related issues

• back2source: Add d2d mapping for Scala mapping #1805
• back2source: Add d2d mapping for Koltin #1806
• JVM d2d: .class file not mapped if .java also exists on deploy side #1807
• JVM d2d: Java cases where .class map with .g/.g4 #1808