Why this change?

See previous

There is no need to change the established ordering: the change you've made just replaces r.getPath() with r.getFileName()

It's far more subtle then that.

Path has knowledge of things like .isFile and .isRegularFile and .isReadable.
String (from r.getFileName) does not, it's just a String.

The original test FileID.isClassFile(Path) doesn't only check that the last Path segment text filename is a *.class, but also checks if the provided Path is a file, not allowing directories with the name *.class to pass as true.

So, for example, the following change would need to be made at a minimum to maintain backward compatibility to the original code.

- if (FileID.isClassFile(r.getPath()))
+ if (Resoures.isReadableFile(r) && FileID.isClassFile(r.getFileName()))

That would need to occur in at least two places.
Once you see that code, you'll quickly understand the order change.

Why the change in order?

To make the conditions easier.

Keeping the order it would have been ...

        if (!r.isDirectory() && FileID.isJavaArchive(r.getFileName()))
        {
            parseJar(handlers, r);
            return;
        }

        if (r.isDirectory())
        {
            parseDir(handlers, r);
            return;
        }

        if (!r.isDirectory() && FileID.isClassFile(r.getFileName()))
        {
            parseClass(handlers, null, r.getPath());
        }

Which reads really weird.
So, it was changed to be anything directory related do this, anything else (a file), do that.

Well, the point of this PR is just to add null protection, so I don't see how a change in the order of the way we consider what to do with a given resource is related. I don't like changing the well-established order of things unless there is an associated bug.

The NPE protection is just part of the change.
The change also moves away from Path to Resource, so that the example of URLResource has an opportunity to work.
That change means we also change this section.
The new parseClass() that uses Resource is another example of this change.
Let me change the title of this PR, the description still has the correct details.

There is no need to change the established ordering: the change you've made just replaces r.getPath() with r.getFileName()