How is multidex support implemented?

[Kristopher Micinski]

Hey all,

I'm currently working on a project where we have to do some binary rewriting of Android apps for research purposes. Our binary rewriter works for Android apps written in the old (standard) style, and pulls data form classes.dex, and rewrites it. But I'd like some clarification on multidex implementation and dex file layout using multidex.

In a standard dex layout, you have all of the class, method, type, etc... definitions in different tables. Things that are dynamically linked (such as those from the Android framework) simply have their method prototypes included, but no code data. Is it true that in a multidex setup, each classesN.dex contains a set of class implementations, and methods that are implemented in other dex files are merely included in the same way as dynamically linked calls? I.e., they have their prototypes included, but no implementation, and are implemented in other dex files?

Thanks!

Kris