[Soot-list] Polyglot DataFlow

[mbatch at cs.mcgill.ca]

Hello,

I am trying to write a simple data flow analysis on the exact java source
that gets read into Soot (before any transformations) for the purpose of
generating source code readability metrics.

I am trying to use Polyglot's DataFlow using the polyglot AST created
before any transformations happen but unfortunately, it is very unclear
how I can run this analysis, given a root AST node. Specifically, I need a
TypeSystem and Resolver to create the DataFlow extended class, which I do
not have at that point in the compilation.

The other option I could consider is an analysis in the jj pack, before
any other transformations happen. It is imperative that my analysis runs
on the un-changed source AST. Is it possible to add an analysis
(transformer) to the jj pack and have it be the _FIRST_ tranform that is
run?

Thanks,

Michael