Re: Points-To-Analysis and BCEL (Byte Code Engineering Library)

[Navindra Umanee <navindra@cs.mcgill.ca>]

Eric Bodden <eric@bodden.de> wrote:
> The problem however I see here is that the representation generated by SOOT
> and the BCEL one are entirely different. Thus one would have to write some
> kind of adapter, which is generally doable but would be a hard job since one
> would have to write such adaptors for several classes. IMHO having a look at
> how the PTA is implemented in SOOT and reimplementing that on BCEL would be
> an easier and more suitable approach.

I have no knowledge of BCEL, but it seems to me that doing the
instrumentation in Soot itself would be easier than implementing PTA
in BCEL?

Also note that you could probably save the results of your PTA in Soot
in tags/attributes and read this in from BCEL (assuming that's easily
doable).

Cheers,
Navin.