[Soot-list] Using soot within a javaagent

[Prashant Deva]

Hi Eric,
 Yes, I would very much like to work with you on restructuring soot to
allow it to work as a javaagent.

As for resolution of dependencies, I dont require it and i guess anyone
using it as a java agent wont either.

I personally am trying to get only intra procedural alias analysis
information.
If i am correct, this would require converting the bytecode to jimple and
for each method separately getting the 'points-to' data.

Any guidance on where to start restructing the codebase to achieve the
required and get rid of dependency resolution would be helpful and as per
the terms of lgpl i would be happy to share my results with you.



Prashant


On Thu, Dec 29, 2011 at 6:26 AM, Eric Bodden <eric.bodden at ec-spride.de>wrote:

> Hello Prashant.
>
> On 28 December 2011 00:33, Prashant Deva <prashant.deva at gmail.com> wrote:
> > All the examples/tutorials I see of soot show it being used as a command
> > line tool.
> >
> > Is it possible to use it as part of a javaagent to analyse classes
> > individually as they are loaded?
>
> That's an interesting question. Indeed this would be very helpful.
> Soot is not currently set up to do this and it would require some
> restructuring to make this work. The problem ist that when Soot loads
> a class, transforming it into Jimple, at the same time it tries to
> resolve (and therefore load) referenced classes up to a certain
> amount. This is somewhat different from the class-loading process as
> it happens in the JVM. I am reasonably sure Soot could be restructured
> to make this work, and it would indeed be great to see this work, but
> I have never tried, nor do I know of anyone else who has.
>
> > In this case the only input available to soot would be the byte array of
> the
> > class being loaded at the moment.
> > Not that since the class would not have been loaded at this point, if
> soot
> > tries to use .class methods on the class being analyzed, those will
> throw an
> > exception.
>
> That's not an issue. Soot does not really need anything else but the
> byte array. The problem is rather in the resolution of dependencies,
> as mentioned above. If you were to give this a try I would be happy to
> assist you!
>
> Best wishes,
> Eric
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.mcgill.ca/pipermail/soot-list/attachments/20111229/468b714c/attachment.html