[Soot-list] Possible to use SootClass object outside of transform() ?

[Joe Roback]

that was our thought also. if an soltution doesn't present itself
rather quickly, we will probably go the thread route. thanks again
eric.

Joe

On Wed, Nov 26, 2008 at 2:44 PM, Eric Bodden <eric.bodden at mail.mcgill.ca> wrote:
> From your description, I think that it would be easiest to prevent
> Soot from "finishing" in the first place, i.e. bring it to block just
> after all your transformations have run. Then, when you need to
> perform additional work with Soot you awaken its blocked thread and
> have it perform the transformations (e.g. obfuscations) you want to
> do.
>
> Eric
>
> 2008/11/26 Joe Roback <robackja at cs.arizona.edu>:
>> Hi Eric.
>>
>> How about a 2 second overview of the project. Client/Server. Client
>> asks Server for program blocks to run an application. Server provides
>> them. Server is continuously obfuscating the blocks as it sends them
>> to the Client. That is a high-high-level overview, skipping over lots
>> of details.
>>
>> And we do run some code within the control flow of soot main method to
>> start with. But its the continuous replacement part, where we want to
>> continuously obfuscate the generated program blocks. Program flow
>> looks like this
>>
>> Server starts
>>  - Run SOOT over program, spliting program into blocks using custom
>> soot tranformer
>>  - At this point, soot.Main() has finished.
>>  - Opens Socket and between serving Blocks to Client
>>
>> The idea here is each time the same block is served to the client, its
>> changed using some obfuscation technique. Re-running Soot each time is
>> not feasible, since it takes 2 minutes+ to run even for a simple
>> program on a fast machine (Soot is generating the call graph to map
>> dependencies between program blocks -- old code, stuff I didn't
>> personally write)
>>
>> I keep looking into it. :) I was just really looking for a "maybe" or
>> "absolutely not" feasible.
>>
>> thanks for quick reply eric.
>>
>> Joe
>>
>> On Wed, Nov 26, 2008 at 2:14 PM, Eric Bodden <eric.bodden at mail.mcgill.ca> wrote:
>>> Hi Joe.
>>>
>>> I have never tried to add functionality after main(..) has completed
>>> running. I am sure it can be done but it could be painful. What
>>> exactly are you trying to achieve? Why can't you run your code from
>>> within the control flow of Soot's main method?
>>>
>>> Eric
>>>
>>> 2008/11/26 Joe Roback <robackja at cs.arizona.edu>:
>>>> Hello All.
>>>>
>>>> I am working on a continuous-replacement project with Christian
>>>> Collberg here at the University of Arizona. In the project we need to
>>>> keep SootClass objects around and manipulate them after-the-fact (ie
>>>> after soot.Main is run). Is this feasible?
>>>>
>>>> if I try to even write out a SootClass using JasminClass outside the
>>>> transform() I get errors of "no active body". Backtrace looks like
>>>> this:
>>>>
>>>> java.lang.RuntimeException: method: <init> has no active body!
>>>>        at soot.AbstractJasminClass.emitMethod(AbstractJasminClass.java:687)
>>>>        at soot.AbstractJasminClass.<init>(AbstractJasminClass.java:568)
>>>>        at soot.jimple.JasminClass.<init>(JasminClass.java:62)
>>>>        at server.blockify.SootOutput.writeByteArray(SootOutput.java:74)
>>>>        at server.blockify.BlockifiedBlock.getClassBytes(BlockifiedBlock.java:68)
>>>>        at server.blockify.BlockifiedBag.getBlock(BlockifiedBag.java:47)
>>>>        at server.components.Scheduler_c.processClientRequest(Scheduler_c.java:98)
>>>>        at server.components.BlockProvider_c.run(BlockProvider_c.java:37)
>>>>        at server.components.ServerInstance_c.startServerInstance(ServerInstance_c.java:78)
>>>>        at server.control.ServerApplication_c.run(ServerApplication_c.java:115)
>>>>        at java.lang.Thread.run(Thread.java:613)
>>>>
>>>> So, is this feasible thing to do with SOOT? or must all SOOT related
>>>> manipulations happen within the SOOT tranformations?
>>>>
>>>> thanks. much appreciate the help.
>>>>
>>>> Joe Roback
>>>> _______________________________________________
>>>> Soot-list mailing list
>>>> Soot-list at sable.mcgill.ca
>>>> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
>>>>
>>>
>>>
>>>
>>> --
>>> Eric Bodden
>>> Sable Research Group, McGill University, Montréal, Canada
>>> Got an interesting job offer? http://www.bodden.de/hire-me/
>>>
>>>
>>
>
>
>
> --
> Eric Bodden
> Sable Research Group, McGill University, Montréal, Canada
> Got an interesting job offer? http://www.bodden.de/hire-me/
>
>