[Soot-list] How to set -process-dir and -cp correctly in this case? I got the error: class small.DataClass read in from a classfile in which DataClass was expected.

[Khanh Nguyen]

Sorry for my ignorance but where exactly the -x option is processed. I mean
if -x option is processed before you running any pack then the excluded
classes should not be in the list of application class. That's what -x
does, right?

Also, PackManager.runPacksForOneClassAtATime() only called if
Options.v().oaat() is true. Not sure what oaat() is but probably in
runPacks(), PackManager should take the else branch where it will take the
whole program packs and run them

Thank you,

Khanh Nguyen


On Fri, Nov 22, 2013 at 3:12 PM, Guru Devanla <gdevan2 at uic.edu> wrote:

> Hi Khanh,
>
> I looked at the current version of Soot. I think the problem is in
>
> PackManager.runPacksForOneClassAtATime().  This methods loops through all
> files it files under -process-dir and *does not check/respect*
> attributes set with -x option.
>
> The looping structure in that method can have a conditional check to see
> if the class being processed is part of the excluded package.
>
> Eric et all: So, is it okay to exclude classes at this method level. Does
> anyone see any incompatibility issues with excluded class under process-dir
> not being part of application class list?
>
> Thanks
> Guru
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> On Fri, Nov 22, 2013 at 1:20 PM, Khanh Nguyen <khanhtn1 at uci.edu> wrote:
>
>> Even with -no-bodies-for-excluded I still see my transformation touches
>> the excluded packages
>> My command is this:  -w -pp -keep-line-number -f J -p jb
>> use-original-names:false  -process-dir C:\test\bin -exclude small
>> -no-bodies-for-excluded -cp C:\test\bin
>> Anything that is unusual?
>>
>> Thank you,
>>
>> Khanh Nguyen
>>
>>
>> On Fri, Nov 22, 2013 at 12:12 PM, Guru Devanla <gdevan2 at uic.edu> wrote:
>>
>>> I think you also need to use -no-bodies-for-excluded.  I have seen
>>> excluded packages being touched for different analysis when I just use the
>>> -x option.
>>>
>>> Hope that helps.
>>>
>>> -Guru
>>>
>>>
>>> On Fri, Nov 22, 2013 at 10:38 AM, Khanh Nguyen <khanhtn1 at uci.edu> wrote:
>>>
>>>> Hi Eric,
>>>>
>>>> Unfortunately, that doesnt work for me. I set the "-x small" then
>>>> use Scene.v().getApplicationClasses() to get the application classes to
>>>> check. And there are classes from package "small" being analyzed.
>>>> I'm using the released soot 2.5. My analysis is in pack "wjtp" if that
>>>> info helps.
>>>>
>>>>
>>>>
>>>> Thank you,
>>>>
>>>> Khanh Nguyen
>>>>
>>>>
>>>> On Fri, Nov 22, 2013 at 12:20 AM, Bodden, Eric <
>>>> eric.bodden at sit.fraunhofer.de> wrote:
>>>>
>>>>> Hi Khanh.
>>>>>
>>>>> When you set Soot's classpath or process dir then the path must always
>>>>> point to the root of the package hierarchy, not to some particular package
>>>>> folder, i.e., in your setting it must point to C:\test\bin\
>>>>> You can then the include/exclude options to include/exclude certain
>>>>> packages from analysis. See here:
>>>>> http://www.sable.mcgill.ca/soot/tutorial/usage/index.html
>>>>>
>>>>> Does that work for you?
>>>>>
>>>>>
>>>>> Cheers,
>>>>> Eric
>>>>>
>>>>>
>>>>>
>>>>> On 18.11.2013, at 08:47, Khanh Nguyen <khanhtn1 at uci.edu> wrote:
>>>>>
>>>>> > Hello,
>>>>> >
>>>>> > I know that this question has been in the list since a long time
>>>>> ago, I read all those posts and got what I wanted to do but still, I want
>>>>> it work in another way (if possible)
>>>>> > The background is this: In C:\test\bin I have 2 folders, "simple"
>>>>> and "small" (equivalent to 2 different packages) in "small", I have one
>>>>> class: DataClass.class, in "simple", I also have one class called
>>>>> DataClass.class
>>>>> >
>>>>> > I want to apply my analysis only on folder "small". Thus I used the
>>>>> command line :-process-dir C:\test\bin\small -cp C:\test\bin\small.
>>>>> > I understand that by setting -process-dir C:\test\bin\small, the
>>>>> class Soot will find to load is DataClass.class (i.e. no info about the
>>>>> package 'small') which is okay, that's why I thought the -cp must be set as
>>>>> C:\test\bin\small. So now the file structure is identical. But I got the
>>>>> error instead("class small.DataClass read in from a classfile in which
>>>>> DataClass was expected.")
>>>>> > Is it because even when setting -cp: C:\test\bin\small the class
>>>>> loaded is still be considered as small.DataClass? (because the actual class
>>>>> is read and thus, the name is small.DataClass?)
>>>>> >
>>>>> >
>>>>> > Of course I can make my analysis to work only on "small" by deleting
>>>>> the "simple" and set both -process-dir and -cp to C:\test\bin (so that
>>>>> -process-dir sets the target class is actually small.DataClass and -cp will
>>>>> let me see small\DataClass.class)
>>>>> > But, even with only "small" under C:\test\bin, I cant do
>>>>> -process-dir C:\test\bin -cp C:\test\bin\small, either (Soot will complain
>>>>> there is no small.DataClass on classpath- why? now -cp only see
>>>>> DataClass.class?)
>>>>> >
>>>>> > But, bottom line is, isnt there anyway to just set the target of the
>>>>> analysis restricted to only 1 package like what I wanted?
>>>>> >
>>>>> > Thank you,
>>>>> >
>>>>> > Khanh Nguyen
>>>>> > UC Irvine
>>>>> > _______________________________________________
>>>>> > Soot-list mailing list
>>>>> > Soot-list at sable.mcgill.ca
>>>>> > http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
>>>>>
>>>>> --
>>>>> Prof. Eric Bodden, Ph.D., http://sse.ec-spride.de/ http://bodden.de/
>>>>> Head of Secure Software Engineering  at Fraunhofer SIT, TU Darmstadt
>>>>> and EC SPRIDE
>>>>> Tel: +49 6151 16-75422    Fax: +49 6151 16-72051
>>>>> Room 3.2.14, Mornewegstr. 30, 64293 Darmstadt
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Soot-list mailing list
>>>> Soot-list at sable.mcgill.ca
>>>> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.mcgill.ca/pipermail/soot-list/attachments/20131122/aed09a8c/attachment-0001.html