[Soot-list] Empty catch range for exception handler

[Don Manuel]

Hi Alexandre,

We were looking into this thing and did not want to engage you guys with out finishing our checks.

I could see that the error was because we used Dex2Jar instead of Dexpler!!
Yesterday we tried out dexpler and did not get the error!! So all is back to normal and good!!
 


On Mar 18, 2013, at 7:31 AM, Alexandre Bartel <alexandre.bartel at uni.lu> wrote:

> Hi Don Manuel,
> 
> Thank you for reporting this.
> 
> I am not able to reproduce the bug using the information you provide
> (five lines of Java source code).
> 
> How do you generate the Android application?
> Do you instrument Java bytecode?
> Do you instrument Dalvik bytecode?
> 
> Can you send the source code and the .apk file of one of your Android
> applications which raises the exception?
> 
> Thanks,
> /Alexandre
> 
> Le vendredi 15 mars 2013 à 10:06 -0400, Don Manuel a écrit :
>> Hello,
>> 
>> 
>> I am getting the below error when I try to run some android apks with
>> soot.
>> Exception in thread "main" java.lang.RuntimeException: Empty catch
>> range for exception handler
>> at soot.coffi.CFG.jimplify(CFG.java:1049)
>> at soot.coffi.CFG.jimplify(CFG.java:955)
>> at soot.coffi.CoffiMethodSource.getBody(CoffiMethodSource.java:100)
>> 
>> 
>> 
>> 
>> I could see that the error occurs when I put a string assignment like
>> below in the android code
>> try { 
>> String name = "testname";
>> }
>> catch (Exception e) {
>> }
>> However the code will not error out if it were an integer assignment
>> instead of the string assignment shown above.
>> 
>> 
>> 
>> 
>> The error is caused by the soot code in CFG.jimplify 
>> if(startIns == endIns)
>>    throw new RuntimeException ("Empty catch range for exception
>> handler");
>> 
>> 
>> 
>> 
>> I also tried with the latest nightly, and got the same error. 
>> 
>> 
>> Can anyone give me a helping hand? We are in the final stages of our
>> analysis, and cannot move further without solving this error as 25% of
>> the applications we analyze are erring out due to this and that is a
>> significant failure rate which we cannot ignore.
>> 
>> 
>> 
>> 
>> Thanks in advance,
>> Don Manuel
>> _______________________________________________
>> Soot-list mailing list
>> Soot-list at sable.mcgill.ca
>> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
> 
>