java-dev

Christian Roth wrote:

> I was using ProGuard 3.4 for a long time and never had any issues with
> our application. After installing the latest Java update, the jar
> produced by ProGuard started to throw an ArrayIndexOutOfBounds  
> exception
> (when the unprocessed one does not).


You should look very closely at the exception information (i.e. the  
reported bad index and array), and at the bytecodes where the  
exception occurred.  If nothing else gives you bytecodes, the 'javap'  
command will.

Like Ulli Kortenkamp, I have seen occasional problems with ProGuard  
optimizations.  In one case, the bytecode sequence appeared valid,  
though they were obviously altered from the normal compiler output in  
a direct comparison.  A particular sequence was apparently triggering  
a JITC problem.  I investigated this at the time by running the code  
with the -Xint JVM option to force always-interpreted mode.  It ran  
at ~10% of JITC'ed speed, but it worked in that mode.  Since we  
couldn't afford to run at 10% of usual speed, and since we hadn't  
seen any speedups we could directly attribute to ProGuard's  
optimizations, we decided to omit ProGuard's optimizations.

  -- GG

_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list      ([EMAIL PROTECTED])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/alexiscircle%40gmail.com

This email sent to [EMAIL PROTECTED]