Hi,
> What exactly is happening here, can you share more details on how to avoid this.
The instructions are being reorganized now that new code has been inserted by AspectJ. Destinations for branch targets that moved (probably due to the insertion of advice) are being recomputed. If the jump destination is a long way away for some instructions it is possible to switch to a 'wide' version of the instruction to encode the 'further' destination (so use 4bytes instead of 2bytes). If the instruction was a GOTO we could use a GOTO_W instead. In your case it isn't a goto, it is IFLE (if less than or equal) and we can't fit the expected destination into the space we have available to encode the target.
The 'fix' in AspectJ is to dynamically adjust the IFLE to a nearby branch destination and then use a GOTO_W to get to the real target, but we haven't had time to implement that. Basically you have a big method and you are weaving stuff into it and just pushing the branches too far apart. Do you need to weave so much into the method in question? Does any extra exception info tell you the method in question? I can't recall.
cheers,
Andy