[Soot-list] Bytecodeoffset Tag probably broken in nightly builds

[Steven Arzt]

Hi Manas,

 

This happens because the new ASM frontend does not yet support BytecodeOffsetTag. If you really need it, you can switch back to coffi using –coffi command-line option. Sadly, it is not trivial to get bytecode offsets from ASM without patching the ASM class loader. The problem is that ASM never really gives access to such low-level information as it operates on a higher level of abstraction.

 

I have just looked into the problem and was able to extract the bytecode offsets for those locations that are ASM labels. The location of a label is not exactly the location of the next instruction, though. I have not yet fully understood why there is a tiny offset, but it exists. Furthermore, if there is no ASM label (i.e., we are in the middle of a basic block and there are no line number annotations), we are out of luck anyway. I’m not happy with that, but I don’t see a quick way to fix that at the moment.

 

@All: Some ideas?

 

Best regards,

  Steven

 

 

Von: soot-list-bounces at CS.McGill.CA [mailto:soot-list-bounces at CS.McGill.CA] Im Auftrag von Manas Thakur
Gesendet: Dienstag, 18. November 2014 13:49
An: soot-list at CS.McGill.CA
Betreff: [Soot-list] Bytecodeoffset Tag probably broken in nightly builds

 

Hi

 

I have been using the BytecodeOffsetTag in soot-2.5.0. I had updated my Soot version from a nightly build last week. When I checked today, the BytecodeOffsetTag was always null. I downloaded the latest nightly build again, and found that the problem is still there.

 

Can the developers please check if it’s really broken, or there is some new way introduced to access the BytecodeOffset of the units?

 

Regards,

Manas

 

 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.CS.McGill.CA/pipermail/soot-list/attachments/20141121/e1ffaa55/attachment.html