Manu Sridharan

> @msridhar I made some attempts. It is still a bit hacky, but want to hear your opinions first. It's a good start! I added a couple initial comments. >...

Ok thanks @liblit. I'm not sure what is the best way to proceed here. When I revamp documentation a bit more I will think more about it.

Sorry for my slow response. This looks like a clear bug to me. Any chance you could submit a PR with a fix?

@RalfHerzog what is the scope of this problem? Is Shrike’s bytecode instrumentation broken in cases where the input bytecodes have `pop2` and `dup2`? Is the WALA SSA IR affected? Thanks...

Thanks again for digging into this! Unfortunately I do not have deep expertise in the implementation of Shrike. Further, I don't think we have a lot of users making use...

@RalfHerzog thanks for digging further into this. As I mentioned before I’m not an expert on these issues. Maybe you could look at the ASM code to see how they...

@RalfHerzog How often do these bytecodes show up? If they are rare, I would be open to an implementation approach that is less performant. But I'm guessing they are not...

@RalfHerzog we really appreciate all the work you've put in on this issue! Those failing test cases correspond to languages other than Java getting compiled to JVML. So there may...

@juliandolby with this version bump this test is failing: https://github.com/wala/WALA/blob/7790a8c4453a4035390d30bd4aa7c1a226cbd5c6/com.ibm.wala.cast.js.rhino.test/harness-src/com/ibm/wala/cast/js/test/TestSimplePageCallGraphShapeRhinoJericho.java#L33-L39 With Jericho 3.2 this test would pass since the parse errors for the garbage HTML would get ignored. With Jericho...

/cc @MadhuNimmo