[abc] POPL submission / Soot and abc release

Hi all.

As mentioned earlier, Patrick, Laurie and I wrote a POPL submission on
flow-sensitive optimizations for tracematches. This basically replaces
"stage 3" of the ECOOP paper with something that actually works. We
got pretty good speedups and found out other interesting stuff so it
might be worth a look.

You can download it here:
http://www.sable.mcgill.ca/~ebodde/pubs/blh07flowsensitive-draft.pdf

A TR version will follow shortly.

Further, I would like to prepare a joint Soot and abc release very
soon. The Soot release because of this bug we fixed with the points-to
analysis and the abc release so that at least the ECOOP analysis is
out there when I give the talk. (All stuff relevant to running the
ECOOP analysis is already in the abc HEAD. The POPL analysis I
currently keep in our SVN at McGill.)

However, the analysis for POPL was still impeded in some cases by an
apparent bug in Manu's context-sensitive points-to analysis. (It fails
to produce context information in some even smaller test cases.) I
will see if Manu and I can get this fixed by next week or so, so that
then we can hopefully do a release.

Eric

-- 
Eric Bodden
Sable Research Group
McGill University, Montréal, Canada