Re: bug in generating variable names

[Patrick LAM <plam@sable.mcgill.ca>]

I sent Florin a mail with the new CommonSubexpressionEliminator.  It will
be in the next release; it's also available at

http://www.sable.mcgill.ca/~plam/CommonSubexpressionEliminator.java


On Fri, 8 Jun 2001, Florin Mateoc wrote:

> I would like to ask something related to my
> experiment: the jimple output files kept improving
> with every iteration. Do you have an idea why (I mean
> why didn't the first pass find these jimple-specific
> optimization opportunities)?

I'm quite surprised by your finding, especially because you're not using
whole-program optimization (-W).  Soot shouldn't be doing anything
nontrivial with just -O, and certainly not anything that doesn't show up
in the first pass.  Can you show me an example?

> And an unrelated question: is there a description
> somewhere of what was generated automatically with
> sablecc and what needed to be done manually on top of
> that ? Say if sablecc improves, how easy would it be
> to regenerate an "improved" soot ?

The only part of Soot which is generated from SableCC is the jimple
parser, which isn't even executed when you're compiling from
classfiles.  The jimple parser is in soot.jimple.parser, and you
regenerate by running sablecc on the jimple.scc file in the main
directory.

John Jorgensen did some profiling of Soot.  I'd be interested in reading
about his results, if he would post them somewhere.  (Hi John?)

pat