On 7/5/05, Leopold Toetsch <[EMAIL PROTECTED]> wrote: > > On Jul 5, 2005, at 8:50, Curtis Rawls wrote: > > > > > I seem to have come across a bug in the compute_dominators() algorithm > > with the following PIR: > > It's not unlikely that there are bugs in that part of imcc. It was > contributed by Angel Faus, who seems to be not participating in Parrot > dev anymore. Some times ago, I was pondering some tests for CFG-related > items, but no one (including me) seems to have the tuits to attack it. >
Yes, I'm running into a lot of areas of the CFG/optimizer that I would like to modify, and some CFG tests would be great to help avoid introducing new bugs. Basically just checking the cfg and dominator dump outputs for correctness. I have looked at the tests in t/ and imcc/t/, but have not found a way to test the compiler debug output. Any suggestions? Also, I am working on some new (and hopefully more correct) implementations of some of the CFG algorithms, such as compute_dominators(). I would appreciate having someone to run these by. Any takers? -Curtis
