[LLVMbugs] instcombine fix

[Chris Lattner]

On Thu, 5 Dec 2002, Casey Carter wrote:

> I noticed that one of the reassociate regression tests --
> 2002-07-09-DominanceProblem -- was failing.  instcombine was transforming
...
> without reexamining %Y.neg.ra to see that it is clearly replacable by
> %A.  The problem is that instcombine's runOnFunction does not add uses
> of a constant-folded instruction into its worklist for examination.

That's a great catch!  Thanks for noticing the bug:
http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20021202/001841.html
http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20021202/001842.html

-Chris

-- 
http://llvm.cs.uiuc.edu/
http://www.nondot.org/~sabre/Projects/