[llvm-commits] [llvm] r128333 - /llvm/trunk/lib/Transforms/Scalar/DeadStoreElimination.cpp

[Frits van Bommel]

On Sat, Mar 26, 2011 at 4:39 PM, Dan Gohman <gohman at apple.com> wrote:
> On Mar 26, 2011, at 2:32 AM, Bill Wendling wrote:
>> Simplification noticed by Frits.
>>
>> Modified:
>>    llvm/trunk/lib/Transforms/Scalar/DeadStoreElimination.cpp

>> -  if ((EarlierOff == LaterOff && Earlier.Size <= Later.Size) ||
>> -      (EarlierOff > LaterOff &&
>> -       EarlierOff + Earlier.Size <= LaterOff + Later.Size))
>> +  if (EarlierOff >= LaterOff &&
>> +      EarlierOff + Earlier.Size <= LaterOff + Later.Size)
>>     return true;
>
> I don't have time to fully investigate, but the testcase passes without the fix,

What testcase? This commit doesn't have a testcase (and doesn't need
one, given the next point).

> and the new code is now equivalent by De Morgan's law to the old code.

Hint: the commit message called it a "simplification" because it *is*
equivalent (but simpler) code.

I'm not sure where you're getting De Morgan's law from though, since
there's no negation anywhere...
(The only De Morgan's laws I know about are these:
<http://en.wikipedia.org/wiki/De_Morgan%27s_laws>)