[llvm-commits] [llvm] r52537 - /llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp

[Chris Lattner]

On Jun 24, 2008, at 11:02 AM, Matthijs Kooijman wrote:

>> Out of curiosity, why not?  {x} -> x would eliminate extractvalue
>> instructions, so it seems that it would reduce the size of the IR.
> Because that would also perform this transformation on external  
> functions and
> other functions that are intrinsically alive (Since we don't store  
> that they
> are intrinsically alive, only that all of their return values and  
> arguments
> are).

Ah, that would be bad.

> On second thought, however, it might be better to store intrinsically
> alive functions as such, which should also reduce the size of the  
> Uses map and
> reduce lookup times.

Makes sense!

>> typo: chaning -> changing
> I think Dan pointed this out somewhere as well, I didn't get around  
> to fixing
> those yet :-)

ok :)

-Chris