[llvm-commits] [llvm] r151466 - in /llvm/trunk: include/llvm/Analysis/Dominators.h lib/Analysis/InstructionSimplify.cpp

[Rafael Espíndola]

> The most important thing for me is that dominance of blocks and dominance of instructions follow the same rules.  Beyond that, I'm pretty ok with anything.  If you think that returning true makes sense, feel free to make it so (just don't break anything ;-)

OK, so the definition of dominates(BBA, BBB) would become:

dominates(BBA, BBB) is true iff  any definition in BBA would dominate
any use in BBB. With the note that an invoke in BBA is not a
definition in BBA (it is on the normal edge) and a phi in BBB is not a
use in BBB (it is on the predecessors).

Do you agree with such a definition? I will try to implement that as
soon I am able to build and test deal.ii since it looks like there
might have been a regression from the last change.

> -Chris
>

Cheers,
Rafael