[llvm-commits] [PATCH] Fix undefined behavior in APInt

[Chris Lattner]

On Feb 24, 2012, at 2:33 AM, Ahmed Charles wrote:

> APInt currently executes UB in the lshr function which is intended to
> result in a 0 value when shifted by more than the number of total
> bits, but this isn't done in a conforming way. It seems like the
> original code had this in mind, but got the check wrong.

Makes sense, please do.

-Chris