Patch to BitcodeReader

[Tobias Grosser]

On 09/12/2013 10:50 PM, Tom Roeder wrote:
> Yes, that triggers it. Here's an updated patch that uses the test you
> suggest and uses FileCheck to make sure that the right IR is generated from
> llvm-dis. I'm new to the LLVM testing infrastructure, so please tell me if
> there's a better way to do this check.

It is probably enough to CHECK: just for the line that contains the select.

Also, please add a short comment to the test case (or call the function 
accordingly) such that it is clear we test the CS select and not any select.

With those changes applied, this patch seems good to go.

Thank you,
Tobias