[llvm] r182832 - Temporary fix to get rid of gcc warning.

[Rafael Espíndola]

> Generally we prefer to simply disable or fix bad warnings rather than
> workaround them in code. (which usually amounts to disabling GCC warnings
> and fixing Clang warnings)
>
> Which warning was the problem here & can we just disable it? Does Clang have
> an equivalent (& better/correct/not-false-positive-on-this-code) warning?

I think it was a valid warning about the comparison always being true
and if remember the email correctly it actually found a bug on the
enum values they use. Vincent?

Cheers,
Rafael