[clang] [clang][bytecode] Check for integral types in evaluateStrlen (PR #185481)

[Oliver Hunt via cfe-commits]

ojhunt wrote:

> > The fact that this PR is necessary seems to point to the current name resulting in it being used incorrectly.
> 
> How does that make sense? This PR is necessary because there was _no_ such check before, not because there was the wrong check.
> 
> Let me check where fixed point types are needed in `isIntegralType()`.

Sorry I think that statement was due to the addition of `isRealIntegralType()` coupled with the difference from `isIntegralType()` elsewhere making me completely forget what I was actually commenting (basically I fixated on the new function, rather than the actual intent of the patch).

https://github.com/llvm/llvm-project/pull/185481