r254251 - ARM v8.1a adds Advanced SIMD instructions for Rounding Double Multiply

[Renato Golin via cfe-commits]

On 1 December 2015 at 19:45, David Blaikie <dblaikie at gmail.com> wrote:
> These cases I don't understand - if they're features LLVM supports we'd just
> test them straight up. If you run our test suite against another compiler it
> should tell you if it's compatible with our compiler (does it offer the same
> functionality), so we don't have to maintain a distinction between
> functionality we support and other compilers support, and functionality we
> support and no other compiler support.

I see what you mean. If the other compiler is not "compatible" with
the results we expect from Clang, the tests should fail, and that's
the *expected* result.

This is one way of looking at it. I personally prefer the current
state, where the test-suite is a black-box and the code is deemed
correct if the output is equals the expected value, because that can
be achieved with any compiler, and should solely rely on the source
code and the compiler following "the standard" (whatever it is). I
know it's not that simple, but having a simple premise makes it easier
to classify "bugs in the test" from "bugs in the compiler". :)

cheers,
--renato