r263048 - ARM & AArch64: convert asm tests to LLVM IR and restrict optimizations.

[Eric Christopher via cfe-commits]

On Wed, Mar 9, 2016 at 1:32 PM Tim Northover via cfe-commits <
cfe-commits at lists.llvm.org> wrote:

> On 9 March 2016 at 13:20, David Blaikie via cfe-commits
> <cfe-commits at lists.llvm.org> wrote:
> > If we are touching these tests at all, as a cleanup, why aren't we
> adjusting
> > them to not pass through the llvm optimizers/code generators at all?
> > Generally we are pretty clear that any test in Clang looking at llvms
> > generated assembly is an anachronism to be removed eventually
>
> This patch should remove all checking (and generation) of assembly.
> But as I said, I think banning all optimizations is going too far and
> leads to unreadable tests (with excess loads and stores being a
> particular problem).
>
>
I still think we should avoid the optimizer, but as (at least) an
intermediate step this seemed like a good compromise.

-eric
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20160309/db8e15e8/attachment.html>