On Fri, Oct 14, 2016 at 11:48 AM, Bernd Schmidt <bschm...@redhat.com> wrote:
> On 10/14/2016 11:33 AM, Richard Biener wrote:
>> Ok, so at least for the GIMPLE FE side I thought it's useful to allow
>> a correctness verification with something simpler than pattern matching
>> on the pass output. By means of doing runtime verification of an expected
>> result (this necessarily includes running followup passes as we have to
>> generate code). I don't see why this shouldn't apply to __RTL -- it might
>> be more difficult to get __RTL testcases to the point where they emit
>> assembly of course.
> Also, if this ends up with a lot of testcases, I still think it would be
> best to have these use a special rtl-test port so that the tests don't
> become a maintenance burden for real ports, and so that we can construct the
> tests any way we like.
> Such a port might not even be able to produce executable code (unless you
> also want to write a sim).
Sure - I don't say running only a single pass doesn't make sense in some cases.
It just adds a less fragile way of testing for correctness.