llvm dev - Jan 2017 - Removed a call to EmitXRayTable() from ARMAsmPrinter

> As Oleg points out, apparently the way we've configured the XRay tests
in compiler-rt/... excludes running tests on non-64-bit platforms. arm7 is
definitely not 64-bit and therefore means we've not been running those
tests. :(
I believe that for this particular case one does not need to have
execution tests - you'd simply need to check for expected output and
that's it.

-- 
With best regards, Anton Korobeynikov
Department of Statistical Modelling, Saint Petersburg State University

> On 10 Jan 2017, at 17:00, Anton Korobeynikov <anton at
korobeynikov.info> wrote:
> 
>> As Oleg points out, apparently the way we've configured the XRay
tests in compiler-rt/... excludes running tests on non-64-bit platforms. arm7 is
definitely not 64-bit and therefore means we've not been running those
tests. :(
> I believe that for this particular case one does not need to have
> execution tests - you'd simply need to check for expected output and
> that's it.
Oh, that's a good point. Yes, we definitely ought to be looking for the
instrumentation map in the codegen tests for arm7.

That's probably something we can set up in the same change, and ought to be
simple to do anyway.

-- Dean

Then I plan to enable back emission of XRay instrumentation map in Arm32
and testing of Arm32 (about 1 line each, but they may break the builds),
and to submit this for review together with the tail call patches.

Serge

On 10 January 2017 at 09:05, Dean Michael Berris <dean.berris at
gmail.com>
wrote:
>
> > On 10 Jan 2017, at 17:00, Anton Korobeynikov <anton at
korobeynikov.info>
> wrote:
> >
> >> As Oleg points out, apparently the way we've configured the
XRay tests
> in compiler-rt/... excludes running tests on non-64-bit platforms. arm7 is
> definitely not 64-bit and therefore means we've not been running those
> tests. :(
> > I believe that for this particular case one does not need to have
> > execution tests - you'd simply need to check for expected output
and
> > that's it.
>
> Oh, that's a good point. Yes, we definitely ought to be looking for the
> instrumentation map in the codegen tests for arm7.
>
> That's probably something we can set up in the same change, and ought
to
> be simple to do anyway.
>
> -- Dean
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20170110/5270bef6/attachment-0001.html>