ARM HW bots: looks for FEATURE_VERSION in the right place

The ARM HW bots weren't running SIMD tests because they weren't looking for the FEATURE_VERSION file at the right place: they were looking in pnacl_newlib instead of pnacl_translator, and the former only exists if the build was local.

R= dschuff@chromium.org
TEST= git try -b nacl-arm_opt_panda
BUG= none

[JF, 2014-06-14 01:01:23]

dschuff: the fix we discussed doesn't seem to have fixed the issue, feature
version is still 0.

See:
 
http://build.chromium.org/p/tryserver.nacl/builders/nacl-arm_hw_opt_panda/bui...

[Derek Schuff, 2014-06-17 21:58:27]

On 2014/06/14 01:01:23, JF wrote:
> dschuff: the fix we discussed doesn't seem to have fixed the issue, feature
> version is still 0.
> 
> See:
>  
>
http://build.chromium.org/p/tryserver.nacl/builders/nacl-arm_hw_opt_panda/bui...

I think it works (seems to work locally) but we still need to put the
FEATURE_VERSION file in the translator package first.

[JF, 2014-06-17 22:10:42]

> I think it works (seems to work locally) but we still need to put the
> FEATURE_VERSION file in the translator package first.

Oh I thought we already did that?

-- 
You received this message because you are subscribed to the Google Groups
"Native-Client-Reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to native-client-reviews+unsubscribe@googlegroups.com.
To post to this group, send email to native-client-reviews@googlegroups.com.
Visit this group at http://groups.google.com/group/native-client-reviews.
For more options, visit https://groups.google.com/d/optout.

[Derek Schuff, 2014-06-17 22:15:06]

Yeah, I just noticed that for some reason it was missing on my local build. my
local test did seem to work though. we'll see what happens with this try job.

[Derek Schuff, 2014-06-17 23:33:51]

On 2014/06/17 22:15:06, Derek Schuff wrote:
> Yeah, I just noticed that for some reason it was missing on my local build. my
> local test did seem to work though. we'll see what happens with this try job.

I figured it out. The first set of tests (including the step you linked to) does
not use the sandboxed translator, it just uses the prebuilt nexe. I had
forgotten we had that. if you look in your try jobs, the later steps (which run
toolchain_tests with sandboxed translator) it does successfully run the
vector_shuffle test. So we can just tweak the logic, maybe to just check for the
presence of the main toolchain file, and fall back to the sandboxed translator
file.

[JF, 2014-06-24 00:20:53]

Closing, dschuff fixed this in https://codereview.chromium.org/341973003/