Pass shader version to varying link check

Pass shader version to varying link check

Rules for invariant varying (or in/out variable) matching are different
for ESSL 3 and ESSL 1. Use a recently added ANGLE API to choose which
rules to use based on shader version.

Tested by WebGL 2 test deqp/data/gles3/shaders/qualification_order.html

BUG=488372
TEST=WebGL 2 conformance tests

Committed: https://crrev.com/fe5270311b555b02e2779541a1577e2fa7ff66a5
Cr-Commit-Position: refs/heads/master@{#330335}

[oetuaho-nv, 2015-05-15 11:25:30]

oetuaho@nvidia.com changed reviewers:
+ bajones@chromium.org, kbr@chromium.org, zmo@chromium.org

[oetuaho-nv, 2015-05-15 11:25:32]

Please review!

Should we try to also update
content/test/gpu/gpu_tests/webgl_conformance_expectations.py in the same patch?
I'm not yet sure whether that starts passing on all platforms with this change,
but it should pass at least on Linux.

[Zhenyao Mo, 2015-05-15 17:19:39]

On 2015/05/15 11:25:32, oetuaho-nv wrote:
> Please review!
> 
> Should we try to also update
> content/test/gpu/gpu_tests/webgl_conformance_expectations.py in the same
patch?
> I'm not yet sure whether that starts passing on all platforms with this
change,
> but it should pass at least on Linux.

LGTM.

You can land this CL and check GPU bots on build.chromium.org to see if the test
is passing on various platforms (wait at least two hours after landing), and
then update the test expectations.

Don't you have try jobs privilege already?  If not, we should get you one. Other
path is to update the conformance test expectations and see how the try bots are
doing, and decide whether to keep that update.

[Zhenyao Mo, 2015-05-15 17:20:21]

Assume ANGLE already rolled in with the extra version param?

[oetuaho-nv, 2015-05-18 07:13:50]

On 2015/05/15 17:20:21, Zhenyao Mo wrote:
> Assume ANGLE already rolled in with the extra version param?

Yep, latest ANGLE roll contains the necessary patch. I'll check the GPU bots
after landing and update the expectations according to the results.

[oetuaho-nv, 2015-05-18 07:14:10]

The CQ bit was checked by oetuaho@nvidia.com

[commit-bot: I haz the power, 2015-05-18 07:14:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1135733006/1

[commit-bot: I haz the power, 2015-05-18 08:18:30]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-05-18 11:33:58]

Patchset 1 (id:??) landed as
https://crrev.com/fe5270311b555b02e2779541a1577e2fa7ff66a5
Cr-Commit-Position: refs/heads/master@{#330335}

[oetuaho-nv, 2015-05-18 12:18:57]

I looked for the results of WebGL 2 conformance tests on the GPU bots but
couldn't find them. I was looking at the "webgl_conformance_tests" stdio log.
Are they somewhere else in there?

[Ken Russell (switch to Gerrit), 2015-05-18 19:00:16]

On 2015/05/18 12:18:57, oetuaho-nv wrote:
> I looked for the results of WebGL 2 conformance tests on the GPU bots but
> couldn't find them. I was looking at the "webgl_conformance_tests" stdio log.
> Are they somewhere else in there?

The WebGL 2 conformance tests aren't yet running as part of the testing on the
GPU bots. zmo's been starting to put this in place.