VEA unittest: remove quality test on chromeos temporarily

media/gpu/video_encode_accelerator_unittest.cc

 // Copyright 2013 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include <inttypes.h>
 #include <stddef.h>
 #include <stdint.h>
 
 #include <algorithm>
 #include <memory>
@@ skipping 2125 matching lines @@
   const int test_type = GetParam();
   ASSERT_LT(test_type, 2) << "Invalid test type=" << test_type;
 
   if (test_type == 0)
     SimpleTestFunc<VEANoInputClient>();
   else if (test_type == 1)
     SimpleTestFunc<VEACacheLineUnalignedInputClient>();
 }
 
 #if defined(OS_CHROMEOS)
+// TODO(kcwu): add back test of verify_output=true after crbug.com/694131 fixed.

[Owen Lin, 2017/03/20 03:14:57]

How about file a new bug for adding this back?

[kcwu, 2017/03/20 03:33:29]

My plan is keeping 694131 open, so unnecessary to file a new bug.

[Owen Lin, 2017/03/20 06:53:02]

It just makes things clear. Maybe we will assign the bug to partner.

I mean, if you're going to fix the issue, I am fine with that. If not, I would
suggest to have a dedicate bug for adding back.

[kcwu, 2017/03/20 07:11:45]

I filed a bug to add it back.
https://bugs.chromium.org/p/chromium/issues/detail?id=703039

 INSTANTIATE_TEST_CASE_P(
     SimpleEncode,
     VideoEncodeAcceleratorTest,
     ::testing::Values(
-        std::make_tuple(1, true, 0, false, false, false, false, false, false),
-        std::make_tuple(1, true, 0, false, false, false, false, true, false)));
+        std::make_tuple(1, true, 0, false, false, false, false, false, false)));

[Owen Lin, 2017/03/20 03:14:57]

Why not filter out the test in autotest? My point is the test is still valid and
maybe we can use it for debugging/verification.

[kcwu, 2017/03/20 03:33:29]

I am also considering filter out in autotest. My reasons to disable it here are:
1. This test is not named explicitly. Its name is "..../1" with index. The index
may be out of sync between veatest and autotest.
2. It seems nice (?) to fix the bug and enable the test later in the same CL.

What do you think?

[Owen Lin, 2017/03/20 06:53:02]

Acknowledged. I think (1) is a good reason. But about (2), do we know the root
cause to the issue, maybe the root cause is not in chromium source code.

 
 INSTANTIATE_TEST_CASE_P(
     EncoderPerf,
     VideoEncodeAcceleratorTest,
     ::testing::Values(
         std::make_tuple(1, false, 0, false, true, false, false, false, false)));
 
 INSTANTIATE_TEST_CASE_P(ForceKeyframes,
                         VideoEncodeAcceleratorTest,
                         ::testing::Values(std::make_tuple(1,
@@ skipping 169 matching lines @@
 
   media::g_env =
       reinterpret_cast<media::VideoEncodeAcceleratorTestEnvironment*>(
           testing::AddGlobalTestEnvironment(
               new media::VideoEncodeAcceleratorTestEnvironment(
                   std::move(test_stream_data), log_path, run_at_fps,
                   needs_encode_latency, verify_all_output)));
 
   return RUN_ALL_TESTS();
 }