Document WebGL2 CTS expectations findings

content/test/gpu/gpu_tests/webgl2_conformance_expectations.py

 # Copyright (c) 2015 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.
 from gpu_tests.webgl_conformance_expectations import WebGLConformanceExpectations
 
 # See the GpuTestExpectations class for documentation.
 
 class WebGL2ConformanceExpectations(WebGLConformanceExpectations):
   def __init__(self, conformance_path):
     super(WebGL2ConformanceExpectations, self).__init__(conformance_path)
@@ skipping 77 matching lines @@
         ['win'], bug=483282)
     self.Fail('conformance2/textures/canvas/' +
         'tex-2d-rgba4-rgba-unsigned_byte.html',
         ['win'], bug=483282)
     self.Fail('conformance2/textures/canvas/' +
         'tex-2d-rgba4-rgba-unsigned_short_4_4_4_4.html',
         ['win'], bug=483282)
     self.Fail('conformance2/textures/misc/tex-unpack-params.html',
         ['win'], bug=483282)
 
+    # Failing because ANGLE using a different interpretation of the spec for
+    # CopyTexImage2D so that for example a copy from RGBA8 to R8 fails.

[Zhenyao Mo, 2016/04/22 18:50:16]

We should resolve this ambiguity in spec interpretation quickly and remove this
large trunk of entries.

Can you create a bug to track this? and update the bug IDs accordingly below.

     self.Fail('conformance2/textures/webgl_canvas/' +
         'tex-2d-r8-red-unsigned_byte.html',
         ['win'], bug=483282)
     self.Fail('conformance2/textures/webgl_canvas/' +
         'tex-2d-rg8-rg-unsigned_byte.html',
         ['win'], bug=483282)
     self.Fail('conformance2/textures/webgl_canvas/' +
         'tex-2d-rgb8-rgb-unsigned_byte.html',
         ['win'], bug=483282)
     self.Fail('conformance2/textures/webgl_canvas/' +
@@ skipping 256 matching lines @@
         ['linux'], bug=483282)
     self.Fail('conformance2/reading/read-pixels-from-fbo-test.html',
         ['linux'], bug=483282)
     self.Fail('conformance2/rendering/draw-buffers.html',
         ['linux'], bug=483282)
     self.Fail('conformance2/textures/misc/tex-unpack-params.html',
         ['linux'], bug=483282)
     # We want to mark this Flaky for all of Linux however we currently skip
     # all the tests on Intel. Tag this with AMD and Nvidia to avoid an
     # expectation conflict that would make this test run on Intel.
+    # Flaky because the virtual gl contexts implementation calls glUseProgram
+    # while a transform feedback is active.

[Zhenyao Mo, 2016/04/22 18:50:16]

Ah, my current CL (adding transform_feedback_manager) is addressing this.

     self.Flaky('deqp/functional/gles3/negativeshaderapi.html',
         ['linux', 'amd', 'nvidia'], bug=483282)
 
     # Linux NVIDIA only.
     self.Fail('deqp/functional/gles3/fbostatequery.html',
         ['linux', 'nvidia'], bug=483282)
     self.Fail('deqp/functional/gles3/vertexarrays.html',
         ['linux', 'nvidia', 'debug'], bug=483282)
 
     # Linux AMD only.
@@ skipping 87 matching lines @@
     # self.Fail(page_name,
     #     ['linux', ('nvidia', 0x1), 'debug', 'opengl'])
     # self.Fail(page_name,
     #     ['win', ('nvidia', 0x1), 'debug', 'opengl'])
 
     # Conflicts if between a generic os condition and a specific version
     # self.Fail(page_name,
     #     ['xp', ('nvidia', 0x1), 'debug', 'opengl'])
     # self.Fail(page_name,
     #     ['win', ('nvidia', 0x1), 'debug', 'opengl'])