Enable CopyTextureCHROMIUM in Blink for Tex{Sub}Image2D with more cases

content/test/gpu/gpu_tests/webgl2_conformance_expectations.py

Index: content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
diff --git a/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py b/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
index ecb27677b8565bb76b086c95ccaca81eccc00746..acc51b775c2161ba38f62e19ed92bd84854aa308 100644
--- a/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
+++ b/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
@@ -51,6 +51,8 @@ class WebGL2ConformanceExpectations(WebGLConformanceExpectations):
     self.Fail('conformance2/textures/misc/' +
         'integer-cubemap-specification-order-bug.html',
         bug=483282) # owner:cwallez, test might be buggy
+    self.Fail('conformance/textures/misc/tex-sub-image-2d-bad-args.html',
+        bug=625738)

[Ken Russell (switch to Gerrit), 2017/04/01 01:13:26]

What issues remain in getting this test to pass, both for WebGL 1.0 and 2.0? Is
some more validation needed in Copy{Sub}TextureCHROMIUM?

[qiankun, 2017/04/05 08:43:14]

We need to do TexSubImage validations in CopySubTextureCHROMIUM. The difficulty
is CopySubTextureCHROMIUM doesn't have format/type arguments or know whether the
caller is TexSubImage or not.

[Ken Russell (switch to Gerrit), 2017/04/06 21:14:22]

OK. I assume we'll continue to work together to resolve this, either by adding
more arguments to CopySubTextureCHROMIUM, or assuming that its caller is
TexSubImage2D and ensuring that all callers using it for a TexImage2D operation
get the format/type correct when defining the texture (if that will work).

 
     # Windows only.
     self.Fail('conformance2/rendering/blitframebuffer-outside-readbuffer.html',