This reverts commit 34087533.

Change-Id: Ide87d79791aa37095127f048929e1b7a925259ab
Reviewed-on: https://chromium-review.googlesource.com/322391Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

In some cases ANGLE flushes state for FBOs during validation. For
testing of the state synching code for FBOs, this makes end-to-end
testing impossible.

Solve this by partially implementing a hidden no_error extension,
hidden to the user by not exposing the extension string, but allowing
us to skip validation of some of the FBO methods that require checking
for complete FBOs.

BUG=angleproject:1280
BUG=angleproject:1260

Change-Id: I708f348ccec6697b974c48cd890ec75a703abe21
Reviewed-on: https://chromium-review.googlesource.com/322210
Tryjob-Request: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

applyShaders will update uniform data by uploading the constant buffer.
In the future uniform data will contain the base level of textures so
that shaders can use it to adjust results of texture functions. The most
natural place to set the base level in uniform data is applyTextures, so
it needs to be run before applyShaders.

updateSamplerMapping needs to be called before applyTextures, but
genericDraw* functions already do that. Remove redundant
updateSamplerMapping call from ProgramD3D::applyUniforms.

BUG=angleproject:596
BUG=angleproject:1092
BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I1828f7bdc75999872a26941a02cb4b6c6a66094d
Reviewed-on: https://chromium-review.googlesource.com/322121Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

ANGLE would previously treat a texture as incomplete whenever its base
level was above 0. This was because the base level wasn't being taken
into account correctly when determining what size the image at a specific
mip level should be. Fix this by taking the base level into account when
determining the expected size.

BUG=angleproject:596
TEST=angle_end2end_tests

Change-Id: I4b05514aed2858797ea46cf3570c3c40e5efc508
Reviewed-on: https://chromium-review.googlesource.com/322132
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Change-Id: Id59fca42b922c5c9bacb6eaed36ceb59e839d35e
Reviewed-on: https://chromium-review.googlesource.com/322133Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tryjob-Request: Andrii Shyshkalov <tandrii@chromium.org>
Tested-by: Andrii Shyshkalov <tandrii@chromium.org>

Also improve debug messages in window_test.

BUG=angleproject:1273

Change-Id: Ie47dc1d7ad3be5a55d2a199334bba606f2b45ad0
Reviewed-on: https://chromium-review.googlesource.com/322350
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

Previously the SwapLayer was kept in the CALayer hierarchy after
WindowSurfaceCGL's destruction which caused SwapLayer's callbacks to be
called after the shared data was freed, causing a segfault.

BUG=angleproject:1233

Change-Id: I71dce45330e6b7d5f06a41198c10890b1d8d520e
Reviewed-on: https://chromium-review.googlesource.com/320711
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

This test makes sure that even if sampler / texture registers get
grouped together based on type in the HLSL backend of the shader
translator, the D3D renderer still has the right register index
information for each ESSL sampler.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: Id09f96cfa903cd20c8e7229980fa363bb07b6045
Reviewed-on: https://chromium-review.googlesource.com/322200Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

With EGL_ANGLE_surface_orientation this is now necessary.

Change-Id: I92a20a128640f7f30b8d1d18173336d40fdad187
Reviewed-on: https://chromium-review.googlesource.com/322265Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: John Bauman <jbauman@chromium.org>

Calling setVisible multiple times in a row can cause a hang while
X11Window waits for events that will never be generted.

Change-Id: I43af8d4c79c7010dae78aa38c9716ed5b7b02865
Reviewed-on: https://chromium-review.googlesource.com/322380Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

Shadow samplers use a different sampler object type in D3D11 HLSL than
other ESSL samplers. They need to be able to coexist in the same shader.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I960f6a41c0237ed912b072527e14aebeca98cf7c
Reviewed-on: https://chromium-review.googlesource.com/322120
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

On some GL driver/window system combinations the default framebuffer may
do SRGB conversion, which affects the values written there by draw
operations and read by readPixels. This is known to happen at least on
some Intel systems. For this reason, use an FBO with a known format
instead in the texture end2end tests.

This enables turning on most TextureTest cases on the OpenGL backend.
Tested on Intel HD 4600.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I5a39bd1dadf15b7c52793a0f66d67e9f61618b83
Reviewed-on: https://chromium-review.googlesource.com/321612Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

Adding the function is very helpful for diagnostic messages. Also
change the name of the debug trace capture file to the more distinct
angle_debug.txt from debug.txt.

Change-Id: I59b3d61568fd3917cd8c4a8125ee1c28d608aa30
Reviewed-on: https://chromium-review.googlesource.com/321871
Tryjob-Request: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Using these methods to compute block offsets instead of the block
layout info could lead to some confusion.

Change-Id: Iaa3b91bdb517632107923076e3ece7efeab278c2
Reviewed-on: https://chromium-review.googlesource.com/321250
Tryjob-Request: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Alexis Hétu <sugoi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

The test must call glDeleteBuffers to delete a buffer instead of
glDeleteTextures. This was one reason behind the texture tests failing
on the GL backend.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I67042f6948146f3c3fca9e01231b0c4935a6f08d
Reviewed-on: https://chromium-review.googlesource.com/321611Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Previously the code didn't resize the depth buffer. Also adds a magic
setNeedsDisplay call when resizing a CALayer.

BUG=angleproject:1233

Change-Id: Id1906312f58d474e95a1b5ebfc15eff4344cd126
Reviewed-on: https://chromium-review.googlesource.com/318450
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

The test deliberately tries to render a scene using an invalid texture.
ANGLE rendered this correctly, but when it bound the dummy 1x1 black texture
that's used instead of invalid textures, it didn't bind a sampler state. This meant
that D3D had no pixel shader samplers bound, and outputted a debug warning.

Also, technically the test should have been setting GL_UNPACK_ALIGNMENT=1,
but that wasn't impacting the result of the test.

BUG=angleproject:1248

Change-Id: I871706d518077ea840a585ae0df8f9176e130cb1
Reviewed-on: https://chromium-review.googlesource.com/321391Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tryjob-Request: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

This adds some coverage for samplers that is missing from dEQP.

The shaders in the added tests both have a sampler array with two
samplers. The test shaders sample both textures, add up the results,
and the tests then check that the results of the shaders are correct.
Only constant indexing is used to access the sampler arrays.

One test accesses the sampler array in the main function, another one
first passes the array as a parameter to a user-defined function.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I9a0100705ac0da4905d50da37c0149fdbdd24c42
Reviewed-on: https://chromium-review.googlesource.com/321623Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

This adds some coverage for samplers that's missing from dEQP.

The test shader passes a sampler to an user-defined function, samples
the texture there and sets the returned sample to gl_FragColor. The
test then checks that the shader result is correct.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: I2cc511e591a4af01515c6fb38fd43c5f15337bcb
Reviewed-on: https://chromium-review.googlesource.com/321622Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

This adds some test coverage that's missing from dEQP.

The test uses two different samplers in a shader, a cube map sampler
and a 2D texture sampler, adds the sampling results together and checks
if the result is correct.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: If21e24c964fc208b5d1ccd446834dabf13194944
Reviewed-on: https://chromium-review.googlesource.com/321621Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

This removes the cube map setup and the draw scale parameter from the
tests that don't need them, and reuses the code for setting up the window
and the shader program for most of the texture tests. The tests are now
structured as follows:

TexCoordDrawTest: Test class that sets up a shader program for drawing
with texture coordinates. Vertex shader source can be overridden in
subclasses, and fragment shader source must be specified in subclasses.

Texture2DTest: Inherits TexCoordDrawTest, sets up a 2D texture and a
shader for drawing from it.

Texture2DTestWithDrawScale: Inherits Texture2DTest, adding a scale
parameter that scales the quad that gets drawn.

TextureCubeTest: Inherits TexCoordDrawTest, sets up a cube map and a 2D
texture and a shader for drawing from them.

Texture2DArrayTestES3: Inherits TexCoordDrawTest. Reserves a texture ID
and sets up an ESSL3 shader for drawing from a 2D texture array.

Also add a few comments about where things being tested are specified.

Also, ANGLETest::drawQuad parameter names are renamed to make their
meaning clearer. The parameters affect the vertex shader attribute
values, which the shader may use for other things besides setting the
vertex position.

BUG=angleproject:1261
TEST=angle_end2end_tests

Change-Id: Id673e36d5883aaaf47f2f830c2a1ad0ca293d578
Reviewed-on: https://chromium-review.googlesource.com/321620Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

This seems to be Clang-only.

Change-Id: I4276a29e3d1575bd47d58d819ed3223b2c66838a
Reviewed-on: https://chromium-review.googlesource.com/321820Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

This error was introduced in 82b5ab60
where the translator's essl and glsl sources are compiled in both
the translator and translator_lib targets. As translator depends on
translator_lib this caused symbols to be defined multiple times.

BUG=angleproject:1250

Change-Id: I50bab7112d1460e14710f50250fa2417b2163236
Reviewed-on: https://chromium-review.googlesource.com/321770Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

BUG=None

Change-Id: I9e417b37b072ea3874cd8b111d6b751295b9b0c2
Reviewed-on: https://chromium-review.googlesource.com/321760Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Reland with a link fix for Clang on Windows

In order to implement the --use-gl=angle Chrome flag on Linux, we need
to have the ANGLE .so directly in the build directory however GYP
places shared_libraries under lib/ on Linux. On the contrary loadable_modules
cannot be linked against so we need to keep the shared_library targets.

This commit adds _ANGLE versions of the libEGL and libGLESv2 targets as
loadable module that depend on their respective shared library.

BUG=522967

Change-Id: Iad968ab786b94b6041c279921de99633d23b48a7
Reviewed-on: https://chromium-review.googlesource.com/318470
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:1272

Change-Id: I20cb8968221c25bb374b3cc71d7ca619a35f1a2f
Reviewed-on: https://chromium-review.googlesource.com/320781Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

ESSL 3.00.4 section 4.6.1 says that using #pragma STDGL invariant(all)
in a fragment shader is an error, so make it an error. This spec
language is not found in ESSL 1.00, and it's been removed in ESSL 3.10.

BUG=angleproject:1276
TEST=angle_unittests

Change-Id: I2022f35475f867304b55dfb142f8568f6df28830
Reviewed-on: https://chromium-review.googlesource.com/321240
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>

This is Windows-only for now, but will be expanded to Linux. Also do
a bit of refactoring in our deqp.gypi file.

BUG=angleproject:1272

Change-Id: I526a99fe3b91a2d9e8901c7e033a4916a06780be
Reviewed-on: https://chromium-review.googlesource.com/320800Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

In WinRT we only use the HLSL code, so the GLSL/ESSL output code
is not necessary and is only adding to our binary size.

BUG=angleproject:1250

Change-Id: I9363ca3981bde50a230f8353c1bcc09f6ea209cb
Reviewed-on: https://chromium-review.googlesource.com/317358Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tryjob-Request: Austin Kinross <aukinros@microsoft.com>
Tested-by: Austin Kinross <aukinros@microsoft.com>

This test checks for OES_texture_half_float, but it's trying to use a
16F texture as a color buffer so it should be checking for
EXT_color_buffer_half_float.

Since EXT_color_buffer_half_float requires more format support (e.g.
R16F, RG16F, RGB16F) than OES_texture_half_float, it's possible for an
implementation to support one extension and not the other. D3D11 Feature
Level 9_3 does this.

Change-Id: I40fb38aa3f419be2606464bd0476cd064afe90fa
Reviewed-on: https://chromium-review.googlesource.com/319642Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=574909

Change-Id: I6dc4b010b67a9c7f3b64215a168e6ed5e894989b
Reviewed-on: https://chromium-review.googlesource.com/320635
Commit-Queue: Byungseon Shin <xingri@gmail.com>
Tested-by: Byungseon Shin <xingri@gmail.com>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>

Fixes a glitch on OSX where the root window would swap twice, causing a
glitch.

Change-Id: I5886105fb18ed8f45e69cd10b5d1cac49b82093f
Reviewed-on: https://chromium-review.googlesource.com/320291
Tryjob-Request: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

This fixes the static library build of this target on Linux.

BUG=angleproject:1270

Change-Id: Id65bd9b7136918829f385920b540a4df83650c6f
Reviewed-on: https://chromium-review.googlesource.com/320615Tested-by: Kenneth Russell <kbr@chromium.org>
Tryjob-Request: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Function prototypes should not be allowed inside other functions
according to ESSL 3.00.4 section 4.2.4.

BUG=angleproject:1068
TEST=angle_unittests,
     dEQP-GLES*.functional.shaders.functions.invalid.local_function_proto*

Change-Id: I54160da4d49b92a6cd7cbee020e67733963d4e10
Reviewed-on: https://chromium-review.googlesource.com/320091
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

The same function can't be declared twice in ESSL 1.00. In ESSL 3.00 this
is allowed.

A function prototype following the definition of that function is not
interpreted as redeclaration, and the shader compiler continues to allow
this.

BUG=angleproject:1067
TEST=angle_unittests,
     dEQP-GLES2.functional.shaders.functions.invalid.double_declare*

Change-Id: I6d2ddafd456d378d92839600f19069ad1cd19aff
Reviewed-on: https://chromium-review.googlesource.com/320082
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This change is pure refactoring. It will make it easier to fix bugs
related to function declarations.

BUG=angleproject:911
BUG=angleproject:1067
BUG=angleproject:1068
TEST=angle_unittests

Change-Id: I031783dc02612d9cf3ff7a9c8291cf8ab33577aa
Reviewed-on: https://chromium-review.googlesource.com/320081
Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

Standalone builds failed to generate projects.

Change-Id: I2897f84242255edabd0606ad858a048f635db0bc
Reviewed-on: https://chromium-review.googlesource.com/320680Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

This fixes member initialization order and unused private member
warnings that broke the Clang build after the last Angle roll.

BUG=82385

Change-Id: If866f352bda6d66093c180fdbec6c16bf902b628
Reviewed-on: https://chromium-review.googlesource.com/320466Reviewed-by: Nico Weber <thakis@chromium.org>
Tryjob-Request: Nico Weber <thakis@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tryjob-Request: Geoff Lang <geofflang@chromium.org>
Tested-by: Nico Weber <thakis@chromium.org>

Move assignment operator of both classes contained a memory leak.
Changed type of mMessage from mutable std::string *to mutable
std::unique_ptr<std::string>

BUG=angleproject:1264

Change-Id: I7d1419b2e9f385b36afebfd6371983be33ee9e61
Reviewed-on: https://chromium-review.googlesource.com/319520Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tryjob-Request: Tibor Ouden, den <tibordenouden@gmail.com>
Tested-by: Jamie Madill <jmadill@chromium.org>

The root_build_dir variable yields the root of the build tree.
But when the component is being built under a non-default
toolchain, its output files are delivered to a toolchain
subdirectory instead.  The copy_compiler_dll rule should put its
output next to where libANGLE and things built using it will go,
which is root_out_dir.  This fixes cases where the angle GN
component is instantiated under more than one toolchain, which
heretofore would produce multiple Ninja rules for the same target
file $root_build_dir/d3dcompiler_47.dll and thus get complaints
from Ninja.

BUG=512869

Change-Id: I9fe83c5c5fa74902ef270654d3707636588af6b7
Reviewed-on: https://chromium-review.googlesource.com/320119Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Roland McGrath <mcgrathr@chromium.org>