This fixes a crash when we try to load a program executable in the
worker thread in the multi-threaded shader compilation. This fixes
the angle_end2end_tests.

BUG=angle:900

Change-Id: Id4494864c7121630c43c2fe64a92674326eabc69
Reviewed-on: https://chromium-review.googlesource.com/245681Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Change-Id: I31ccffddcb04a45f19d3c3eb9a396c30e794b218
Reviewed-on: https://chromium-review.googlesource.com/243951Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Minmin Gong <mgong@microsoft.com>

DISALLOW_COPY_AND_ASSIGN was missing from CompilerImpl,
FramebufferImpl and DefaultAttachmentImpl.

BUG=angle:904

Change-Id: I2e57078cd5b40baf6962f19869c2f3b67ad6dc6e
Reviewed-on: https://chromium-review.googlesource.com/245514Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

GetAs and GetImplAs are template helpers that can replace all of our
custom "makeTextureD3D", etc methods. This will help save code across
different back-ends.

Change-Id: Ib3215c005bfac5a819c5d8f7d60a73a725241332
Reviewed-on: https://chromium-review.googlesource.com/245390Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

The original change caused a Chromium build break due to
"__uncaught_exception" not being defined in concrt.h. This is because Chromium
defines "_HAS_EXCEPTIONS=0" in its GYP, but ANGLE doesn't do this.

This change defines "_HAS_EXCEPTIONS=0" in ANGLE's GYP to match Chromium,
and refines it in ProgramD3D.cpp before including <future>.

Change-Id: Ic324702569bac8f4ae1381f308c4f3f11f190f9e
Reviewed-on: https://chromium-review.googlesource.com/244860Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Change-Id: Iba3d8d51f5fb04958c5f917d56dbc01df70da041
Reviewed-on: https://chromium-review.googlesource.com/245323Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Gregoire Payen de La Garanderie <Gregory.Payen@imgtec.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Also add an extra check in BindBufferBase/Range that XFB is not active.

BUG=angle:898

Change-Id: I0d61ebaa0674b0a375cc35efa614b021a6901eac
Reviewed-on: https://chromium-review.googlesource.com/243892Tested-by: Gregoire Payen de La Garanderie <Gregory.Payen@imgtec.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Was mistakenly returning a boolean instead of EGLint.

BUG=angle:850

Change-Id: I07f0995d66783f62124d37f5ea8e595f516cc84f
Reviewed-on: https://chromium-review.googlesource.com/244844Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <capn@chromium.org>

The check should be "==" instead of "!=". This gave a crash any time
we would generate an Error with a message.

Change-Id: I7c5d57e30053591b4713d8c829e89fd1c0703da0
Reviewed-on: https://chromium-review.googlesource.com/244820Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

The check should be "==" instead of "!=". This gave a crash any time
we would generate an Error with a message.

Change-Id: I9217e6237ee8a2d3a4d6caa799b8508edbf4e774
Reviewed-on: https://chromium-review.googlesource.com/244754Reviewed-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

In some cases, when an unpack buffer is enabled, we would use the
incorrect layer offset, because we would think the pixel pointer
was NULL. Use a new parameter instead of the double "null" meaning
to fix this.

This fixes several tests in the dEQP texture specification group.
(EG: dEQP-GLES3.functional.texture.specification.teximage3d_pbo)

BUG=angle:899

Change-Id: If96bba6987a04653301efc348d32de372a237aa1
Reviewed-on: https://chromium-review.googlesource.com/244092Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

Failing compile on Chromium bots:

FAILED: ninja -t msvc -e environment.x86 -- C:\b\build\goma/gomacc "C:\b\depot_tools\win_toolchain\vs2013_files\VC\bin\amd64_x86\cl.exe" /nologo /showIncludes /FC @obj\third_party\angle\src\libANGLE\renderer\d3d\libANGLE.ProgramD3D.obj.rsp /c ..\..\third_party\angle\src\libANGLE\renderer\d3d\ProgramD3D.cpp /Foobj\third_party\angle\src\libANGLE\renderer\d3d\libANGLE.ProgramD3D.obj /Fdobj\third_party\angle\src\libANGLE.cc.pdb 
c:\b\depot_tools\win_toolchain\vs2013_files\vc\include\concrt.h(4774) : error C3861: '__uncaught_exception': identifier not found
ninja: build stopped: subcommand failed.


This reverts commit 6d51f262.

Change-Id: Iebb2843dfbc3795290fbb33e1a111ddad59c3126
Reviewed-on: https://chromium-review.googlesource.com/244792Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

BUG=angle:900

Change-Id: Iad5dbbcc676e2a2b6dfc3d7bc6ab5957154de33e
Reviewed-on: https://chromium-review.googlesource.com/240490Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

There are currently some failures on the Chromium GPU FYI bots. The
failing machine is using an Intel HD 3000, and the group of failing
tests are the TextureTest.CopySubImageFloat_* group. I'm unable
to repro these failures, but disable the tests until we can fix them.

Also clean up our method of determining Renderer properties in
ANGLETest.

BUG=angle:850

Change-Id: I444b9d3497ea635fbd8957ac243b16c822c19d06
Reviewed-on: https://chromium-review.googlesource.com/244781Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <capn@chromium.org>

We were missing the entries in the def file, and we were missing two
new entry point functions from the new header. This fixes linking
against dEQP.

BUG=angle:901

Change-Id: I56d35bb6e8f3bd9e26943eaec196034605e154fc
Reviewed-on: https://chromium-review.googlesource.com/244444Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Our header was out-of-date with some EGL 1.5 Image-related stuff.
This fixes the new dEQP compile which relies on the entry points
being defined.

BUG=angle:901

Change-Id: Ic6ecf45c9e626e4b8e5f715f5301a9fbba295f3e
Reviewed-on: https://chromium-review.googlesource.com/244443Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Change-Id: I5ec9090e0129c778de8fcdc6cf70ea0fe460a19d
Reviewed-on: https://chromium-review.googlesource.com/244510Tested-by: Cooper Partin <coopp@microsoft.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

These corner case formats can come about from using floating
point textures and trying to copy between them. There currently
isn't a good alternative in WebGL 1 for copies.

BUG=angle:850

Change-Id: I3dc29e42b5ec7dcf071185bc09c6b3e9e3cb3207
Reviewed-on: https://chromium-review.googlesource.com/241048Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Also adds a few assert in SetData and various places to ensure that
images are not dirty before modifying the storage.

BUG=angle:873

Change-Id: I7cc922b33da0d79a1b5aefe216e9ccfaaf86a306
Reviewed-on: https://chromium-review.googlesource.com/241880Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Gregoire Payen de La Garanderie <Gregory.Payen@imgtec.com>

BUG=450257

Change-Id: I7609131d78f255bf9f79b42f39b02303a5c14b60
Reviewed-on: https://chromium-review.googlesource.com/243915Reviewed-by: Nicolas Capens <capn@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

Currently this method doesn't handle partial offsets.
Fixes dEQP tests in the texture specification group.

BUG=angle:899

Change-Id: I20160e49a46957b4f6cf96f529ab4ea7ddf7b343
Reviewed-on: https://chromium-review.googlesource.com/244091Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Some global initializers depend on other globals, for instance a
varying or attribute value. Since we use a static proxy variable for
these varyings, we need to initialize the global static after we
initialize the proxy in the shader preamble. This fixes a long-
standing compiler bug.

We should also add a WebGL test for this.

BUG=angle:878

Change-Id: I71db103a6b8c24fb862e0d8b32293da9bc2e8103
Reviewed-on: https://chromium-review.googlesource.com/243581Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Was mistakenly inserted as getWidth in 691e58cd.

BUG=angle:681

Change-Id: Ibccf056dcb47c8711422efb8bd40fb6f64ff57d9
Reviewed-on: https://chromium-review.googlesource.com/244090Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

ninja expects the output of an action to update the timestamp, however
a plain copy does not do this. So on a second run, ninja sees the
timestamp of the copy .bat is newer than the dll (original publish date)
so it re-copies. To avoid this, `touch` the dll on copy.

BUG=453037
Change-Id: Icb849d2b0faaceeccfc33e2eda509c835c9cdc93
Reviewed-on: https://chromium-review.googlesource.com/244040Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Scott Graham <scottmg@chromium.org>

The dEQP unpack buffer tests were failing because they use a max and
base level of zero with mipmapping enabled. Update our sampler
completeness checks to account for min and max level.

We'll have to dig in to our code to find corner cases in D3D with
base and max level.

Change-Id: I74357c6dc2e1908d0463d2e5cbc8ee91b61a3b7f
Reviewed-on: https://chromium-review.googlesource.com/240763Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Change-Id: I15598e42fa708861c0803acd6df4aa0b9d423ab5
Reviewed-on: https://chromium-review.googlesource.com/243950Tested-by: Minmin Gong <mgong@microsoft.com>
Reviewed-by: Cooper Partin <coopp@microsoft.com>
Reviewed-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>

Previously we would always reference mBody in several intermediate
output methods. This made using these traversals from within the
header, or for utility methods, very difficult. Instead, use a stack
where we write to the top InfoLog, and can push/pop from the stack.
This gives us more flexibility.

BUG=angle:878

Change-Id: I8a6c0382bad18b44d75158274c701db13d4d4e65
Reviewed-on: https://chromium-review.googlesource.com/243580Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

A new test 'sampler-struct-function-arg' fails with ToT WebGL.

Change-Id: Ie62208364ceacd4b980cb16f5de31cc4b5c62fa1
Reviewed-on: https://chromium-review.googlesource.com/243594Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angle:744

Change-Id: I60287f3c612011f505f1bc20632efc4dd17001d3
Reviewed-on: https://chromium-review.googlesource.com/221375Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>

BUG=angle:897

Change-Id: I3d6079b718bb5dd61cbd3098694204aae9bbc932
Reviewed-on: https://chromium-review.googlesource.com/243152Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Gregoire Payen de La Garanderie <Gregory.Payen@imgtec.com>

BUG=angle:877

Change-Id: I3df0fffb19f5ecbe439fbc2a8d6d239a5dc6b638
Reviewed-on: https://chromium-review.googlesource.com/243334Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Avoid rounding intermediate values when the intermediate value is not
consumed. For example, this avoids rounding the return value of
assignment, so "b = a;" becomes "b = frm(a);" instead of "frm(b =
frm(a))".

BUG=angle:874
TEST=compiler_tests

Change-Id: Ifcdb53fb1d07a2cf24e429cc237c2d0262dc32f8
Reviewed-on: https://chromium-review.googlesource.com/241852Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

Still do the actual generation in the Renderer9/11 for now but route it
through the DisplayImpl.

BUG=angle:658

Change-Id: Ib976fbd1b9e349146dc068bbdd4e6a3d2a5ddc69
Reviewed-on: https://chromium-review.googlesource.com/238860Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angle:658

Change-Id: I078d7f9ab4ab42bd7d3cd3865572ead8ed1f0b29
Reviewed-on: https://chromium-review.googlesource.com/234763Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angle:658

Change-Id: I44d95fad4e1be5cc76f443d724defd37da4a268f
Reviewed-on: https://chromium-review.googlesource.com/234762Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

Caused WebGL CTS failures on the texture-npot test:
https://www.khronos.org/registry/webgl/sdk/tests/conformance/textures/texture-npot.html

This reverts commit c67e6e9f.

Change-Id: I089e99859231e0d657084ac3647257c650a9da92
Reviewed-on: https://chromium-review.googlesource.com/243041Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angle:877

Change-Id: I15168ae32605b26aee08274464ffe68adb5a7e87
Reviewed-on: https://chromium-review.googlesource.com/242351Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

Change-Id: Ic229ced16f667882521536db5bd483c76b24d600
Reviewed-on: https://chromium-review.googlesource.com/242905Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

Change-Id: Iab44f6d634bdcd01348aa49de28b96c3b6086df7
Reviewed-on: https://chromium-review.googlesource.com/241945Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

Change-Id: Ibcb72f7b1aabcd98c4bebd91c3a57f5ba57eac3a
Reviewed-on: https://chromium-review.googlesource.com/242850Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>