gl_WorkGroupSize should be written into the AST as a constant node
instead of a symbol node. In correct shaders, local size is guaranteed
to have been declared before any references to gl_WorkGroupSize -
otherwise the shader translator was already generating an error.

This ensures that work group size can be used to size arrays as
specified and also works around a crash issue on NVIDIA Linux OpenGL
driver.

BUG=angleproject:1442
TEST=angle_unittests

Change-Id: I9b1a4bff16ecf2c3db1511c3702756346cdd2f6b
Reviewed-on: https://chromium-review.googlesource.com/418735
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Might be flaky, more suppressions might be needed.

BUG=angleproject:1633

Change-Id: Ic5665aadb7cbbae82639d206f29a2db859622451
Reviewed-on: https://chromium-review.googlesource.com/418399Reviewed-by: Jamie Madill <jmadill@chromium.org>

Compute shaders were added to the core OpenGL spec in version 4.3.

Also don't output GL_ARB_explicit_attrib_location extension directive
for compute shaders. This extension is only relevant for vertex and
fragment shaders.

BUG=angleproject:1442
TEST=angle_end2end_tests

Change-Id: I92021ac048b38ffbc5a61af408b37c387193389b
Reviewed-on: https://chromium-review.googlesource.com/418657Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Previously infinity and NaN resulting from constant folding would be
clamped to finite 32-bit float range when they were written in shader
output. Now they are written as a bit pattern in case the shader
version allows it.

This does not guarantee that NaNs work, but this is fine, since ESSL
3.00.6 spec has very loose requirements when it comes to NaNs.

BUG=angleproject:1654
TEST=angle_end2end_tests

Change-Id: I9997000beeaa8ed22523c22d5cf6929cdfc93f60
Reviewed-on: https://chromium-review.googlesource.com/417301
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Also add a more robust set of unit tests for this internal class.
Because the GL spec allows for succeess when packing any set of
varyings, it's impossible to write negative end to end tests. This CL
also rewrites our disabled varying packing GLSL tests as unit tests.

BUG=angleproject:1296
BUG=angleproject:1638

Change-Id: I78153742517d5c72ddb13ff59dc44ddc4af42fc2
Reviewed-on: https://chromium-review.googlesource.com/415555
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This code doesn't actually call any D3D runtime methods, so it can be
included in our cross platform unit tests. Just include the varying
packing code for now.

BUG=angleproject:1296
BUG=angleproject:1638

Change-Id: I5c0ccccb00e1c13ca98448256be048f7cd0968ce
Reviewed-on: https://chromium-review.googlesource.com/417116
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Many tests were removed from mustpass at some point, and a couple
were expected to fail but passing. Also update some of the global
skip settings to use DEBUG RELEASE as a catch-all.

BUG=angleproject:989
BUG=angleproject:998
BUG=angleproject:1017
BUG=angleproject:1018
BUG=angleproject:1019
BUG=angleproject:1020
BUG=angleproject:1025
BUG=angleproject:1027
BUG=angleproject:1028
BUG=angleproject:1033
BUG=angleproject:1034
BUG=angleproject:1051
BUG=angleproject:1108
BUG=angleproject:1143
BUG=angleproject:1335
BUG=angleproject:1418
BUG=angleproject:1655
BUG=angleproject:1656

Change-Id: I731d3adeed36509ce69f72e5fcc80d0adc89cb67
Reviewed-on: https://chromium-review.googlesource.com/418063Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Otherwise when trying to add the declarations back, things might fail
because the loop initialization is a sequence and not a block.

BUG=668028

Change-Id: I8d84a25c25765e9655c16ce56604ae08f0f8176c
Reviewed-on: https://chromium-review.googlesource.com/414305
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>

Many test expectations listed were not in the mustpass list, so
weren't even necessary.

BUG=angleproject:989
BUG=angleproject:1015
BUG=angleproject:1016
BUG=angleproject:1020
BUG=angleproject:1025
BUG=angleproject:1027
BUG=angleproject:1029
BUG=angleproject:1030
BUG=angleproject:1128

Change-Id: Ia18796cb3388e8528b4351cb2c81ff9793a1d250
Reviewed-on: https://chromium-review.googlesource.com/417413Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:1595
TEST=dEQP-GLES31.functional.state_query.integer.draw_indirect_buffer_binding*

Change-Id: Ib8f712fdf10411ef0b7b63742d17c3caca99137b
Reviewed-on: https://chromium-review.googlesource.com/416193Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

BUG=668223

Change-Id: I2f38b358dd4a6b9327562704dacce37f2b16acaa
Reviewed-on: https://chromium-review.googlesource.com/417192Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Constant folding arithmetic operations that involve infinity are now
handled correctly in the cases where the result is infinity or zero.
The implementation mostly relies on C++ to implement IEEE float
arithmetic correctly so that unnecessary overhead is avoided.

Constant folding arithmetic operations that result in overflow now
issue a warning but result in infinity. This is not mandated by the
spec but is a reasonable choice since it is the behavior of the
default IEEE rounding mode.

Constant folding arithmetic operations that result in NaN in IEEE will
generate a warning but the NaN is kept. This is also not mandated by
the spec, but is among the allowed behaviors.

There's no special handling for ESSL 1.00. ESSL 1.00 doesn't really
have the concept of NaN, but since it is not feasible to control
generating NaNs at shader run time either way, it should not be a big
issue if constant folding may generate them as well.

TEST=angle_unittests
BUG=chromium:661857

Change-Id: I06116c6fdd02f224939d4a651e4e62f2fd4c98a8
Reviewed-on: https://chromium-review.googlesource.com/414911Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Change "DEBUG RELEASE" modifier to "OPENGL D3D11" to be more specific.

Add D3D11 expectations.

BUG=angleproject:1442

Change-Id: If3a9704e7f16cc473fd6ea410d10eb01ae1a8008
Reviewed-on: https://chromium-review.googlesource.com/416960Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

ES 3.0 doesn't require non-constant sampler array indexing, and we
can't require it for ES 2 without breaking WebGL for some users.

BUG=671280

Change-Id: Ib04446284acc92410a90c683a3d222871de554a8
Reviewed-on: https://chromium-review.googlesource.com/416195
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>

BUG=angleproject:1366

Change-Id: I9e44679754eb704b390191c28206dedc3dc7cc4f
Reviewed-on: https://chromium-review.googlesource.com/367082
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

UNIMPLEMENTED debug spam was causing the tests time time out.  We can mark them
as FAIL again once the dEQP test setup/tear down code doesn't emmit so many
messages.

Implement a couple validation cases for ES 3.1 to greatly reduce the spam.

BUG=angleproject:1647
BUG=angleproject:1442

Change-Id: Ie7b4ac8737a2df1c0ada6ad53154ddf2f37d9c3c
Reviewed-on: https://chromium-review.googlesource.com/415520
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Change-Id: I2f96baedf10d346eaa150bab04f8f6ca3ba573b9
Reviewed-on: https://chromium-review.googlesource.com/414272Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1645

Change-Id: If35d277705ce9f2882d9ba059bd56e494ca9d123
Reviewed-on: https://chromium-review.googlesource.com/415613Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>

PointSpritesTest.PointWithoutAttributesCompliance:

https://build.chromium.org/p/chromium.gpu.fyi/builders/Win7%20Debug%20%28New%20Intel%29/builds/1741

Showing as failed now that flaky tests are not retried.

BUG=angleproject:1346

Change-Id: Ic0fce361f37394da8771fa82288b0a9206d3a347
Reviewed-on: https://chromium-review.googlesource.com/415448Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:1442

Change-Id: I49299d2345f67b7ed3e2c80db0ec85589fc7583c
Reviewed-on: https://chromium-review.googlesource.com/415526Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

BUG=None

Change-Id: Id4b9674b6e98b8313c47f3fbc78cc93d137ec14c
Reviewed-on: https://chromium-review.googlesource.com/415569Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The mechanism for requesting extensions is now a separate extension.
Added a way to query the extensions that support enabling.

BUG=angleproject:1523

Change-Id: I2efaa9f6d67b12ecae325f455404e34ba04d0e7c
Reviewed-on: https://chromium-review.googlesource.com/414529
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Change-Id: If3883de8f49a3fb33ecc598ba0b86c658cd3f0ca
Reviewed-on: https://chromium-review.googlesource.com/415565Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

This is an work-in-progress CL to prototype using ComPtr.

It also has a new design for internal errors that doesn't use
FormatString, preferring a stream-based approach.

One thing to be aware of is that the address operator does not
behave as expected with ComPtr - we should use ::AddressOf.

BUG=angleproject:530
BUG=angleproject:1644

Change-Id: If5643e9e5726fd9aa5cbd422fca12ae169eb5b1f
Reviewed-on: https://chromium-review.googlesource.com/415027
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

The constant folding test classes are moved into a separate file in
test_utils. This will enable adding multiple test files that use
constant folding test classes, so that constant folding tests can be
organized better.

TEST=angle_unittests
BUG=chromium:661857

Change-Id: I00bf25a4b941bdc1364ff5aa9bee2d571e4b0ea0
Reviewed-on: https://chromium-review.googlesource.com/414910Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Change-Id: I63aa1c1b3bbc52b0b522e4a703e6d20574eb21ad
Reviewed-on: https://chromium-review.googlesource.com/414588Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:422

Change-Id: I1a1c4551f2fc55b74afd66e7f18ea61e77250eb6
Reviewed-on: https://chromium-review.googlesource.com/414634
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:1589
TEST=dEQP-GLES31.functional.program_uniform.*

Change-Id: I27fffa755fd277ed918746259cac88ab9e349c41
Reviewed-on: https://chromium-review.googlesource.com/412193
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

These tests became flaky when google test retries are turned off.

BUG=chromium:669196
BUG=angleproject:1643
BUG=angleproject:1541

Change-Id: I7cc10f532662928e5436a4e88a42fc109c00bfe8
Reviewed-on: https://chromium-review.googlesource.com/414216Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Choose to use std::async for now to implement multi-threaded compiles.
This should work across platforms and also be usable for other
threading tasks. Note that std::async does not have a good way to
wait for multiple std::futures. Also the Linux compile of std::async
is broken due to a bug in an old STL version, so disable it on this
platform.

The implementation uses a static polymorphism approach, which should
have very good performance (no virtual calls).

This design leaves the door open for other future implementations,
such as a Win32 thread pool, or one based on angle::Platform.

BUG=angleproject:422

Change-Id: Ia2f13c3af0339efaca1d19b40b3e08ecca61b8e8
Reviewed-on: https://chromium-review.googlesource.com/413712
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1634

Change-Id: I3822b99b59d4653e4d9a2c1d3dd16734f2050fae
Reviewed-on: https://chromium-review.googlesource.com/414437
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:1442

Change-Id: I18bee163e7c7ce159cd23d6136486c49b7d9ec94
Reviewed-on: https://chromium-review.googlesource.com/414950Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Change supported ES version to 3.1 on win.

Without GL_TEXTURE_2D_MULTISAMPLE enum, it will reach UNREACHABLE()
when running dEQP ES 3.1 tests on Win.

BUG=angleproject:1442

Change-Id: I55114854c4e9cba1caac10d3da5b416ef66a0d3e
Reviewed-on: https://chromium-review.googlesource.com/414606
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

IEEE rules for generating zero or infinity are now checked for the
constant folding of several built-in functions except for the cases
where ESSL 3.00.6 explicitly states that the results are undefined.

BUG=chromium:661857
TEST=angle_unittests

Change-Id: I2ce427229a5583039694d060ea6db29c5bdace97
Reviewed-on: https://chromium-review.googlesource.com/414370Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

BUG=chromium:660670

Change-Id: I51c7e93f071041663f6b9a39ee75c0cca2264015
Reviewed-on: https://chromium-review.googlesource.com/414447Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:422

Change-Id: I10a2192c5c91360a1b7fe5e4d014bcf1fa84c96e
Reviewed-on: https://chromium-review.googlesource.com/413713Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This also makes BlitGL work correctly on OpenGL ES (provided vertex
arrays are available)

BUG=angleproject:1356

Change-Id: Icb7cef35bebfe6672220aa0b312ab89187dbf585
Reviewed-on: https://chromium-review.googlesource.com/412452
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:1636

Change-Id: Idea0e4f4a4415351b83e8f1a39b697139ff9ebe6
Reviewed-on: https://chromium-review.googlesource.com/414205
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Add a configurable limit for how many nested MacroExpander objects can
be created in the preprocessor, so that stack overflow can be
prevented in case of malicious shaders. By default the limit is set to
1000. In unit tests the limit is set lower to make the test run
faster.

Includes refactoring of most of the preprocessor tests so that they
use utility functions provided by the test class instead of repeating
the same code for initializing the preprocessor.

BUG=angleproject:1600
TEST=angle_unittests

Change-Id: I23b5140d9f2dc52df96111650db63150f7238494
Reviewed-on: https://chromium-review.googlesource.com/413986
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:1468

Change-Id: I3cf0d0a9486778b1d4384e00359034e08e425191
Reviewed-on: https://chromium-review.googlesource.com/414284Reviewed-by: Corentin Wallez <cwallez@chromium.org>