ESSL 3.00 allows a wider variety of types of input/output variables
than ESSL 1.00, but there are still specific restrictions on structs,
matrices and arrays.

Some of the checks need to be implemented twice: once for array syntax
where the brackets are after the type, and another time for array syntax
where the brackets are after the variable name.

This requires fixes to constant folding unit tests which were previously
incorrectly using matrix outputs in fragment shaders.

New unit tests are added for several of the rules introduced, but some
cases are also covered by dEQP.

TEST=angle_unittests, dEQP-GLES.functional.shaders.linkage.varying.rules.*
BUG=angleproject:1061

Change-Id: I655b054cfe56d376db775b96a2bb41b3ac5740b0
Reviewed-on: https://chromium-review.googlesource.com/285482Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

All the previously implemented surfaces used to preserve the content of
the surface on a swap but this is no longer the case with CGL.

BUG=angleproject:891

Change-Id: I264c579bcbbd22b4eeb09e78aa95751d223694fc
Reviewed-on: https://chromium-review.googlesource.com/289871Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

BUG=516027

Change-Id: Ibbf45c863570581ea79e198604cd4c5cc1836620
Reviewed-on: https://chromium-review.googlesource.com/289996Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

These macros replace a lot of repeated timing code, and mirror how
Chromium operates.

BUG=516027

Change-Id: Ib03e35d0f3b8638b9b0e8785c397c8381f16b041
Reviewed-on: https://chromium-review.googlesource.com/289995Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

The previous skeleton was name NSGL because it seemed the easier route
to go, however Chrome uses CGL internally, which we want to match.

BUG=angleproject:891

Change-Id: I00eb597080c8b735545750dfcd6ad80dcea79300
Reviewed-on: https://chromium-review.googlesource.com/289870Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

std::to_ullong is c++11, failing on Mac.

BUG=angleproject:1040

This reverts commit 37a9e712.

Change-Id: I5cecd1fe95b889feffb556f67d082cec585256f4
Reviewed-on: https://chromium-review.googlesource.com/290211Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

This class allows us to quickly notify Impl/Renderer layers of
which state changed.

BUG=angleproject:1040
TEST=angle_unittests (Mac/Win/Linux)

Change-Id: I2f27bc65b4f627a261e94e3e3b7095b169986cc6
Reviewed-on: https://chromium-review.googlesource.com/289557Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Adds more profiling help for startup shader compiles in Chrome.

BUG=510151

Change-Id: I550b55b1280542d1eaab9ed6166387b663cafffe
Reviewed-on: https://chromium-review.googlesource.com/290091Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Change-Id: I225f38b4616ac47e690dccbcb67074fedc571d61
Reviewed-on: https://chromium-review.googlesource.com/225642Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

This will allow us to show more extended error info in about:gpu
in Chromium.

BUG=angleproject:966
BUG=515229

Change-Id: I24c8f33ac3576395ff13dbf7f4664c649af18e33
Reviewed-on: https://chromium-review.googlesource.com/289590Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

This reverts commit 0018c85e.

Change-Id: Ieab61c0a4cd1b24d606263e9bf2bae31ea2e8ac4
Reviewed-on: https://chromium-review.googlesource.com/289486Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

to be consistent with how we start vertex attributes. A null pointer
indicates we're using the streaming buffer.

Will also aid the dirty state bits refactor.

BUG=angleproject:1040
TEST=WebGL CTS, end2end_tests, unittests

Change-Id: I7a9167282b60dbe0cbb9f0e5d9e3770890ffeb71
Reviewed-on: https://chromium-review.googlesource.com/284619Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Causing failures in the WebGL1 test: conformance/glsl/misc/shader-with-non-reserved-words.html

This reverts commit b5f88853.

Change-Id: I2105c8040057665abda00435e8c0ff8a83af3645
Reviewed-on: https://chromium-review.googlesource.com/289192Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1111

Change-Id: I0c71b765e8589a7387f4fd40f9dc453ed06c05b9
Reviewed-on: https://chromium-review.googlesource.com/289591Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

This test is having flaky timeouts on the bots. Disable it for now.
The re-land fixes the test syntax, and ensures we will see any errors
in test config as test failures on the bots.

BUG=angleproject:1107
TEST=angle_deqp_gles2_tests runs

Change-Id: Ia4fa8affb70e4474d8e43ff87ee86328aa47ebeb
Reviewed-on: https://chromium-review.googlesource.com/289580Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Our validation wouldn't correctly handle single-element arrays because it
was checking size > 1 instead of isArray().  This caused issues in
ProgramGL on some drivers that optimized the array size (AMD) if some elements
were unused.

BUG=angleproject:882

Change-Id: I417d13cd86380e2c6caa688f6398709a74692e21
Reviewed-on: https://chromium-review.googlesource.com/289201Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

With ESSL 3.00, names of built-in functions cannot be redeclared as
functions and therefore overloading or redefining builtin functions
is an error.
This is fixed by inserting unmangled built-ins into the symbol table
and then checking if the new function declaration matches any of the
built-in in symbol table.

BUG=angleproject:1066
TEST=angle_unittests(new: ParseESSLFunctionsTest), dEQP Tests
dEQP-GLES3.functional.shaders.functions.invalid.overload_builtin_function*
(2 tests started passing with this change)

Change-Id: I0e027e588664e604f29c130028178315c3e21631
Reviewed-on: https://chromium-review.googlesource.com/287801Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Incorrect syntax is making the bots skip all tests.

BUG=angleproject:1107

This reverts commit e8c8de2d.

Change-Id: I4469b6b320f8754af576875128e4e3fb210c67b6
Reviewed-on: https://chromium-review.googlesource.com/289501Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

These files were a relic of a previous structure, we can safely
remove them now that chrome uses angle_unittests.gypi.

BUG=None
TEST=compiles standalone and with chromium

Change-Id: If9a09122f664b740099b4270f5acc5627fc5ac35
Reviewed-on: https://chromium-review.googlesource.com/289343Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:884

Change-Id: Ia201ac60e337fd45ef72cc254bcb36d333418613
Reviewed-on: https://chromium-review.googlesource.com/289333Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1109

Change-Id: I551dbf79b98798f06fd5cbe1b005dc2410daaa6b
Reviewed-on: https://chromium-review.googlesource.com/289274Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

FBConfigs where missing some filtering that wasn't needed on NVIDIA:
the multisample and accumulation buffers must be the same as the
context's config otherwise glXMakeCurrent will fail. See
FBConfigCompatibility.md for more details.

BUG=angleproject:1109

Change-Id: Ied3336f9f5227594b143c884c4546fd603238b47
Reviewed-on: https://chromium-review.googlesource.com/289273Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

BUG=

Change-Id: Id5f98332129b3cc854254187b494bf6a97921dbd
Reviewed-on: https://chromium-review.googlesource.com/289272Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

Instead of trying to encapsulate the GL texture format workarounds into
tables, use functions that do manual checks for specific cases.  Simplifies
the logic.

Fixes:
 * conformance/extensions/ext-sRGB.html
 * conformance/extensions/oes-texture-half-float.html
 * conformance/extensions/oes-texture-half-float-with-canvas.html
 * conformance/extensions/oes-texture-half-float-with-image.htm
 * conformance/extensions/oes-texture-half-float-with-video.html

BUG=angleproject:884

Change-Id: Ifb719fff908680fddc7c53a544e2284a42a58356
Reviewed-on: https://chromium-review.googlesource.com/289082Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:892

Change-Id: I0999b0d3b043d412c092350f7d20560cb8b87ff7
Reviewed-on: https://chromium-review.googlesource.com/287450Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

On desktop GL, using an internal format of GL_RGBA produces GL_RGBA8
textures even if format is set to GL_FLOAT.  Use sized internal formats
whenever possible to ensure that the internal format of the texture is
the expected format.

Work around issues with Intel and AMD drivers that always round up when
writing to a 1-bit alpha framebuffer format.

Fixes:
 * conformance/extensions/oes-texture-float-with-image-data.htm
 * conformance/extensions/oes-texture-float-with-image.html
 * conformance/extensions/oes-texture-float.html
 * conformance/extensions/oes-texture-half-float-linear.html
 * conformance/extensions/oes-texture-half-float-with-image-data.html
 * conformance/extensions/oes-texture-half-float-with-image.html
 * conformance/extensions/oes-texture-half-float.html

Reland of https://chromium-review.googlesource.com/#/c/288350/

BUG=angleproject:884

Change-Id: Ia2f9251d12dd3bbc7b6918d46e61623d0abedfac
Reviewed-on: https://chromium-review.googlesource.com/288584Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

This test that using glDrawElements with an offset and an index buffer
works correctly with or without a static copy of the buffer in D3D.

BUG=510585

Change-Id: I4574a4058cc803e2ee0306e5bac52b6df6a71530
Reviewed-on: https://chromium-review.googlesource.com/288352Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

Previously, once a dynamic buffer had an associated static data, it would
be considered as static and the static data would be always be
reinitialized just after having been invalidated.

BUG=510585

Change-Id: I403e91d35d11efe17f52947b2182f7b8febd7922
Reviewed-on: https://chromium-review.googlesource.com/288351Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:884

Change-Id: I9e48803f6150b9d8cd70a36ab9197f78fd01a50f
Reviewed-on: https://chromium-review.googlesource.com/288583Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

It is not referenced outside of the d3d classes.

BUG=angleproject:884

Change-Id: Ibeec034b1b78b569bea2f1c77d6a686b58b8e010
Reviewed-on: https://chromium-review.googlesource.com/288582Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

The std::map became a significant bottleneck for draw calls.

BUG=angleproject:959

Change-Id: I28176d1474fe5cff05018f89878ea8b2a3e3f9f1
Reviewed-on: https://chromium-review.googlesource.com/288702Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

This test is having flaky timeouts on the bots. Disable it for now.

BUG=angleproject:1107

Change-Id: Iaa15672fb96f4aa3463a0613980f8a941aeafe1f
Reviewed-on: https://chromium-review.googlesource.com/288405Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

This test is failing on the bots.

BUG=angleproject:1108

Change-Id: Iaef3f0810c5d076b6c49bc356340c9bde014c4f5
Reviewed-on: https://chromium-review.googlesource.com/288518Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Causing failures on the Intel and AMD bot.

This reverts commit 8d8044f7.

Change-Id: I7dc6634eaf6f57ad84c315d74aa74d3ef2d4016f
Reviewed-on: https://chromium-review.googlesource.com/288325Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

On desktop GL, using an internal format of GL_RGBA produces GL_RGBA8
textures even if format is set to GL_FLOAT.  Use sized internal formats
whenever possible to ensure that the internal format of the texture is
the expected format.

Fixes:
 * conformance/extensions/oes-texture-float-with-image-data.htm
 * conformance/extensions/oes-texture-float-with-image.html
 * conformance/extensions/oes-texture-float.html
 * conformance/extensions/oes-texture-half-float-linear.html
 * conformance/extensions/oes-texture-half-float-with-image-data.html
 * conformance/extensions/oes-texture-half-float-with-image.html
 * conformance/extensions/oes-texture-half-float.html

BUG=angleproject:884

Change-Id: I48f220a4da518fff5831155a45c04cafcdffd33b
Reviewed-on: https://chromium-review.googlesource.com/288350Tested-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=505308

Change-Id: I34ae3aa65ce55c4e04eecfef3c9cde9841e47140
Reviewed-on: https://chromium-review.googlesource.com/288245Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Daniel Cheng <dcheng@chromium.org>

Causing failures on the Intel and AMD bots which seem to have incorrect rounding for 1-bit alpha formats.

This reverts commit 0bcb68f3.

Change-Id: I00d333b5c10ba3f41d54bec95769f4baf07b0771
Reviewed-on: https://chromium-review.googlesource.com/288206Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

Using git cl format.

BUG=None

Change-Id: Ia0ce03ae8115f45c3bcc4b022acd7e3e88518062
Reviewed-on: https://chromium-review.googlesource.com/287833Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

This bug was exposed after a refactoring from the translator.

BUG=None
TEST=angle_unittests,angle_end2end_tests

Change-Id: I13fddcbe84f87826068a557f139f6e35c674571e
Reviewed-on: https://chromium-review.googlesource.com/287832Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

Relying on the driver to round in one direction or the other is not
reliable.

Change-Id: Ide96864607ee6a22d2523544d6cfe8fae3116fcd
Reviewed-on: https://chromium-review.googlesource.com/288289Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>