@@ LAYOUT-xx @@ and @@ QUALIFIER-xx @@ macros are generated by the
compiler when emitting Vulkan GLSL.  These macros are replaced at link
time in the Vulkan backend.

Previously, this replacement was done through calls to
angle::ReplaceSubstring, reiterating over the whole source on every
replacement.  This CL does a prepass on the input source and chunks it
up in blocks.  Search is optimized as only blocks of a certain type are
string-compared (skipping large chunks of shader text).  Replace is
optimized as the whole shader is not shifted left or right on every
replacement.

Additionally, this CL modifies the layout macro to the following format:

	@@ LAYOUT-xx(extra, args) @@

This is used in a follow up CL to have the compiler provide additional
layout qualifiers.

Bug: angleproject:3220
Change-Id: I6367e781c3304d5f2e0a406e4fb4e6feb4c45f1d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592070
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

If an image was used by a different context previously, it's stored serial
is not relevent.  We always know that any work involving the swap chain image
has been completed by the time we aquire a new one.

BUG=angleproject:2464

Change-Id: I0b12b9ff4b9a68593d661344d2d1e2c9d94e4755
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592043Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

`out` variables in the vertex shader and `in` variables in the fragment
shader were not decorated with `layout(location=?)` as they should
according to the Vulkan GLSL spec.  This change makes sure these
decorations are present regardless of whether the compiler sees these
variables as EvqVaryingIn/Out or EvqVertexOut/FragmentIn.

Bug: angleproject:3412
Change-Id: I66473e876cb989a60b0c6d6a5850a8b2c763d8e2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1590694
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This CL adds a command line option to angle_end2end_tests that will
iterate over all test configs. For each config it'll fork a new child
process that will run only a single config. This will allow us to
isolate each config to a specific child process. Hopefully this will
reduce test flakiness due to driver issues with multiple configs.

The command line option is "--separate-process-per-config".

Note that there are about 25 configs right now.

Bug: angleproject:3393
Change-Id: Ia117b371bbe159c1b0d28d82befffeb0f40467a9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1591428
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

By updating YYLTYPE_IS_TRIVIAL to true, YYSTACK_RELOCATE can be enabled, so
we can dynamically allocate stack size based on needs.

BUG=angleproject:3028

Change-Id: I1b9cc0768cd0c6955589695e703595c56d43f24a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1593954
Commit-Queue: Yi Xu <yiyix@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Consolidate these tests into ANGLETest. Add a new parameter to
PlatformParameters that lets us skip normal test setup.

This removes the last 'configless' tests from angle_end2end_tests.

Bug: angleproject:3393
Change-Id: I87186698ade90f95577534eb8ed1dfd4245f740e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1590467
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This allows us to run angle_end2end_tests with a single config without
using gtest_filter. It will also allow us to run each test config in
a separate process.

Bug: angleproject:3393
Change-Id: I09aaf9cfe55a117b0af8d79ecfd129f3d0f1d7c8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1591427
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

If the shader contains code such as the following:

    struct S {
        vec4 field;
    };
    out S varStruct;

The layout qualifier macro is defined as @@ LAYOUT-varStruct @@.
However, the Vulkan backend was replacing @@ LAYOUT-field @@.

Bug: angleproject:3220
Change-Id: Iae15003867e0bed2cc939159a6653429c7a431e0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1571389Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>

Moving the Platform Methods init to GetDisplay allows us to override
the platform methods as soon as possible. This lets Chromium override
the methods via ANGLEGetDisplayPlatform. It also lets us reset the
platform methods when making a new display. Note that because the
methods pointer is global it's a bit tricky to manage between swapping
configs in ANGLE tests. Using a separate test process for each config
should help.

Bug: angleproject:3393
Change-Id: I351f86422e53075563f6d8fce32efbd40023cd23
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592045
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

These results are somewhat redundant. Don't need to be shown in the
public doc.

Bug: angleproject:3369
Change-Id: I4994cd5714e727afbfe63fb082dab074ed137d43
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1593959Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>

This helper was redundant with the image utils library. Switch the test
to use the existing methods. Doesn't affect test correctness.

This removes more 'configless' tests from angle_end2end_tests.

Bug: angleproject:3393
Change-Id: I49cfbd95c3713b2b4b25c4716a7a279d84ddf994
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1591426Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Add support for GCC Compilers, so it can be build on Ubuntu.

BUG=angleproject:3392

Change-Id: I2f5ac7355a40f52123bd465f86b1e966c59c943d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592062Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Yi Xu <yiyix@chromium.org>

We don't need to use upper case. Lower case can be more readable.

Skip a test on Linux/NV/Vulkan that became flaky after changing
the test order.

Bug: angleproject:3393
Change-Id: Ie027e08ddc9d47b9484cf5b1aaca72a05dd60517
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1574676
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools.git/+log/c8b09744c6a1..5fc5303eeceb


Created with:
  gclient setdep -r ./third_party/spirv-tools/src@5fc5303eeceb

The AutoRoll server is located here: https://autoroll.skia.org/r/spirv-tools-angle-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=jmadill@google.com

Change-Id: I9885ba82ef201582c2b0496ba49438d7ce12192b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1594474Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

* Removes D3D11 Warp and Reference configs from tests.
* Removes several permutations of OpenGL back-end specific configs.
* Removes FL 9_3 since it is no longer supported.
* Removes present path "Copy" since it's redundant with normal D3D11.

Reduces number of configs from over 60 to 25.

Bug: angleproject:3393
Change-Id: Ia5a23de3c4865b17ee50673a4066757b901a4b5a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1574675
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This test was failing due to patch errors. This could be because the
bot uses DEPS ANGLE instead of ToT. Removing the bot because it was
causing valid CLs to fail on the CQ.

Bug: angleproject:3146
Change-Id: I7b8b120160fe435832cba919c256ce185ade4d90
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592602Reviewed-by: Jamie Madill <jmadill@chromium.org>

https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools.git/+log/32af42616abe..c8b09744c6a1


Created with:
  gclient setdep -r ./third_party/spirv-tools/src@c8b09744c6a1

The AutoRoll server is located here: https://autoroll.skia.org/r/spirv-tools-angle-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=jmadill@google.com

Change-Id: Ibb2108c988ea4b27108da68498b5a51e72ff0f0c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1593015Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

This CL moves all of the EGLDisplay configuration into an EGL platform
struct. Consolidating display configuration in the struct allows us to
move configuration out of the test constructor. Then when we filter
test configs we don't need to wait for the individual test setup.

Bug: angleproject:3393
Change-Id: I5bd06dcdc9f2867ebc43c1d4984077ada35cafc8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1574674
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Context has the invariant that if it is not current anywhere, mCurrentSuface is
null.  If makeCurrent fails in ContextImpl::onMakeCurrent, the surface is not
reset, breaking this invariant.

BUG=angleproject:3414

Change-Id: Icb13c8e2d1edd98d87a8af091a3fbacebadb395c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592064Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Also add a missing arrow from VertexArray to Context.

Bug: angleproject:3002
Change-Id: I2a30df0bb1c9681b4cf5178b75f1264ddf4ca2e2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592065Reviewed-by: Jamie Madill <jmadill@chromium.org>

The diagram source is available as a Google Driver document. It can be
opened with the draw.io web app.

Bug: angleproject:3002
Change-Id: Ic7885c8da1aacac5082eea05d0562d3d62f801ba
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592252Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This was previously returning maxCombinedTextureImageUnits instead of
maxCombinedUniformBlocks.  Fixing that exposed a bug in the GL backend
where the combined values were sometimes capped to vertex+fragment
values and sometimes not capped at all.

The reasoning for such capping was that the combined limits as queried
from GL contains stages not available in GLES (such as tessellation).
However, the capping failed to take geometry shader limits into account.

This change adjusts such capping to vertex+fragment+geometry.  It also
applies the cappping to all combined limits, rather than a select few.

Bug: angleproject:2099
Change-Id: I7231058b5d7f80b1b2452d9f87d4b0ab6e1cdb17
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1572487Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>

None of these functions needed to be member functions. Also make the
naming more consistent.

Bug: angleproject:3393
Change-Id: I7aafe2269a48af703a87bd9a8cf4cfab9e177dd3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1574673Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Also clean up the comments formatting for generate_stats.py

Bug: angleproject:3369
Change-Id: Ifaa26cb898915038cf47e88341db5057b59dd937
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1588315
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

The swap interval can be changed independent of the Surface config.
Thus it makes more sense to set it explicitly in test setup. This
simplifies the test config.

Also updates some of the API for GLWindowBase. Return an explicit
error from makeCurrent.

Bug: angleproject:3393
Change-Id: Ic62b33018e872bc0e38f2848e2427ed898b60749
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1574672Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Required a few build tweaks because GN check does not understand
#ifdef includes.

Bug: angleproject:3411
Change-Id: I1aa7b97626767e3dc973bfa55f5b2a511c806ec9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1589148Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The the user hasn't called SwapBuffers explicitly, we should not write to
the window surface.  This is especially a problem when the native window has
already been destroyed.

BUG=angleproject:2464

Change-Id: Ib4289de471ba2ad10e5cc21a8c2af946642f138c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1586355Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools.git/+log/736376dbf989..32af42616abe


Created with:
  gclient setdep -r ./third_party/spirv-tools/src@32af42616abe

The AutoRoll server is located here: https://autoroll.skia.org/r/spirv-tools-angle-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=jmadill@google.com

Change-Id: I89036ebd9c06aeb79b7b5b358636f5adb4176548
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1589164Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

Also modify or remove some tests to sync up with the expected behavior
stated in spec.

Related to https://github.com/KhronosGroup/WebGL/pull/2780
 If any draw buffer with an attachment does not have a defined fragment shader output,
 draws generate INVALID_OPERATION.

Also remove Framebuffer masking for inactive outputs.

This workaround is no longer necessary as the WebGL spec has changed.
It also was never fully working and had bugs with certain orders of
calls.

Bug: angleproject:2872
Bug: chromium:927908
Bug: chromium:943538
Change-Id: I73715a6ab851ae3db7096f49ea0a9fdd6f576703
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1530018
Commit-Queue: Shrek Shao <shrekshao@google.com>
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>

If vkAcquireNextImageKHR fails, the aquireImageSemaphore would be leaked,
wrap it in a vk::Scoped so it is always cleaned up.

BUG=angleproject:2464

Change-Id: Ic4a0e16c89ea7c35cf060e5601760422e673c080
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1585318Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

This allows contexts to flush and update state when they are no longer
going to be current.

BUG=angleproject:2464

Change-Id: Ie577475a94090631a0208542b32a12a239bdeb75
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1553824Reviewed-by: Geoff Lang <geofflang@chromium.org>

The lists of tests can't be added to the sheets if more than 50,000
lines. Since the goal is to get rid of all unexpected passes/failures,
it's reasonable to clip the lists by length. Other formatting changes
too.

Bug: angleproject:3398
Change-Id: Ice461862fb4266e0ad1280f20de85224fc6d7e97
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1585612Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>

A nearly direct port of the https://www.khronos.org/registry/webgl/sdk/tests/conformance2/rendering/framebuffer-texture-changing-base-level.html
test for easier debugging.

BUG=angleproject:2291

Change-Id: Ibc8a7e77629f1ea3f15e92ea93f1210fa850c524
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1584361
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shrek Shao <shrekshao@google.com>

The stats generation should pick up a list of unexpected failed/passed
tests, so the tests should be printed with the format used by this
script.

Bug: angleproject:3398
Change-Id: I96a8eba6c2e294f64ca59ec60042cc688d1133c5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1584363Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>

KHR_robustness says that certain queries that poll such as GetSynciv with
SYNC_STATUS should still generate errors after context loss but also return
that the result is available.

BUG=angleproject:3379

Change-Id: Ibf61f8481bc7c3d1b4fa8979dc404d0ca3952fcd
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1566142
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Add more explicit comments and change some behavior slightly.

Bug: angleproject:3369
Change-Id: I76f9054564e8d33127e8ff8233b317e82e06c7de
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1582760Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>

TextureD3D_Cube::isFaceLevelComplete skips checks for the base level
because it assumes that isCubeComplete has already been called to
verify the base level.  This assumption is not correct for some cases
like glCopyTextureCHROMIUM which attempts to create a render target
from cube maps that may not be complete.

This caused us to create render targets from incomplete cube maps and
then fail to upload data to the cube faces because the sizes or formats
did not match between the images and render target.

TEST=conformance/more/conformance/quickCheckAPI-S_V.html

BUG=angleproject:3384

Change-Id: Iceb9415debfa07436fea1eb4bc23d0ec05679839
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1578957
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>

D3D11 cannot allow the same (sub)-resource bound as both
a SRV and an UAV at the same time. If render pipline and
compute pipeline access the same texture, must unset
conflicting SRVs and UAVs before drawing and dispatching
on D3D backend. Currently, these cases could not pass
on D3D11 backend.

Bug: angleproject:3152
TEST=angle_end2end_tests.ComputeShaderTest.*

Change-Id: I3ed19d7586625aac7fe9f46f86341d136d148019
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1578131Reviewed-by: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Xinghua Cao <xinghua.cao@intel.com>

Add target for angle_white_box_perftests to DevSetup.md Android build
instructions. Generously comment the test source to clarify various
tests and how to run them.

Bug: angleproject:3136
Change-Id: Ie4f0ec6951d354ebac861f727c9d550a93ad48c9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1577740Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Tobin Ehlis <tobine@google.com>

The StreamDraw usage flag was changed to a static type in
http://crrev.com/50eb1e6f77b3b0ad89d2cd00efc0d1d673365296 but it appears to
have regressed performance in several Chrome benchmarks because
Skia uses it as a usage flag for buffers that change data frequently.

BUG=angleproject:3366
BUG=953788
BUG=954028

Change-Id: Ib6805d94714bcd1679f3c9c49f1b70574f0b658a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1582786Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Commit-Queue: Geoff Lang <geofflang@chromium.org>