Commits · ce8bb2fa5827323fbf3107d81907dcbcbbdbd93f · Chen Yisong / angle

05 Jan, 2015 4 commits

Improve standards conformance of ANGLE's testing code. · ce8bb2fa

authored Dec 30, 2014

ANGLE's testing code recently got enabled in Chromium's builds. While it
builds fine with cl.exe, it isn't quite standards-conformant and doesn't
build with clang. Fix this.

There were three issues:

1. ANGLE_TYPED_TEST_CASE() is a variadic macro that tries to use
__VA_ARGS__ as argument to a variadic template and then pass
that template to another macro. However, [cpp.replace] describes
that ANGLE_TYPED_TEST_CASE(Test, int, float) should be expanded
to TYPED_TEST_CASE(Test, ::testing::Types<int, float>) which
should be interpreted as a "call" of TYPED_TEST_CASE with the
3 macro arguments `Test`, `::testing::Types<int`, and `float>`.
As a fix, use a typedef for the variadic template and refer to
it through the typedef in the macro call.

2. `#version` was used on its own line in a substitution of the
SHADER_SOURCE macro. [cpp]p1 says that every line starting with
a `#` is a preprocessing directive, and [cpp.replace]p11 says
"If there are sequences of preprocessing tokens within the list
of arguments that would otherwise act as preprocessing directives,
the behavior is undefined" (with a footnote that this includes
non-directives -- # followed by unknown text). As a fix, merge the
`#version` line with the previous line. Now the line doesn't start
with `#` and things are fine.

3. Unqualified lookup usually doesn't look into dependent bases. If
this is desired, one usually has to make the call qualified, a
good explanation for this is at
http://eli.thegreenplace.net/2012/02/06/dependent-name-lookup-for-c-templates
cl.exe doesn't implement this rule, and clang tries to emulate
cl.exe's behavior to a certain extend when in Microsoft mode.
However, that seems to not work for member templates with explicit
types (filed http://llvm.org/PR22066 for this, but since it's not
needed to parse Microsoft headers and not standards-conformant,
I'm not sure if we'll fix that). As a fix, don't provide an explicit
type, the inferred type is the same. This is also consistent with
all the other tests in this file. (We might clean up -Wmicrosoft
warnings in the future; if so I'll add the explicit this->s that
are missing in this file when we do.)

BUG=chromium:445406

Change-Id: I77a2f3ab9601a1f0f39b56ed3d05217f123155b8
Reviewed-on: https://chromium-review.googlesource.com/238090Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

ce8bb2fa

Fix DisplayD3D not initializing mRenderer. · 5b072780

authored Dec 11, 2014

BUG=angle:658

Change-Id: If865910d52154f574f4d488d4c45fcac9a87b050
Reviewed-on: https://chromium-review.googlesource.com/234761Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

5b072780

Add format size information for DXGI_FORMATs. · 932b465b

authored Dec 17, 2014

Change-Id: I14d0b8ebd631e5c9c3c3b63c047c9ef7c11bf310
Reviewed-on: https://chromium-review.googlesource.com/236304Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

932b465b

Add format size information for D3DFORMATs. · eca8561e

authored Dec 17, 2014

Change-Id: Id8d884df31726546e871d75a719bc419dd767d6b
Reviewed-on: https://chromium-review.googlesource.com/236303Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

eca8561e

29 Dec, 2014 2 commits

Fix a translator crash with index expressions. · 21c1e456

authored Dec 29, 2014

This crash happened with certain bad shaders which used temporary
values as array (or other) index expresisons.

Fixes the crash covered in the WebGL test
"conformance/bugs/undefined-index-should-not-crash"

BUG=angle:857

Change-Id: I13e2ba6d5f1ab0846ac902021bc0b57cbb37d759
Reviewed-on: https://chromium-review.googlesource.com/237460Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

21c1e456

Enable conditional InfoLog for HLSL link failures. · 97399235

authored Dec 23, 2014

We don't always have an InfoLog available for dynamic shaders, but we
can still capture data in some cases for default shaders.

Change-Id: Iccd022f24d5c2b8922e9254580705675cd4e405f
Reviewed-on: https://chromium-review.googlesource.com/237328Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>

97399235

23 Dec, 2014 3 commits

Implement support for the binary operator '%' in the translator. · be954a23

authored Dec 23, 2014

BUG=angle:854

Change-Id: If116de132dc83d93255749b54c1919a75abcb65c
Reviewed-on: https://chromium-review.googlesource.com/236330Tested-by: Gregoire Payen de La Garanderie <Gregory.Payen@imgtec.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

be954a23

Update WebGL expectations. · be12fa22

authored Dec 23, 2014

Change-Id: I69795b97ef032838cd43e889119405aaee66fa4f
Reviewed-on: https://chromium-review.googlesource.com/237329Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

be12fa22

Disable FL9_3 tests because of Windows 7 failures. · 48825bea

authored Dec 23, 2014

These tests were failing on SwapChain11 with errors relating to
the Depth Stencil backbuffer.

BUG=angle:856
BUG=435726

Change-Id: Idd371a380e9b302d7d5e40437e9d76589872b750
Reviewed-on: https://chromium-review.googlesource.com/237450Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>

48825bea

22 Dec, 2014 3 commits

Update ANGLE's translator to Bison 3. · 185de884

authored Dec 22, 2014

BUG=angle:462

Change-Id: I2c1c18027dee1c3b4efb87374caaadbf58367841
Reviewed-on: https://chromium-review.googlesource.com/236930Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

185de884

Support remaining vertex attrib formats on D3D11 9_3 · d5663b1c

authored Dec 22, 2014

Change-Id: Id0d3ad6ec7764ed84ec504ad92eca81dff29aac8
Reviewed-on: https://chromium-review.googlesource.com/236043Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

d5663b1c

Prevent usage of GSSetConstantBuffers on D3D11 Feature Level 9_3 · 366db7cd

authored Dec 16, 2014

Calling GSSetConstantBuffers with real data causes a device removed
error on 9_3. We therefore ensure that ANGLE doesn't call it on 9_3.

Change-Id: I151a74ee7ea81fc4b795d8bf10d5e5c42bfe7c86
Reviewed-on: https://chromium-review.googlesource.com/236042Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

366db7cd

19 Dec, 2014 1 commit

Refactored test targets into angle_unittests and angle_end2end_tests. · 2609bf4c

authored Dec 18, 2014

angle_unittests subsumes the previous:
  angle_compiler_tests
  angle_preprocessor_tests
and will subsume angle_implementation_unit_tests in a follow-on CL.

angle_end2end_tests subsumes the previous:
  angle_tests
  angle_standalone_tests
  angle_implementation_unit_tests

The bulk of these two targets have been factored into their own
.gypi files, so that a different main.cpp can be trivially
plugged in for the gtest harness. This refactoring has been
tested both in ANGLE standalone builds, and builds within
Chromium.

BUG=chromium:435726

Change-Id: I231a3c1989c17e188cc469dcf80fe78b052afe78
Reviewed-on: https://chromium-review.googlesource.com/236681Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Kenneth Russell <kbr@chromium.org>

2609bf4c

18 Dec, 2014 3 commits

Move blit to the Framebuffer object and Impl. · 54bd5a46

authored Dec 01, 2014

BUG=angle:841

Change-Id: I482e53a90606d9d6b105c7006234215d51ab1a6b
Reviewed-on: https://chromium-review.googlesource.com/232692Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

54bd5a46

Move ReadPixels to the Framebuffer object and Impl. · bce529e4

authored Dec 01, 2014

BUG=angle:841

Change-Id: I71deac9e755b5dfa010596cd1f8a213c24d895bf
Reviewed-on: https://chromium-review.googlesource.com/232691Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

bce529e4

Move Framebuffer clearing from Renderer to the Framebuffer object. · b04dc82e

authored Dec 01, 2014

BUG=angle:841

Change-Id: I95c9cbdc2d1c99731e19c48e18117358d22b9e94
Reviewed-on: https://chromium-review.googlesource.com/232690Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

b04dc82e

16 Dec, 2014 1 commit

Fix normalized GL_BYTE vertex attributes on D3D11 9_3 · c9878cb1

authored Dec 11, 2014

Feature Level 9_3 doesn't support as many formats for Input Layouts
as 10_0+. On 9_3, we have to make sure that GL vertex attributes are
converted into formats that 9_3 supports.

Change-Id: I27b9a85a6eb21a37bd36e60bf011b83fce743fd0
Reviewed-on: https://chromium-review.googlesource.com/234523Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

c9878cb1

15 Dec, 2014 1 commit

Unify duplicated code for both depth and stencil buffer in ValidateBlitFramebufferParameters() · 44b422c1

authored Dec 09, 2014

Tested with angle_*_tests and WebGL CTS in Chrome Canary on Windows. Passed with no regressions.

Change-Id: Ied0a32ea75565ef19b8e87bed8bdd555b74edee3
Reviewed-on: https://chromium-review.googlesource.com/234110Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Dongseong Hwang <dongseong.hwang@intel.com>
Reviewed-by: Kenneth Russell <kbr@chromium.org>

44b422c1

12 Dec, 2014 2 commits

Link libANGLE against translator_static. · 4ed1c450

authored Dec 12, 2014

If there are multiple users of translator.dll then one could ShFinalize
and destroy the memory of all the others. libANGLE needs it's own copies
of the global pool allocator to make sure this doesn't happen.

There should be no functional changes for non-component builds.

BUG=angle:823

Change-Id: Id6f7d50e86d8950834399b16a7ab14e6e3198d12
Reviewed-on: https://chromium-review.googlesource.com/234594Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

4ed1c450

Prevent D3D11 Feature Level 9_3 from sampling from SV_Position in Pixel Shaders · 588434c4

authored Dec 10, 2014

D3D11 FL9_3, like Shader Model 2.0 in D3D9, doesn't support reading from SV_Position in the pixel shader. We have to reconstruct gl_FragCoord using dx_ViewCoords instead.

Change-Id: I7e898038d210d73a9d224dcc18b033e5cd4a56f5
Reviewed-on: https://chromium-review.googlesource.com/234277Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

588434c4

11 Dec, 2014 3 commits

Encapulate the ESSL compiler into a GL object that is per-context. · 492a7e43

authored Nov 05, 2014

 * Allows for multiple contexts have to have different client versions, caps
   and extensions without causing shader compilation failures.

BUG=angle:823

Change-Id: I523679e90be031b0b7fa385d46d6839b1cf3029f
Reviewed-on: https://chromium-review.googlesource.com/227710Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

492a7e43

Move some structs and enums from Renderer.h to RendererD3D.h · 797ff4c0

authored Dec 11, 2014

Change-Id: Ib334d2506031466a73f1a233ee68c1658841a882
Reviewed-on: https://chromium-review.googlesource.com/234720Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

797ff4c0

Remove D3D11_BIND_STREAM_OUTPUT usage on D3D11 Feature Level 9_3 · 62588f5b

authored Dec 10, 2014

Change-Id: Ib9d2d155a6e062488ff5bb5ce578036b6187ca70
Reviewed-on: https://chromium-review.googlesource.com/234276Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

62588f5b

10 Dec, 2014 3 commits

In Clear11, use ID3D11DeviceContext1::ClearView when appropriate · 53c1a456

authored Dec 05, 2014

Change-Id: I8d9054b5582cfe78beaff07146adc73b4716eaf9
Reviewed-on: https://chromium-review.googlesource.com/233680Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

53c1a456

Fix cube map rendertargets. · eb32a2e9

authored Dec 10, 2014

We were using the entire level parameter of the cube map, instead
of using the helper functions for cube map ImageIndexes, which
set the layer corresponding to the cube map face.

BUG=angle:849
BUG=440701

Change-Id: Id78db5c8281b6b644392bb961d69a7f869755a34
Reviewed-on: https://chromium-review.googlesource.com/234380Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>

eb32a2e9

On D3D11 Feature Level 9, use IDXGIAdapter2::GetDesc2. · b1237ed7

authored Dec 08, 2014

On Feature Level 9_3, GetDesc() returns "Software Adapter" in the hardware
description string, instead of a description of the actual GPU. As a
result, glGetString(GL_RENDERER) on 9_3 wouldn't return any useful info
about the current hardware.

Change-Id: Ie4d329a08a03ee044fc0b594308aaf0bf235c74e
Reviewed-on: https://chromium-review.googlesource.com/233663Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

b1237ed7

09 Dec, 2014 1 commit

Add basic D3D11 Feature Level 9.3 support · 0dbda057

authored Dec 04, 2014

Change-Id: I660c74791ddb9917bd4796bb652eefab2a3e4863
Reviewed-on: https://chromium-review.googlesource.com/233660Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

0dbda057

08 Dec, 2014 3 commits

Rename Framebuffer::completeness to checkStatus and add an Impl method. · 748f74ef

authored Dec 01, 2014

BUG=angle:841

Change-Id: I04b4ffd086424569a15aa21447dd552e0a898928
Reviewed-on: https://chromium-review.googlesource.com/232394Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

748f74ef

Add set*Attachment methods to FramebufferImpl. · 9dd95808

authored Dec 01, 2014

BUG=angle:841

Change-Id: Ie819c253a900d105d768aee168a6a2de89754ccc
Reviewed-on: https://chromium-review.googlesource.com/232393Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

9dd95808

Update generated D3D shaders. · f1b85f31

authored Dec 08, 2014

We neglected to regenerate with our switch to the 8.1 SDK.

Change-Id: Idd67dce1147ad0acec9bf933a0fd01f421c784ef
Reviewed-on: https://chromium-review.googlesource.com/233844Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

f1b85f31

05 Dec, 2014 10 commits

Move libEGL to compile on all platforms. · 16545669

authored Dec 05, 2014

BUG=angle:773

Change-Id: I377c6df9b1f10e2756b7f000099847dc910c965c
Reviewed-on: https://chromium-review.googlesource.com/232793Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

16545669

Allow implementation tests to compile on all platforms. · 1a05ce93

authored Dec 05, 2014

BUG=angle:773

Change-Id: I9fb56c1d8332b1529460eb3962bae1b756f58e91
Reviewed-on: https://chromium-review.googlesource.com/232791Reviewed-by: Brandon Jones <bajones@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>

1a05ce93

Move the Image::copy implementation to ImageD3D. · 62a4be24

authored Dec 05, 2014

This method uses rx::RenderTarget, which is D3D-only.

BUG=angle:773

Change-Id: I2cb523fa98bda96661260d37a8c69c2ed8e7aa84
Reviewed-on: https://chromium-review.googlesource.com/233441Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

62a4be24

Move GetAttachmentRenderTarget and GetAttachmentSerial to FramebufferD3D.h. · b5d8f23e

authored Dec 04, 2014

BUG=angle:841

Change-Id: Ic341239e848f1daf7cf38f233d79a58f221780f4
Reviewed-on: https://chromium-review.googlesource.com/232969Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

b5d8f23e

Move some D3D-specific stuff into ProgramD3D. · 437d2662

authored Dec 05, 2014

BUG=angle:773

Change-Id: I48b42e7a3e82a43d3dde16a8d1016d28280eae39
Reviewed-on: https://chromium-review.googlesource.com/232968Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

437d2662

Fix texture format support queries on AMD-D3D9. · 8e7c8a13

authored Dec 04, 2014

We were allowing calls to the device to check support for D3DFMT_NULL,
which AMD reports as a usable texture type. This was causing crashes
with the WebGL CTS and angle_end2end_tests, when the user tries to use
an sRGB or other type, which we incorrectly reported as available.

BUG=angle:839

Change-Id: I4941baac6a2b14a09c0ad2c924f29189ccfc6a07
Reviewed-on: https://chromium-review.googlesource.com/232728Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

8e7c8a13

Mark the exported entry points as stdcall. · 130e6e8f

authored Dec 05, 2014

Because they may be returned as stdcall from eglGetProc address, the entry
point functions must be stdcall.

Fixes issues on 32-bit builds.

BUG=angle:733

Change-Id: Ia20046a67fdc28c702837ec59b8299c6bb722006
Reviewed-on: https://chromium-review.googlesource.com/233471Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

130e6e8f

Have libGLESv2 compile tls.cpp. · babd8caf

authored Dec 05, 2014

BUG=angle:733

Change-Id: Ief3ddabdc441eb7e46581337a511cec2a90ede48
Reviewed-on: https://chromium-review.googlesource.com/233431Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

babd8caf

Use the static CRT. · 8300538f

authored Dec 03, 2014

Remove a test that expects a crash, it has become flaky.

BUG=angle:733

Change-Id: Ia1d4e0b3dd09fa755f678b97b8deceee3ef0a35b
Reviewed-on: https://chromium-review.googlesource.com/232963Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

8300538f

Implement all entry points in libGLES and have libEGL act as a shim. · e7c6e43c

authored Dec 03, 2014

This allows libANGLE to only be included in libGLESv2 and moves all TLS
data to libGLESv2.dll.

BUG=angle:733

Change-Id: I34f0b47987a5efbe906c290d3ca656142e69ea9a
Reviewed-on: https://chromium-review.googlesource.com/232962Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

e7c6e43c