Commits · f4126d2b9eb008d4e792fff2592f338e19c68569 · Chen Yisong / angle

14 Aug, 2014 1 commit

Fix nameless struct handling. · f4126d2b

authored Aug 11, 2014

BUG=401296

The addition of layout qualifier handling in es3 development inadvertently
broke handling of nameless structs.

Change-Id: I805bab7a981d1f7f6227ae043720296fc3454662
Reviewed-on: https://chromium-review.googlesource.com/211860Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Shannon Woods <shannonwoods@chromium.org>

Conflicts:
	src/compiler/translator/UniformHLSL.cpp
Change-Id: I2f9fca71cef383c1a48c9bcdbe1083ac3222fac0
Reviewed-on: https://chromium-review.googlesource.com/212153Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Shannon Woods <shannonwoods@chromium.org>

f4126d2b

11 Aug, 2014 1 commit

Avoid D3D11 primitive restart index using 32-bit indices · 75551cf9

authored Aug 11, 2014

D3D11 interprets an index value of 0xFFFF in a 16-bit index buffer as a
primitive restart marker. This behavior can't be toggled off.
http://msdn.microsoft.com/en-us/library/windows/desktop/bb205124(v=vs.85).aspx
We work around it by converting to 32-bit indices.

BUG=angle:708

Change-Id: Ibc92d6ba98e5f11a98d76cae14f90ca050a19964
Reviewed-on: https://chromium-review.googlesource.com/211426Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Nicolas Capens <capn@chromium.org>

75551cf9

25 Jul, 2014 1 commit

Use commit_id.py on Windows, and handle missing git. · c333af9c

authored Jun 12, 2014

This allows us to delete the Windows batch file.

Changes the commit_id script to take the working directory so that it
can be called from a different working directory than the angle
repository is in.

Renames the generated commit header to angle_commit.h. This is being
written to the shared generated code directory for the entire build,
and "commit.h" is insufficiently unique.

BUG=angle:669
BUG=395705

Change-Id: I35e80411a7e8ba1e02ce3f6a4fc54ed4dbc918f3
Reviewed-on: https://chromium-review.googlesource.com/209617Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

c333af9c

14 Jul, 2014 3 commits

Fix HLSL compiler error with else-rewriting in functions. · ebba7d30

authored Jul 07, 2014

In functions with return types where we would use if-else rewriting,
we would potentially generate a spurious HLSL error that warned of
branches with no return value in the function.

This was causing a maps regression where overlays would not draw
in Earth mode.

BUG=346463
BUG=391697

Change-Id: I9f4fa959057a3a2dab6cdd98f8381b5871cabf03
Reviewed-on: https://chromium-review.googlesource.com/206824Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <capn@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/207834

ebba7d30

Add a new TIntermRaw node type to translator. · 9093114b

authored Jul 07, 2014

This raw node stores text strings that we directly copy to the
output. This allows for more tricky substitutions that don't fit
in to the HLSL/GLSL shared parsing model.

BUG=346463
BUG=391697

Change-Id: Ibbde6db4fc98ef6d892f219631ca1a258a902a86
Reviewed-on: https://chromium-review.googlesource.com/206823Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <capn@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/207833

9093114b

Fix not rewriting else-if blocks. · 5ec99a79

authored Feb 26, 2014

We would miss expanding the else-if clauses in a chain
of selection statements.

BUG=346463
BUG=391697

Change-Id: Iee644b875cf68d0ed3dc0b73542efc49ecb23242
Reviewed-on: https://chromium-review.googlesource.com/206822Reviewed-by: Nicolas Capens <capn@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/207832

5ec99a79

09 Jul, 2014 3 commits

Fix npot Texture level 0 validation. · b3319dfe

authored Jun 23, 2014

We were rejecting npot textures for level 0, even though it is
valid in GLES 2 for non-mipped textures to have npot size.

BUG=381495

Change-Id: Iacc3ab50d6487ecba804fd8963aa0a2ada2f1cae
Reviewed-on: https://chromium-review.googlesource.com/205220Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Change-Id: Iad3f5d007e878e60a87180bba3a2c1e246c311ca
Reviewed-on: https://chromium-review.googlesource.com/207121Reviewed-by: Nicolas Capens <capn@chromium.org>

b3319dfe

Restrict the size of shader arrays. · a1bb2d5a

authored Jun 06, 2014

This prevents overflow issues in the HLSL translator and some drivers. The
limit it hard-coded to 65536 to be larger than the Shader Model 5 register
limit (4096) to account for register allocation optimizations and future
hardware.

BUG=379799

Change-Id: I20f05c9dd230fbfc1c92af52533cd2b50f2ff994
Reviewed-on: https://chromium-review.googlesource.com/202939Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Nicolas Capens <nicolascapens@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/207123Reviewed-by: Nicolas Capens <capn@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

a1bb2d5a

Fix buffer overflow error on buffer resize. · ea77ce22

authored Jun 16, 2014

We were using the new, larger size for our buffer data copy step,
instead of the old, smaller size. This bug was causing a crash on
older nVidia drivers during normal browser usage.

Reproducible with the index-validation-large-buffer WebGL test.

BUG=angle:667
BUG=384420

Change-Id: I98ee893e0d8ba0bfc9adfe5a338da9b940248879
Reviewed-on: https://chromium-review.googlesource.com/203776Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/207122Reviewed-by: Nicolas Capens <capn@chromium.org>

ea77ce22

10 Jun, 2014 3 commits

Prevents multisample depthstencil from being bound as an SRV · afeeda93

authored Jun 06, 2014

BUG=381413

Multisample textures can only be bound as both depth stencil
views and shader resource views at feature level 10_1 and above.
We don't need to bind textures as SRVs if they're multisample
depth targets, so we should avoid the SRV flag in that case.

Change-Id: I28ca98bdc4a86eb5442b39ea3cf2b42aa9216aab
Reviewed-on: https://chromium-review.googlesource.com/202921Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Shannon Woods <shannonwoods@chromium.org>

afeeda93

Move FBO attachment code to new source file. · afc21c06

authored Jun 04, 2014

Since FBO attachments and Renderbuffers are semantically distinct,
make a new location for attachments to live.

BUG=angle:660

Change-Id: I51753f8a814e89641637c5d8293f7e9a573a7ba5
Reviewed-on: https://chromium-review.googlesource.com/201833Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

afc21c06

Rename Renderbuffer to FramebufferAttachment. · 3c7fa226

authored Jun 05, 2014

Part of the refactoring effort to clean up our classes for FBO
attachments and Renderbuffers.

BUG=angle:660

Change-Id: Id23df904f56499568159611b66c8922b9dce6a3d
Reviewed-on: https://chromium-review.googlesource.com/201832Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

3c7fa226

06 Jun, 2014 2 commits

Fix up some validation bugs in Texture functions. · 6f38f823

authored Jun 06, 2014

In several places we weren't handling certain errors, or were
returning the wrong error code. Several dEQP negative API
tests are affected in functional.negative_api.texture.

BUG=angle:571

Change-Id: I19d7250aab2c15d709fd591b8df90ce883f9ac60
Reviewed-on: https://chromium-review.googlesource.com/201250Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

6f38f823

Redesign HLSL scoped structures to a unique ID. · bfa91f47

authored Jun 05, 2014

A unique ID gives a more flexible renaming scheme than our
current method of using nested scope identifiers. The reduced
complexity allows for fewer points of breakage and fixes an
outstanding bug with scoped structures (with added test).

BUG=angle:618

Change-Id: I6551248bb9fa2d185ab67248721f898dd50151f0
Reviewed-on: https://chromium-review.googlesource.com/202183Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

bfa91f47

05 Jun, 2014 3 commits

Regenerate public projects. · 0abd523d

authored Jun 05, 2014

Change-Id: I57817666faca834e075216e63d7b0353d7a4d72c
Reviewed-on: https://chromium-review.googlesource.com/202596Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

0abd523d

Fix unused variable in validationES.cpp. · 04c68600

authored Jun 05, 2014

Broken in 560a8d88.

BUG=angle:571

Change-Id: Idb2414e11ad397cc45c2bdf5f357fd1b8e7fe6c2
Reviewed-on: https://chromium-review.googlesource.com/202602Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

04c68600

Refactor CopyTexImage validation. · 560a8d88

authored May 21, 2014

Move the common validation code to a shared base function.

BUG=angle:571

Change-Id: Id70b413b408a21f0a8933cfd4a09e261e637bae1
Reviewed-on: https://chromium-review.googlesource.com/200559Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

560a8d88

02 Jun, 2014 2 commits

Fix missing bracket in translator source. · 74be6a84

authored May 29, 2014

Broken in ed14b79d.

BUG=angle:628

Change-Id: Ia6b92dd13362bb51f0aa17a14278039d568f5434
Reviewed-on: https://chromium-review.googlesource.com/202182Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

74be6a84

Fix edge case scoped structures name conflict. · 96509e47

authored May 29, 2014

Structures with names ending in "_#" such as "_0" could conflict
with the internally rewritten scoped structures. Fix this by using
a prepending rule instead of appending.

Also includes a test, and fixes a WebGL test in Firefox. (Chrome is
not affected because of the variable hashing step.)

BUG=angle:618

Change-Id: I3d441f1de268b6d7e74a0834b43e889b7bfe578c
Reviewed-on: https://chromium-review.googlesource.com/201468Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>

96509e47

29 May, 2014 4 commits

Move the GL_APICALL and similar defines to the project level. · d7e7d735

authored May 27, 2014

BUG=angle:658

Change-Id: Ib4fa10bb89a6658efbc20d5a763d8ec9c3a5506f
Reviewed-on: https://chromium-review.googlesource.com/201465Tested-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

d7e7d735

Add include of common_defines.gypi for the angle_path variable. · 13b0d70a

authored May 28, 2014

BUG=skia:2272

Change-Id: I29540e009c41a1638e1a0399950621abbbbf2904
Reviewed-on: https://chromium-review.googlesource.com/201839Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

13b0d70a

Fix uniform bug from duplicate D3D9 shader pointers. · a5c9a144

authored May 26, 2014

In D3D9, sometimes the compiler would return the same shader
pointer on different calls to compile, if the programs were
similar. Work around this quirk by storing an applied serial
for the program, so we know when to refresh uniform data.

BUG=angle:661

Change-Id: I9750ee2298701dbe3f121dac6626198d24a2d1b2
Reviewed-on: https://chromium-review.googlesource.com/201562Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

a5c9a144

Don't allocate unused swap chain buffers · 673e680c

authored May 28, 2014

According to
msdn.microsoft.com/en-us/library/windows/desktop/bb173075(v=vs.85).aspx
the front buffer is the desktop and is not a dedicated part of the swap chain.
Thus we should only need 1 swap chain buffer instead of the currently allocated
2. Samples in the DirectX SDK also only allocate just one.

BUG=angle:662

Change-Id: If88eeab1954cbc92b39ac7cf42f924515c4b77f9
Reviewed-on: https://chromium-review.googlesource.com/201860Tested-by: Nicolas Capens <nicolascapens@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

673e680c

28 May, 2014 3 commits

Fix warnings about unreferenced local variables. · 9cd1915c

authored May 28, 2014

BUG=skia:2272

Change-Id: Ibf03efedc662fea2a389ad2dc5af5b7b014181a8
Reviewed-on: https://chromium-review.googlesource.com/201900Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

9cd1915c

Fix register packing for structs in d3d9. · 24d8d675

authored May 27, 2014

Fixes WebGL test misc/shader-with-array-of-structs-uniform

BUG=angle:656

Change-Id: If79a19db4d40bab9110422f7876de32e7f85e506
Reviewed-on: https://chromium-review.googlesource.com/200620Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>

24d8d675

Add a compiler query for the translator output type. · 68fe74aa

authored May 27, 2014

This is useful for determining if we are compiling to a D3D9 or D3D11
shader outside of the internal translator classes.

BUG=angle:656

Change-Id: Ib1c1d3de569edaa2b65c24c09d05aa4dd229d3e4
Reviewed-on: https://chromium-review.googlesource.com/201564Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

68fe74aa

27 May, 2014 1 commit

Use 'inout' qualifiers in generated HLSL for 'out' parameters. · 5f5320a9

authored May 26, 2014

With our added padding to some structures, using only 'out' in the
generated HLSL can cause warnings because not all of the structure members
are initialized.

BUG=angle:286

Change-Id: I866960883de487f19883fb2bc086c02703ea6271
Reviewed-on: https://chromium-review.googlesource.com/201364Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

5f5320a9

26 May, 2014 2 commits

Regenerate public projects. · afc10260

authored May 26, 2014

Change-Id: Ibaccaa46457e9ee4379c137182673153a17f207d
Reviewed-on: https://chromium-review.googlesource.com/201461Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

afc10260

Fix placement of windows.h include for Win 8.1 SDK. · d61e4726

authored May 26, 2014

The windows header must be included before versionhelpers.h.

BUG=angle:649

Change-Id: Ifee95b2918ea039f9a1c93ad0244bf2d35bd8570
Reviewed-on: https://chromium-review.googlesource.com/201462Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

d61e4726

23 May, 2014 4 commits

Make available EGL surface parameters queryable. · 3501c165

authored May 21, 2014

BUG=angle:331

Change-Id: Idb3c3ddc6ddd6bfc21b15b7e8942975ce4cd051f
Reviewed-on: https://chromium-review.googlesource.com/200811Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Nicolas Capens <nicolascapens@chromium.org>

3501c165

Move state query validation out of Context. · 893ab087

authored May 16, 2014

Generate all GL errors in the validation helper functions, instead of
within the state manipulation logic and internals of Context.

BUG=angle:571

Change-Id: I7a3f540e2ae0f5f8c7126e2593717cc3200dd7e5
Reviewed-on: https://chromium-review.googlesource.com/200551Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

893ab087

Automate the DEQP tests by wrapping them in the gtest suite. · 5df9f523

authored May 16, 2014

BUG=angle:497

Change-Id: If0a72c053bccccc4369ec78dd70173bbadb1be7b
Reviewed-on: https://chromium-review.googlesource.com/200044Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

5df9f523

Automate the es conformance tests by wrapping them in the gtest suite. · 6ba49d94

authored May 07, 2014

BUG=angle:497

Change-Id: If12b2dd79f9f666c5d686237d5663f316171b15c
Reviewed-on: https://chromium-review.googlesource.com/200043Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

6ba49d94

22 May, 2014 4 commits

Fix arrays-of-struct and nested struct uniforms. · c600c8c3

authored May 16, 2014

Our computation of register offsets had a few errors that
messed up nested structs and arrays of structs.

This fixes a WebGL test, as well as several dEQP tests.

BUG=angle:505
TEST=WebGL CTS 1.0.3, dEQP-GELS3.functional.uniform-api

Change-Id: Id5a0f39481e079fe5ef5ef856943dc1f91ee3639
Reviewed-on: https://chromium-review.googlesource.com/200045Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

c600c8c3

Move all uniform validation from ProgramBinary to API layer. · 3639892d

authored May 20, 2014

This will have the contract of having a successful API call once we
call into the ANGLE internals, having successfully passed valiation.

BUG=angle:571

Change-Id: Ia8d6eccc1e42bdf9e9cc5c10094a08729b42778a
Reviewed-on: https://chromium-review.googlesource.com/200075Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

3639892d

Refactor Uniform validation to validationES.cpp. · aa981bd7

authored May 20, 2014

This patch also incidentally fixes some bugged validation in the
new ES3 unsigned int uniform API entry points.

BUG=angle:571

Change-Id: I89bd909467bd9418fe5ce3f246561765cf27a5e9
Reviewed-on: https://chromium-review.googlesource.com/200074Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

aa981bd7

Refactor UniformMatrix validation. · d7c7bb27

authored May 20, 2014

Move the shared validation logic to validationES.cpp.

BUG=angle:571

Change-Id: I2fb06a012703c647de2176914238b48665145c6f
Reviewed-on: https://chromium-review.googlesource.com/200073Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

d7c7bb27

21 May, 2014 2 commits

Fix build warnings in x64. · 82b42271

authored May 20, 2014

A std::min/max with size_t and unsigned int.

Popped up in f98e63a5.

BUG=

Change-Id: I25f26ddbe7fb333f895be8a8dce3d3d73044c3f9
Reviewed-on: https://chromium-review.googlesource.com/200484Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

82b42271

Fix build on Windows 8.1 SDK. · fc2521e1

authored May 16, 2014

The Windows 8.1 SDK includes a new DLL compiler DLL, and is useful
to test against, so we should fix the few deprecated functions and
warnings.

BUG=angle:649

Change-Id: I28bcf49564f3f7d7403d939d61dd78d3972281ba
Reviewed-on: https://chromium-review.googlesource.com/199341Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shannon Woods <shannonwoods@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

fc2521e1

20 May, 2014 1 commit

Adds resource string generation to ANGLE · 2d76e5f6

authored May 16, 2014

This will avoid Chrome source having to be updated each time that ANGLE's
ShBuiltInResources changes.

BUG=374942

Change-Id: If54dba8351de9b261ff269e885f231547c08ff0a
Reviewed-on: https://chromium-review.googlesource.com/200171Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@chromium.org>
Tested-by: Shannon Woods <shannonwoods@chromium.org>

2d76e5f6