kc3-lang/angle/src/compiler
Branch :
Log
| Author | Commit | Date | CI | Message |
|---|---|---|---|---|
 |
bc709339 | 2015-04-14 15:32:19 | Revert "Support constant folding of trigonometry built-ins" Part of a chain causing compile errors on Mac. Example: ../../third_party/angle/src/compiler/translator/IntermNode.cpp:1216:89: error: no member named 'sin' in namespace 'std'; did you mean 'sinf'? http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33747 This reverts commit 1767e6b4f9148a6aa462d23e3838810c0080ce78. Change-Id: Ie39b1bf9a08da61aa0b16e219b34d1ba0b6f6c0f Reviewed-on: https://chromium-review.googlesource.com/265587 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
3c772f36 | 2015-04-14 15:30:45 | Revert "Support constant folding of exponential built-ins" Part of a chain causing compile errors on Mac. Example: ../../third_party/angle/src/compiler/translator/IntermNode.cpp:1216:89: error: no member named 'sin' in namespace 'std'; did you mean 'sinf'? http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33747 This reverts commit 62e2c8d2bc55c41db4e6095eac012b8861131585. Change-Id: Id20f7f4efbc3df7756161b192d4858caeeb0572f Reviewed-on: https://chromium-review.googlesource.com/265627 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
d1e9a0f4 | 2015-04-14 15:30:16 | Revert "Support constant folding of common built-ins" Part of a chain causing compile errors on Mac. Example: ../../third_party/angle/src/compiler/translator/IntermNode.cpp:1216:89: error: no member named 'sin' in namespace 'std'; did you mean 'sinf'? http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33747 This reverts commit 2b1da6ed1dbe5960bed8c1b2fa3d33cb00c6c116. Change-Id: I5d48932a6254d6d1e78966bb3891913d9450e08e Reviewed-on: https://chromium-review.googlesource.com/265612 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
b775778a | 2015-04-14 15:29:29 | Revert "Use nullptr consistently in TIntermConstantUnion::fold" Part of a chain causing compile errors on Mac. Example: ../../third_party/angle/src/compiler/translator/IntermNode.cpp:1216:89: error: no member named 'sin' in namespace 'std'; did you mean 'sinf'? http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33747 This reverts commit 72ed20d4bb5a5c55d9909d6cc9b7a0e89e4a81a2. Change-Id: Ia6119f6081b3184deef3deef25b9e3f43ff2e2a6 Reviewed-on: https://chromium-review.googlesource.com/265586 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
 |
72ed20d4 | 2015-04-08 14:26:25 | Use nullptr consistently in TIntermConstantUnion::fold Newly added code in TIntermConstantUnion::fold uses nullptr where as remaining part of the function uses NULL. This change removes NULL and consistently uses nullptr in complete function. Change-Id: I600dcf11b686d8d72b2bcdac88b7158288c68105 Reviewed-on: https://chromium-review.googlesource.com/265395 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
2b1da6ed | 2015-04-06 17:35:38 | Support constant folding of common built-ins This change adds constant folding support for unary common built-ins: abs, sign, floor, trunc, round, roundEven, ceil and fract. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common* (80 out of 210 tests started passing with this change) Change-Id: I06800ed0e03764c0f7aab6bcd45c4f122de5a3c1 Reviewed-on: https://chromium-review.googlesource.com/265394 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
62e2c8d2 | 2015-04-06 17:29:48 | Support constant folding of exponential built-ins This change adds constant folding support for unary exponential built-ins - exp, log, exp2, log2, sqrt and inversesqrt. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.exponential* (48 out of 56 tests started passing with this change) Change-Id: Ie7808cd57d6ed7598c642a9a0940b8f5bd293741 Reviewed-on: https://chromium-review.googlesource.com/265393 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
1767e6b4 | 2015-04-02 11:45:09 | Support constant folding of trigonometry built-ins This change adds constant folding support for trigonometry built-in functions. Constant folding for these functions also fixes constant expression issues where constant initializer is a built-in trignometry function whose arguments are all constant expressions. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.angle_and_trigonometry* (112 out of 120 tests pass with this change) Change-Id: I66275b2ae9faecef63d76763d21a9b67d9bb68fa Reviewed-on: https://chromium-review.googlesource.com/265392 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
 |
2935c581 | 2015-04-08 14:32:06 | Simplify code related to variable declaration Rename nonInitErrorCheck to declareVariable to clarify that it declares variables. Merge arrayErrorCheck with that so that logic that is common between array and non-array declarations is only in one place. This will simplify adding array initializer handling. This also enables redeclaring gl_LastFragData using ESSL3 array type syntax. Comments in executeInitializer claimed that the TVariable object was needed for error recovery, but that was not actually true, so it can also use the new declareVariable method. Make "variable" a local variable instead of a parameter to executeInitializer, since the parameter was never used by callers of the function. TEST=angle_unittests, WebGL conformance tests BUG=angleproject:941 Change-Id: Ie133be62afc3e1f997370803cf21cada4e738935 Reviewed-on: https://chromium-review.googlesource.com/264674 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
b4a058bb | 2015-04-10 14:50:51 | Revert "Support constant folding of trigonometry built-ins" Part of a chain breaking the clang build: http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33588 This reverts commit af930db15010cacfcdc74bee630372b8f6eb3eeb. Change-Id: Ic0bf09b4088a1ee285fed0fbd77dfc4c682fcd12 Reviewed-on: https://chromium-review.googlesource.com/265144 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
50b7178d | 2015-04-10 14:50:15 | Revert "Support constant folding of exponential built-ins" Part of a chain breaking the clang build: http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33588 This reverts commit 1c89caef47a27c63a7d0f988e74816692a9b9b39. Change-Id: I36e080d8f64d4d4ca5de04bebbd119db83956d0e Reviewed-on: https://chromium-review.googlesource.com/265185 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
3cf271fe | 2015-04-10 14:49:36 | Revert "Support constant folding of common built-ins" Part of a chain breaking the clang build: http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33588 This reverts commit 0273c68c507a10af3e45c6e4a52a44adcd0cf742. Change-Id: I00d7617a7c380f2a5ea3f38ec14948a598aa5c1a Reviewed-on: https://chromium-review.googlesource.com/265184 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
7c9e8c96 | 2015-04-10 14:48:08 | Revert "Use nullptr consistently in TIntermConstantUnion::fold" Part of a chain breaking the clang build: http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Linux%20Builder/builds/33588 This reverts commit 5f2feee42971fdbbd7cb389fffe2f632060ec6e3. Change-Id: I5ad54b05148ae2d4f19e50a67aded2bad4f55f91 Reviewed-on: https://chromium-review.googlesource.com/265143 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
3873cd0b | 2015-04-10 15:00:55 | Fix issues with computing discontinuous loops AST analysis to narrow down usage of [[loop]] and [[unroll]] failed to account for break statements inside switch statements. Add switch statement handling. This fixes asserts/crashes in dEQP tests. BUG=angleproject:937 TEST=dEQP-GLES3.functional.shaders.switch.* Change-Id: I04fdfe4733772a2a234934123bbfacf0376df562 Reviewed-on: https://chromium-review.googlesource.com/265191 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
5f2feee4 | 2015-04-08 14:26:25 | Use nullptr consistently in TIntermConstantUnion::fold Newly added code in TIntermConstantUnion::fold uses nullptr where as remaining part of the function uses NULL. This change removes NULL and consistently uses nullptr in complete function. Change-Id: Ieaada3cf46bf3e16d3ef91b300c86537f778c33f Reviewed-on: https://chromium-review.googlesource.com/264632 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
0273c68c | 2015-04-06 17:35:38 | Support constant folding of common built-ins This change adds constant folding support for unary common built-ins: abs, sign, floor, trunc, round, roundEven, ceil and fract. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common* (80 out of 210 tests started passing with this change) Change-Id: I09f2f99b3108e0d2fb1919a0631e4317b6a28be3 Reviewed-on: https://chromium-review.googlesource.com/263709 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
1c89caef | 2015-04-06 17:29:48 | Support constant folding of exponential built-ins This change adds constant folding support for unary exponential built-ins - exp, log, exp2, log2, sqrt and inversesqrt. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.exponential* (48 out of 56 tests started passing with this change) Change-Id: I22af56876d1b7ce305697bf9bf43ad9ec5d8a3a5 Reviewed-on: https://chromium-review.googlesource.com/263708 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
af930db1 | 2015-04-02 11:45:09 | Support constant folding of trigonometry built-ins This change adds constant folding support for trigonometry built-in functions. Constant folding for these functions also fixes constant expression issues where constant initializer is a built-in trignometry function whose arguments are all constant expressions. BUG=angleproject:913 TEST= dEQP tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.angle_and_trigonometry* (112 out of 120 tests pass with this change) Change-Id: I2ed4360532469fe0d70048d5f2300a8db6f9fcda Reviewed-on: https://chromium-review.googlesource.com/263679 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
6ed7bbe8 | 2015-04-07 18:08:46 | Clean up arrayErrorCheck Remove some unnecessary TPublicType/TType conversions from the code, and clean up code style. voidErrorCheck is changed to take TBasicType so that it can be used with both TType and TPublicType. TEST=angle_unittests BUG=angleproject:941 Change-Id: I6f6cbc0761a4fc971299bad48864309009c54e7f Reviewed-on: https://chromium-review.googlesource.com/264673 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
 |
4176cd64 | 2015-04-02 11:14:38 | Remove now unused DetectDiscontuinity.h/.cpp BUG=angleproject:937 BUG=395048 Change-Id: Ie8e520267d594a595201cb7c2438a2202d9adf25 Reviewed-on: https://chromium-review.googlesource.com/263745 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
1239ee94 | 2015-03-19 14:38:02 | Use the AST analyses to narrow the usage of [[loop]] and [[unroll]] These attributes are now used exactly in the loops and ifs that require them, limiting the number of failed compilations due to excessive unrolling and flattening. Also output Lod0 functions only when needed. Adds unit tests for LOOP, FLATTEN and Lod0 generation. The patch was tested against the WebGL CTS 1.0.4 for which all the failures existed prior to this patch and seem to be unrelated to this change. It also works correctly on the following sites that had trouble with [[loop]] and [[unroll]]: * dev.miaumiau.cat/rayTracer "Skull Demo" * The turbulenz engine particle demo * Lots of ShaderToy samples (including "Volcanic" and "Metropolis") * Google Maps Earth mode * Lots of Chrome Experiments * Lagoa * madebyevan.com/webgl-water * SketchFab * Unit Tests BUG=angleproject:937 BUG=395048 Change-Id: I856de9025f10b79781929ec212dbffc2064a940e Reviewed-on: https://chromium-review.googlesource.com/264791 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
 |
794e0009 | 2015-04-07 18:31:54 | Fix and enable warning C4244 (Conversion from 'type1' to 'type2', possible loss of data) Change-Id: Id0e06d7d6600344d858f00dabc219d79289bbc82 Reviewed-on: https://chromium-review.googlesource.com/265020 Tested-by: Minmin Gong <mgong@microsoft.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
b3584fb4 | 2015-04-09 17:34:21 | Revert "Fix and enable warning C4244 (Conversion from 'type1' to 'type2', possible loss of data)" Causing a build failure on Mac/Clang: ./Tokenizer.cpp:551:7: error: extra tokens at end of #else directive [-Werror,-Wextra-tokens] #else if defined(_MSC_VER) http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Mac%20Builder/builds/29136 This reverts commit 3b26e231d99154814eb428f75a67bbe7a21adadc. Change-Id: I2d11ddcc18130d908fd2ec3d6f5ab890cfccd5e7 Reviewed-on: https://chromium-review.googlesource.com/264983 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
3b26e231 | 2015-04-07 18:31:54 | Fix and enable warning C4244 (Conversion from 'type1' to 'type2', possible loss of data) Change-Id: I73d9a2b9ad16f032be974b9c819de0dc1247c2ea Reviewed-on: https://chromium-review.googlesource.com/264533 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
47fadcb9 | 2015-04-08 14:13:53 | Remove code related to redeclaration of arrays It should not be possible to redeclare arrays. This seems to be another thing left over from earlier erroneous code which allowed declaring unsized arrays. TEST=angle_unittests BUG=angleproject:941 Change-Id: I711565230b35df077f268cec6fdeac6c4c03b4cd Reviewed-on: https://chromium-review.googlesource.com/264672 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
3739d235 | 2015-04-08 12:23:44 | Make checking for array non-constness depend on input version First, remove duplicate check for const qualifier on arrays. Only keep the check inside arrayQualifierErrorCheck(). Second, ESSL3 will introduce array initializers and by extension constant arrays, so it should allow const qualifier on arrays. These checks are somewhat superfluous in ESSL1 as well, since the parser already checks for missing initializers, but it's useful to keep the informative error messages around. Add a few tests to make sure that when the ESSL3 implementation progresses, it still checks for missing initializers on constant arrays. TEST=angle_unittests BUG=angleproject:941 Change-Id: Id871c872c5b92e2a5bf81c00080ac23004916a75 Reviewed-on: https://chromium-review.googlesource.com/264671 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
938f0029 | 2015-04-08 19:35:40 | Revert "Use the AST analyses to narrow the usage of [[loop]] and [[unroll]]" Caused linking failures on mac because the unit test asks for TranslatorHLSL which is not compiled. This reverts commit 3342e01f2a29343ea95961f0194f9d4f422cb840. Change-Id: I02b2f54ca5b90611f11b7a549e75bf2e8310639d Reviewed-on: https://chromium-review.googlesource.com/264790 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
3342e01f | 2015-03-19 14:38:02 | Use the AST analyses to narrow the usage of [[loop]] and [[unroll]] These attributes are now used exactly in the loops and ifs that require them, limiting the number of failed compilations due to excessive unrolling and flattening. Also output Lod0 functions only when needed. Adds unit tests for LOOP, FLATTEN and Lod0 generation. The patch was tested against the WebGL CTS 1.0.4 for which all the failures existed prior to this patch and seem to be unrelated to this change. It also works correctly on the following sites that had trouble with [[loop]] and [[unroll]]: * dev.miaumiau.cat/rayTracer "Skull Demo" * The turbulenz engine particle demo * Lots of ShaderToy samples (including "Volcanic" and "Metropolis") * Google Maps Earth mode * Lots of Chrome Experiments * Lagoa * madebyevan.com/webgl-water * SketchFab * Unit Tests BUG=angleproject:937 BUG=395048 Change-Id: If7baddae1cdae0b3a414aa49e5a4c4babedbfe50 Reviewed-on: https://chromium-review.googlesource.com/261263 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
50931459 | 2015-03-19 10:39:13 | Implement discontinuous loops AST analysis This will allow narrowing down which usages of [[flatten]] and [[unroll]] are actually useful. BUG=angleproject:937 BUG=395048 Change-Id: I091e647e3053d22edadd0cabb7c50bd5efa690b2 Reviewed-on: https://chromium-review.googlesource.com/263776 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
f4eab3b9 | 2015-03-18 12:55:45 | Implement gradient operation AST analysis This will allow narrowing down which usages of [[flatten]] and [[unroll]] are actually useful. BUG=angleproject:937 BUG=395048 Change-Id: Ib8d7b98431b8cd3563e1eff8ecc9ed5df1a9b7d6 Reviewed-on: https://chromium-review.googlesource.com/263775 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
a094a8a9 | 2015-04-07 11:53:06 | Add a compiler option to prune unused function and prototypes Also adds a simple unit test checking the pruning BUG=angleproject:937 BUG=395048 Change-Id: I88440378f66178dcebebcd596f8f80235903f20e Reviewed-on: https://chromium-review.googlesource.com/264568 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
78b3a8b9 | 2015-04-08 16:04:24 | Revert "Add a compiler option to prune unused function and prototypes" Compile error on Mac: http://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Mac%20Builder/builds/29051 This reverts commit e423d9ca6e1b340ae06d543419b8a515de5ac3f2. Change-Id: Ie08d7f2bf86089a006f3177480aa7491a9405257 Reviewed-on: https://chromium-review.googlesource.com/264585 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
e423d9ca | 2015-04-07 11:53:06 | Add a compiler option to prune unused function and prototypes Also adds a simple unit test checking the pruning BUG=angleproject:937 BUG=395048 Change-Id: I49904c34d1a72949cdc579569967d99c736c7237 Reviewed-on: https://chromium-review.googlesource.com/264415 Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
71d147f6 | 2015-02-11 11:15:24 | Implemented a CallDAG to allow for more AST analysis The CallDAG preprocesses the AST to construct a DAG of functions that can be used for several analyses. Use it to implement check for recursion and max call depth. It will also be used to limit the usage of [[flatten]] and [[unroll]]. BUG=angleproject:937 BUG=395048 Change-Id: I8578703f2d49513f315aecccbcff34914562e4ff Reviewed-on: https://chromium-review.googlesource.com/263774 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
693c9aa0 | 2015-04-07 17:50:36 | Replace setArray function in TPublicType with functions from TType Having only one way of setting or clearing arrayness of types clarifies code dealing with arrays. TEST=angle_unittests BUG=angleproject:941 Change-Id: I98cb7c44fd66440c9de8b4c6c4a02827e9300db7 Reviewed-on: https://chromium-review.googlesource.com/264361 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
93a90fd0 | 2015-04-07 18:14:07 | Remove half-baked non-sized array support ANGLE used to accept shaders with non-sized arrays in initializer lists. Fix this. Proper support for implicitly sized arrays will be implemented later for ESSL3. TEST=angle_unittests BUG=angleproject:941 Change-Id: I53c5ccf0f7ef09ad30e142f8350812959fc4846c Reviewed-on: https://chromium-review.googlesource.com/264360 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
9638c352 | 2015-04-01 14:34:52 | Support assigning array constructors in HLSL output This adds support to the following type of shader code, where an array constructor is used as a source of assignment: float a[3]; a = float[3](0.0, 1.0, 2.0); The assignment gets replaced in the HLSL code by a function call where the array is an out parameter and the constructor parameters are regular parameters. Constructors cannot yet be used in initializers. With this change in place, some of the relevant dEQP tests start passing. TEST=dEQP-GLES3.functional.shaders.arrays.constructor.* BUG=angleproject:941 Change-Id: I13ed603c02a30d9f9950a61c4d562ae9fac058ec Reviewed-on: https://chromium-review.googlesource.com/263403 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
f40319e9 | 2015-03-10 14:33:00 | Add basic support for array constructors Add limited support for parsing array constructors and writing them out as GLSL code. Still missing from this version: HLSL output, array support in initializer lists, arrays with implicit size. BUG=angleproject:941 Change-Id: I7febf80923c4cd0b730399f1f49f9456cf3668e9 Reviewed-on: https://chromium-review.googlesource.com/260572 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
711d8663 | 2015-04-02 19:43:57 | Revert "Implement gradient operation AST analysis" This reverts commit 2fc57a2ace58b7ae3106344c56be94948f761b2a. Change-Id: I619297090c8441c1b90099437f8764263cdd68cc Reviewed-on: https://chromium-review.googlesource.com/263728 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
|
c457e7bb | 2015-04-02 19:43:52 | Revert "Implement discontinuous loops AST analysis" This reverts commit f30a9ba96bdfa014fc182060c714a2c75451f4cf. Change-Id: I577cfd1e7d5d6138413c29205cbf4d35b6861226 Reviewed-on: https://chromium-review.googlesource.com/263754 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> | |
 |
186160b2 | 2015-04-02 19:34:38 | Revert "Implemented a CallDAG to allow for more AST analysis" failures in angle_unittests and compilation failure on clang (see: https://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Mac%20Builder%20%28dbg%29/builds/30016/steps/compile/logs/stdio) . This reverts commit b34d1d12969496b1b0fb53934b8ce3200304f900. Change-Id: Ia995fb2db0e891294f3461de01617cb13e5ae381 Reviewed-on: https://chromium-review.googlesource.com/263727 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org> | |
|
f30a9ba9 | 2015-03-19 10:39:13 | Implement discontinuous loops AST analysis This will allow narrowing down which usages of [[flatten]] and [[unroll]] are actually useful. BUG=angleproject:937 BUG=395048 Change-Id: Icc37abf51a1407c9e7bc708dc08d3ba4050a496a Reviewed-on: https://chromium-review.googlesource.com/261189 Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> | |
|
2fc57a2a | 2015-03-18 12:55:45 | Implement gradient operation AST analysis This will allow narrowing down which usages of [[flatten]] and [[unroll]] are actually useful. BUG=angleproject:937 BUG=395048 Change-Id: Ic31e434c7b5bb6c32eff2336e5949391bfee25a5 Reviewed-on: https://chromium-review.googlesource.com/260930 Reviewed-by: Nicolas Capens <capn@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> | |
|
b34d1d12 | 2015-02-11 11:15:24 | Implemented a CallDAG to allow for more AST analysis The CallDAG preprocesses the AST to construct a DAG of functions that can be used for several analysis. Use it to implement check for recursion and max call depth. It will also be used to limit the usage of [[flatten]] and [[unroll]]. BUG=angleproject:937 BUG=395048 Change-Id: I643e0ed605ad095e4fd7da4977d842be5e6a12e0 Reviewed-on: https://chromium-review.googlesource.com/229352 Reviewed-by: Nicolas Capens <capn@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> | |
|
b2d6a9be | 2015-03-30 16:00:53 | Stub simplifying array assignment for HLSL output Add an AST traverser that rejects shaders with expressions that use the return value of an assignment where an array is assigned to another array. In the future this should be made into a tool that can simplify these expressions so that return values of array assignments are not used. In its current form this code prevents OutputHLSL from producing garbage code when the original code contains expressions like a = (b = c); where a, b, and c are all arrays. BUG=angleproject:960 Change-Id: I11353d7ed7160c853e58a0ef3471ca439cb314c8 Reviewed-on: https://chromium-review.googlesource.com/263070 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
1269076c | 2015-03-31 12:55:28 | Add basic support for assigning arrays in HLSL output Implement support for assignments where the return value of the assignment is not used in another part of the expression. TEST=WebGL conformance tests BUG=angleproject:960 Change-Id: Ibf9d71a75d27d139d2aabb5162ab04a0974321d3 Reviewed-on: https://chromium-review.googlesource.com/263222 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
f6c694bc | 2015-03-26 14:50:53 | Assign built-in function return type in promote() This finishes the refactoring of unary math operation handling so that IntermUnary::promote has the complete code for setting the return type of the node. BUG=angleproject:952 TEST=angle_unittests, WebGL conformance tests Change-Id: I19bd8d53029e24f734c9436eceb446b37e7fcf26 Reviewed-on: https://chromium-review.googlesource.com/262416 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
dca3e796 | 2015-03-26 13:24:04 | Refactor unary math operator handling to clarify responsibilities Shuffle the code around so that each part has a clear responsibility: IntermUnary::promote is responsible for setting the return type of the node, Intermediate::addUnaryMath is responsible for creating the node object, and ParseContext::createUnaryMath is responsible for validating the operand type. This removes duplicated bool type check for logical not. BUG=angleproject:952 TEST=angle_unittests, WebGL conformance tests Change-Id: I9f5a0abb6434ad2730441ea9199ec3f5382ebcda Reviewed-on: https://chromium-review.googlesource.com/262415 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
5321c808 | 2015-04-02 17:08:16 | Add unsigned literal support to GLSL output BUG=angleproject:964 Change-Id: I8eb3cfd8d49beded4ac55ead97b50c7691e4980a Reviewed-on: https://chromium-review.googlesource.com/263710 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
f0d10f89 | 2015-03-31 12:56:52 | Replace non-copyable macro with a helper class. This class provides a simpler scheme for blocking default copy and assignment operators. It also reduces the amount of code needed since it's inherited to child classes. This also fixes the conflict between our macro and the same-named macro in Chromium code. BUG=angleproject:956 Change-Id: If0dc72aa3f63fbc7b8fa34907418821c64c39e2f Reviewed-on: https://chromium-review.googlesource.com/263257 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Brandon Jones <bajones@chromium.org> | |
|
fc7cfd1c | 2015-03-31 14:46:18 | Fix array function parameter syntax in HLSL equality function The fact that passing shaders to the platform compiler is deferred until linking is performed allowed this bug to pass undetected in earlier manual testing. TEST=WebGL conformance tests BUG=angleproject:941 Change-Id: Ic12b5d0c5547057551df90086453e897a161d559 Reviewed-on: https://chromium-review.googlesource.com/263221 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Nicolas Capens <capn@chromium.org> | |
|
69c11b5d | 2015-03-26 12:59:00 | Move validation from Intermediate::addUnaryMath to ParseContext Intermediate should only have logic for creating node objects, validation of parameter types belongs in ParseContext. BUG=angleproject:952 TEST=angle_unittests, WebGL conformance tests Change-Id: Ie90697641fabb2a837ccc4571a93616d63ea64e6 Reviewed-on: https://chromium-review.googlesource.com/262414 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> | |
|
5e5c826c | 2015-03-26 14:04:54 | Fix float-int conversion return type precision tracking Float-int conversion functions are defined to always return highp values. BUG=angleproject:865 TEST=angle_unittests Change-Id: Idf243b483f7b5edfcb54de2755af11e17c3756d3 Reviewed-on: https://chromium-review.googlesource.com/262413 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> | |
|
ff699005 | 2015-03-23 14:38:42 | Disallow operations on structs containing samplers ESSL 1.00 spec section 5.9 says that equality operators don't operate on structs containing sampler types. Section 5.7 also suggests this. ESSL 3.00 doesn't have a similar restriction. ESSL 1.00 spec section 4.1.7 says that structs containing samplers can't be used as l-values. This is interpreted to apply also in the case of ESSL 3.00, which similarly disallows samplers as l-values, but doesn't explicitly mention structs. BUG=angleproject:954 TEST=angle_unittests, WebGL conformance tests Change-Id: I73f74962a192e8d9449990ffa5f3d8c851491601 Reviewed-on: https://chromium-review.googlesource.com/261822 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
ae37a5c8 | 2015-03-20 16:50:15 | Support equality between structs containing arrays in HLSL output This requires sorting all equality functions together so that struct equality functions can have dependencies on array equality functions, but support for array equality functions that have dependencies on struct equality functions is also maintained. There's no automated test coverage for this specifically. The change was tested by manually inspecting shader output and ensuring that there were no test regressions in tests listed below. TEST=dEQP-GLES3.functional.shaders.* BUG=angleproject:954 Change-Id: If7199ab2446804afae50f103bb625101172882b9 Reviewed-on: https://chromium-review.googlesource.com/261550 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
7fb4955d | 2015-03-18 17:27:44 | Support array equality operator in HLSL output This requires adding functions to the shader source that can do the comparison for a specific array size. There's no automated test coverage specifically for this functionality, since all deqp tests that cover this also require array constructors to be supported. The change has been tested by manually inspecting shader output. No regressions were seen in automated tests listed below. TEST=dEQP-GLES3.functional.shaders.*, angle_unittests BUG=angleproject:941 Change-Id: Ie2ca7c016a3f0bcb3392a96d6d20d6f803d28bf0 Reviewed-on: https://chromium-review.googlesource.com/261530 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
9dd217bc | 2015-03-20 14:24:31 | Disallow operations on structures containing arrays in ESSL1 Comparing structures that contain arrays to each other and assigning structures that contain arrays is "not defined" in ESSL 1.00 (section 5.7). Sections 5.8 and 5.9 further suggest that these operations are not allowed. Additionally some platform drivers on Linux seem to reject shaders produced by ANGLE which compare structures containing arrays. This might require changing the output GLSL version for ESSL 3.00. BUG=angleproject:954 TEST=angle_unittests Change-Id: I5f3a016f360f940f2fc1ec1ff8e60d13a977eb69 Reviewed-on: https://chromium-review.googlesource.com/261531 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
47fd36a7 | 2015-03-19 14:22:24 | Move some validation from IntermBinary::promote to ParseContext This makes the role of promote() in the system clearer and helps to make the code more understandable, since more of the checks are in the same logical place. BUG=angleproject:952 TEST=angle_unittests, WebGL conformance tests Change-Id: Idb2de927d872e46210d71cf6de06a6f8c1fc5da1 Reviewed-on: https://chromium-review.googlesource.com/260803 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
 |
c8ef69d2 | 2015-03-18 16:43:22 | Fix C4702 issues (unreachable code) in ANGLE, excluding <xtree> Change-Id: Ia7603139af266fd7b14efc8c3465225738456e67 Reviewed-on: https://chromium-review.googlesource.com/261038 Tested-by: Austin Kinross <aukinros@microsoft.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
e79904c3 | 2015-03-18 16:56:42 | Accept equality and assignment for arrays in parsing This is enough to support the operations in GLSL output. HLSL output will likely require additional work to support this. BUG=angleproject:941 Change-Id: I728d511ab07af94bc3382dc2796c1e9ac79d1442 Reviewed-on: https://chromium-review.googlesource.com/260801 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
1ade24db | 2015-03-19 16:29:32 | Add #version 300 es to ESSL output when needed BUG=angleproject:953 Change-Id: Ib9e45fab1d35f8c78abd0ac2209ac4f63afe4465 Reviewed-on: https://chromium-review.googlesource.com/261170 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> | |
|
d6b14287 | 2015-03-17 14:31:35 | Move binary op array check to ParseContext Also piece together an addAssign function in ParseContext that uses the binary op array check. This will make it easier to change the array-related checks in the future to use shaderVersion. Moving validation out from IntermBinary::promote also makes the architecture clearer, promote()'s role should be mainly to determine the type of the return value of the binary operation, not to do validation. BUG=angleproject:941 TEST=angle_unittests, WebGL conformance tests Change-Id: If1de33ea250893527be7f0d7930d4678a0864684 Reviewed-on: https://chromium-review.googlesource.com/260571 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
b2983c95 | 2015-03-18 14:02:46 | Correctly insert unmangled function names to symbol table This fixes detecting some cases of function parameter mismatch (previously a mangled function name corresponding to no function parameters was added to the symbol table for each user-defined function, and this was returned when doing function lookups with no parameters). Also fixes detection of reusing a function name as a variable/struct name. New unit tests are added to ensure that these fixes don't regress. BUG=angleproject:936 TEST=angle_unittests, WebGL conformance tests Change-Id: I2dadde9dcc01c7a4a653c1982c36377b89e6d437 Reviewed-on: https://chromium-review.googlesource.com/260800 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
fc1806e1 | 2015-03-17 13:03:11 | Move most of addBinaryMath from Intermediate to ParseContext Some type checks for binary math will be different based on the shading language version, which is easily accessible in ParseContext. Because of this and also for architectural simplicity it makes more sense to have the checks in ParseContext. BUG=angle:941 TEST=angle_unittests, WebGL conformance tests Change-Id: I92a499f47e1cbc6a7b6391ce0fa04284803e7140 Reviewed-on: https://chromium-review.googlesource.com/260570 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
6341939f | 2015-03-13 11:51:07 | Fix code duplication in OutputHLSL While debugging, I came across the piece of code in OutputHLSL::OutputHLSL which is unnecessarily duplicated, this change just fixes it. Change-Id: I0549d58696417ccf651bacfa19f489b290d4ba6e Reviewed-on: https://chromium-review.googlesource.com/259960 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
2728eda4 | 2015-03-16 10:23:50 | Revert "A special state to track field selection is not necessary." This reverts commit f8dc4fb66d2ee473b6ba37626f02756422863f6c. The special state to track field selection is actually necessary to avoid reduce/reduce conflicts when array constructors are added to the grammar. BUG=angleproject:941 TEST=WebGL conformance tests, angle_unittests Change-Id: I55476483c9e83241e8978cd58f05ef303c7c8680 Reviewed-on: https://chromium-review.googlesource.com/260260 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
4f52da4a | 2015-03-11 10:48:41 | Remove no-op functionality from IdentifyBuiltIns and refactor Adding array built-ins like gl_FragData and gl_LastFragData to the symbol table with a name that included their size in brackets was incorrect and unnecessary. Remove this. The array built-ins should only be added with their proper name and with an array type. Also refactor the code to reduce unnecessary repetition of conditions. The previous version attempted to split the built-ins that had dependencies on resources into their own part of the code, but this split was not clearly defined and only made the code harder to follow and prone to issues. TEST=WebGL conformance tests Change-Id: I7a77865a594864a22f3b566f9d3da1d0ead46466 Reviewed-on: https://chromium-review.googlesource.com/258751 Reviewed-by: Nicolas Capens <capn@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
61ffec88 | 2015-03-11 10:35:01 | Fix formatting of IdentifyBuiltIns Fix indentation of case labels, fix too long lines, put braces on their own lines and remove extraneous spacing and comments that didn't contribute to the readability of the code. TEST=WebGL conformance tests Change-Id: Ie33d0a7c4412c448081f5d879fa9f64baaf8d537 Reviewed-on: https://chromium-review.googlesource.com/258750 Reviewed-by: Nicolas Capens <capn@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
44efa0b8 | 2015-03-04 17:11:05 | Conditionally enable IEEE strictness for isnan() ANGLE's ESSL3 built-in function isnan() implementation just uses HLSL intrinsic function isnan(). For HLSL intrinsic function isnan() to work properly, IEEE strictness needs to be enabled for D3D compiler. This change detects use of isnan() in shaders and passes compiler flag D3DCOMPILE_IEEE_STRICTNESS whenever isnan is used in shaders. This change also moves existing workarounds in D3DWorkaroundType to D3DCompilerWorkarounds. BUG=angle:927 TEST= dEQP tests dEQP-GLES3.functional.shaders.builtin_functions.common.isnan.* Change-Id: I1ce5b1a7a825fdd720a37dc9aeb71320e55162d9 Reviewed-on: https://chromium-review.googlesource.com/255834 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Jamie Madill <jmadill@chromium.org> | |
 |
092e1a12 | 2015-03-09 07:40:11 | Initialize fragmentPrecisionHigh in constructor. This was the only field not initialized by the constructor, which looks like an oversight. Change-Id: If61424d574c663113a4c6d31e94b6ec13b412884 Reviewed-on: https://chromium-review.googlesource.com/257360 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org> | |
|
fc7fab72 | 2015-03-06 12:03:18 | Fix matrix compound multiplication Matrix compound multiplication was lacking transposes. Add them so that results match normal multiplication. BUG=angleproject:938 TEST=dEQP-GLES3.functional.shaders.matrix.mul_assign.* Change-Id: I4c5159de0727afb4ee2831586f04fa817619de7a Reviewed-on: https://chromium-review.googlesource.com/256366 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> | |
|
51585037 | 2015-03-05 19:40:42 | Remove unnecessary ifdef APPLE from built-in emulation The emulation is already toggled by a compilation flag, so having it behind an ifdef only adds unnecessary complexity, particularly when testing changes to the emulation code. BUG=angleproject:865 Change-Id: Idf7854c2b5323609880b527ed806440dcc8ac091 Reviewed-on: https://chromium-review.googlesource.com/256365 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> | |
|
8efc5ad5 | 2015-03-03 17:21:10 | Initialize BuiltInFunctionEmulator outside Compiler This moves GLSL output specific code from the Compiler class to the GLSL/ESSL translators. BUG=angleproject:865 Change-Id: I2d552e9cdb41f7d8ddfee7b0249a99d629a6d7d7 Reviewed-on: https://chromium-review.googlesource.com/255471 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
72ba85bf | 2015-03-04 14:23:26 | Fix regression in addFunctionCallOrMethod Fix a crash when parsing a shader with incorrect constructor arguments. BUG=angleproject:911 TEST=WebGL conformance tests Change-Id: I3459612f36185db861a361fa916cc6b671795037 Reviewed-on: https://chromium-review.googlesource.com/255870 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Zhenyao Mo <zmo@chromium.org> | |
|
95cd3c68 | 2015-03-03 16:45:32 | Move some output out of BuiltInFunctionEmulator subclasses This paves the way for getting rid of the BuiltInFunctionEmulator subclasses in favor of initializing the BuiltInFunctionEmulator dynamically in Compiler subclasses. The eventual goal is getting rid of GLSL output specific functionality in Compiler, which should be language-agnostic. BUG=angle:865 Change-Id: Ibb114b905785c7343b2726c97699268c982536a0 Reviewed-on: https://chromium-review.googlesource.com/255470 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
c4ba3bef | 2015-03-02 14:42:24 | Refactor addFunctionCallOrMethod out of glslang.y TEST=angle_unittests BUG=angle:911 Change-Id: I6d4c0f0bfbf9e6252e241ae5bb1b6acdb463fce6 Reviewed-on: https://chromium-review.googlesource.com/255451 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
 |
5dbece59 | 2015-02-27 20:35:38 | Implement trunc, round, roundEven (Emulated roundEven) BUG=angle:934 Change-Id: If33256bf3707f3f0202147276c71f08c65518205 Reviewed-on: https://chromium-review.googlesource.com/254842 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
2cd7a0ed | 2015-02-27 13:57:32 | Remove switch fall-through on HLSL Remove fall-through from non-empty labels in switch statements. Tested with dEQP-GLES3.functional.shaders.*switch*. All pass except for tests that require dFdx/dFdy, which fail for an unrelated reason. Additional test coverage could still be useful for cases where a label is terminated by a continue or return statement. BUG=angle:921 Change-Id: I4741867789a9308d66d0adeabdaf83907106e2d2 Reviewed-on: https://chromium-review.googlesource.com/254550 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> | |
|
687ef22c | 2015-03-03 10:58:05 | Fix derivatives being related to an extension on ESSL3 dFdx, dFdy and fwidth are parts of the core language in ESSL3. This fixes a regression from the patch "Set symbol extension at insertion". BUG=angle:926 Change-Id: I6c0bbce3afad1f064ff115a16e23ea79bd584e49 Reviewed-on: https://chromium-review.googlesource.com/255450 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
 |
05b6b7fc | 2015-03-02 17:08:09 | Add an SH_GLSL_CORE_OUTPUT profile. So we could generate shaders for Apple using core GL profile. By switching to core profile, we still pass most WebGL conformance tests 1.0.2 on Linux, but not all, so apparently more work is needed. However, I think it's OK to check this CL in because this output profile will be only used behind a chromium switch. BUG=angleproject:933 TEST=webgl conformance tests Change-Id: Iad70e1aebf82349d3fc5f4116c1d6bc4448193fd Reviewed-on: https://chromium-review.googlesource.com/255282 Tested-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
7749005a | 2015-03-03 18:36:05 | Fix min built-in function support One of the recent change that uses generic types to compact the symbol table initialization has introduced a bug in 'min' function implementation, just fixing it here. TEST= dEQP tests: dEQP-GLES3.functional.shaders.operator.common_functions.min.lowp_uvec* Change-Id: I0e3d386c89bc07c3b656ec5029b8ce57f663cf46 Reviewed-on: https://chromium-review.googlesource.com/255369 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
 |
c9d9b30a | 2015-02-20 23:02:15 | Set symbol extension at insertion. BUG=angle:926 Change-Id: I4c96ae89c79d720d99bde803541c3158ba4629ab Reviewed-on: https://chromium-review.googlesource.com/251662 Tested-by: Nicolas Capens <capn@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
05ae50dc | 2015-02-20 10:16:47 | Add basic support for HLSL output of switch and case This patch outputs switch and case statements in HLSL, but ignores fall-through. Switch statements that include fall-through cause the platform HLSL compiler to generate errors. BUG=angle:921 Change-Id: I10efbfb45076cf4fd79840698abafd8111e597d4 Reviewed-on: https://chromium-review.googlesource.com/251527 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
01cd8afa | 2015-02-20 10:39:20 | Add GLSL output of switch and case BUG=angle:921 Change-Id: I0d752440ce6ffdfcc005f1a6123694bdfeb1b067 Reviewed-on: https://chromium-review.googlesource.com/251526 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
c8716df9 | 2015-02-26 17:17:22 | Fix an issue with handling an error case inside ValidateSwitch In case a condition with a wrong type is given to a case statement, it generates an error but the compiler recovers. This caused ValidateSwitch to assert. Fix this. BUG=angle:921 Change-Id: I7949798cab923c2b168817471896470c6c611878 Reviewed-on: https://chromium-review.googlesource.com/254080 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
482907eb | 2015-02-23 16:56:33 | Set symbol operations at insertion. BUG=angle:926 Change-Id: Ifc8f62ebe1277214a4d3fbd5c56eb5faeff8e389 Reviewed-on: https://chromium-review.googlesource.com/251650 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Nicolas Capens <capn@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
e9cc469f | 2015-02-19 16:00:13 | Accept mismatching auxiliary interpolations. The ES3.1 spec, and discussion on Khronos.org, confirm that dEQP is correct in accepting mismatching centroid specifiers in shader linkage. Mismatching flat/smooth is still a link error. Fixes: shaders.linkage.varying.rules.differing_interpolation_2 Change-Id: I3016f4147e7c1b16b02371ee95866c8daf826212 Reviewed-on: https://chromium-review.googlesource.com/251205 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
 |
5cc9ac85 | 2015-02-20 11:27:56 | Fix shadow samplers with textureProj and vertex shaders. Fixes: dEQP-GLES3.functional.shaders.texture_functions.textureproj.sampler2dshadow_fragment dEQP-GLES3.functional.shaders.texture_functions.textureproj.sampler2dshadow_bias_fragment dEQP-GLES3.functional.shaders.texture_functions.textureprojoffset.sampler2dshadow_fragment dEQP-GLES3.functional.shaders.texture_functions.textureprojoffset.sampler2dshadow_bias_fragment and some of: dEQP-GLES3.functional.shaders.texture_functions.texture*.sampler2dshadow_vertex BUG=angle:925 Change-Id: I080e379ded95469f0486ae9c8bb1756842118a2d Reviewed-on: https://chromium-review.googlesource.com/251530 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Nicolas Capens <capn@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
ac5274df | 2015-02-20 10:19:08 | Add validation for the structure of switch statements Implement a traverser to check for the following errors: -More than one default or replicated constant expression -No statement between a label and the end of a switch statement -Statements in a switch statement before the first case statement -Mismatch between the type of init-expression and type of a case label -Case or default label nested inside other control flow nested within the corresponding switch Tested by manually disabling shading language version checks for switch and by manually enabling case statements in a Chromium build and checking that the expected errors are generated. BUG=angle:921 Change-Id: I99c49c17c8b520849adbe4d8521e46cb10e20e41 Reviewed-on: https://chromium-review.googlesource.com/251524 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
3c1dfb5a | 2015-02-20 11:34:03 | Add stubs for switch and case output BUG=angle:921 Change-Id: I58bd645a8d53ef5bad9b680e54c8948d50932fca Reviewed-on: https://chromium-review.googlesource.com/251525 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
 |
73941deb | 2015-02-25 14:34:49 | Disable the HLSL code in Angle if it is not being used. We're only using the HLSL code in Windows so it's not necessary to compile and distribute it on other platforms. This adds a defined ANGLE_ENABLE_HLSL that can be checked in files that are used by non-HLSL code as well. Mostly the HLSL code is just not include by the build system. Details of the space savings (heavily truncated) Total change: -165717 bytes =========================== -606 - Source: angle/src/common/utilities.cpp -627 - Source: angle/src/compiler/translator/FlagStd140Structs.cpp -695 - Source: /usr/include/c++/4.8/bits/stl_algo.h -710 - Source: angle/src/compiler/translator/TranslatorHLSL.cpp -713 - Source: angle/src/compiler/translator/IntermNode.h -863 - Source: /usr/include/c++/4.8/bits/stl_map.h -935 - Source: angle/src/compiler/translator/blocklayout.cpp -1515 - Source: angle/src/compiler/translator/BuiltInFunctionEmulator.cpp -1655 - Source: angle/src/compiler/translator/UnfoldShortCircuit.cpp -2375 - Source: /usr/include/c++/4.8/bits/vector.tcc -3135 - Source: angle/src/compiler/translator/RewriteElseBlocks.cpp -4656 - Source: angle/src/compiler/translator/UtilsHLSL.cpp -5265 - Source: angle/src/compiler/translator/BuiltInFunctionEmulatorHLSL.cpp -6505 - Source: /usr/include/c++/4.8/bits/stl_tree.h -11480 - Source: angle/src/compiler/translator/UniformHLSL.cpp -13580 - Source: angle/src/compiler/translator/StructureHLSL.cpp -18964 - Source: ?? (constant strings and a few vtbls) -89332 - Source: angle/src/compiler/translator/OutputHLSL.cpp Change-Id: I23ccc98abd0a21f847dd34f9482800b3ba679d56 Reviewed-on: https://chromium-review.googlesource.com/251528 Tested-by: bratell at Opera <bratell@opera.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> | |
|
53f076fa | 2015-02-20 10:55:14 | Add basic error checks for switch and case Check that the case labels have constant scalar integers, and that switch statements are initialized with scalar integers. Also check that case and default labels do not exist outside switch statements. Allow break statements inside switch statements. Tested by manually disabling shading language version checks for switch in a Chromium build and checking that the expected errors are generated. BUG=angle:921 Change-Id: Ibe83d0db52958c493ded5640d4babf670dc02d55 Reviewed-on: https://chromium-review.googlesource.com/251523 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
a3a36664 | 2015-02-17 13:46:51 | Implement parsing switch statements Put in some groundwork for parsing switch statements and case labels in the parser, including definitions for IntermNode classes. Intermediate functions for adding the statements are stubbed to only generate errors for now. Tested by manually disabling shading language version checks for switch in a Chromium build and checking that the expected errors are generated. BUG=angle:921 Change-Id: I064b3e0c4c1b724a083cf5bc78eebfdd3794eb1b Reviewed-on: https://chromium-review.googlesource.com/250380 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
49300865 | 2015-02-20 14:54:49 | Refactor validating jump statements out of glslang.y Move the logic to ParseContext. This will simplify changing the logic for the break statement that is needed to implement switch. BUG=angle:921, angle:911 TEST=WebGL conformance tests Change-Id: Ib8bbc7075c05d345e90377202b5446fb3307f94c Reviewed-on: https://chromium-review.googlesource.com/251522 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Zhenyao Mo <zmo@chromium.org> | |
|
f3cc4aef | 2015-02-23 13:51:25 | Use generic types to compact the symbol table initialization. Just like in the spec text, use symbolic types that represent multiple concrete types. The few signatures that get generated more than once are only added to the symbol table once because it ignores duplicates. BUG=angle:926 Change-Id: I216f03d22502b724dbefc87c5a021d6021c3a434 Reviewed-on: https://chromium-review.googlesource.com/251620 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Nicolas Capens <capn@chromium.org> | |
|
6b19d765 | 2015-02-19 09:40:39 | Implement missing variants of abs/sign This change adds ANGLE support for abs/sign variants introduced in ESSL3. BUG=angle:923 TEST=Unit tests, dEQP tests Unit Tests: TypeTrackingTest.BuiltInAbsSignFunctionFloatResultTypeAndPrecision TypeTrackingTest.BuiltInAbsSignFunctionIntResultTypeAndPrecision dEQP tests passing 100% because of this change: dEQP-GLES3.functional.shaders.builtin_functions.common.abs dEQP-GLES3.functional.shaders.builtin_functions.common.sign Change-Id: I2a3b028611f0eaaac3c031f8926d34a0e146663d Reviewed-on: https://chromium-review.googlesource.com/251491 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> | |
|
29190088 | 2015-02-20 16:42:54 | Make Angle code 40 KB smaller by using string literals directly. The implicit conversion of hundreds of string literals to TString generated a lot of machine code. By keeping them as string literals all the way the code will be smaller and faster. This is the change with clang for x64 (note VisitUnary in particular): Total change: -41392 bytes ========================== 2 added, totalling +469 bytes across 1 sources 2 removed, totalling -472 bytes across 1 sources 5 shrunk, for a net change of -41389 bytes (54126 bytes before, 12737 bytes after) across 1 sources 279692 unchanged, totalling 51433327 bytes ------------------------------------------------------------------------------------------------------------------------------------ -41392 - Source: /home/bratell/src/chromium/src/third_party/angle/src/compiler/translator/OutputHLSL.cpp - (gained 469, lost 41861) ------------------------------------------------------------------------------------------------------------------------------------ New symbols: +328: sh::OutputHLSL::outputConstructor(Visit, TType const&, char const*, TVector<TIntermNode*> const*) type=t, size=328 bytes +141: sh::OutputHLSL::outputTriplet(Visit, char const*, char const*, char const*) type=t, size=141 bytes Removed symbols: -133: sh::OutputHLSL::outputTriplet(Visit, std::basic_string<char, std::char_traits<char>, pool_allocator<char> > const&, std::basic_string<char, std::char_traits<char>, pool_allocator<char> > const&, std::basic_string<char, std::char_traits<char>, pool_allocator<char> > const&) type=t, size=133 bytes -339: sh::OutputHLSL::outputConstructor(Visit, TType const&, std::basic_string<char, std::char_traits<char>, pool_allocator<char> > const&, TVector<TIntermNode*> const*) type=t, size=339 bytes Shrunk symbols: -388: sh::OutputHLSL::writeEmulatedFunctionTriplet(Visit, char const*) type=t, (was 628 bytes, now 240 bytes) -714: sh::OutputHLSL::visitBranch(Visit, TIntermBranch*) type=t, (was 1017 bytes, now 303 bytes) -9738: sh::OutputHLSL::visitAggregate(Visit, TIntermAggregate*) type=t, (was 17609 bytes, now 7871 bytes) -14132: sh::OutputHLSL::visitBinary(Visit, TIntermBinary*) type=t, (was 17627 bytes, now 3495 bytes) -16417: sh::OutputHLSL::visitUnary(Visit, TIntermUnary*) type=t, (was 17245 bytes, now 828 bytes) Change-Id: Id0f87d72f6d7f1ab7b543f0d28d5a8b7c7db9ec7 Reviewed-on: https://chromium-review.googlesource.com/251090 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: bratell at Opera <bratell@opera.com> | |
|
80d934bb | 2015-02-19 10:16:12 | Enable MSVS warning 4512. Fix up the missing DISALLOW_COPY_AND_ASSIGN macros and various small problems preventing us from turning on this warning. This should ensure we more often use the DISALLOW macro going forward. Change-Id: I2e1a9d23a31a51279a577fad8dffb8c1530e2ee8 Reviewed-on: https://chromium-review.googlesource.com/251100 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Cooper Partin <coopp@microsoft.com> Tested-by: Cooper Partin <coopp@microsoft.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> | |
|
7c1f6f19 | 2015-02-19 18:56:09 | Revert "Implement missing variants of abs/sign" This for some reason is failing angle_unittests: http://build.chromium.org/p/chromium.gpu.fyi/builders/Win7%20Release%20%28NVIDIA%29/builds/13296 Happens across platforms. This reverts commit 3e142e47cc5b40564bcdba4252fd326bc28ca343. Change-Id: Ideb20457546c0dec10c9584182cc8d5ead42309d Reviewed-on: https://chromium-review.googlesource.com/251160 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org> | |
|
b6e07a6a | 2015-02-16 12:22:10 | Implement ESSL3 modf This is the first built-in function that has an out parameter, so l-value checks are added for built-ins. BUG=angle:918 Change-Id: Ifd5befe955224f706f864e25107879c9cdce9e9f Reviewed-on: https://chromium-review.googlesource.com/250780 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> |