1134 Commits

Author SHA1 Message Date
John Kessenich
5e56423046 Front-ends: Remove now defunct afterEOF and related, use scanner's instead.
Code using atEndOfFile was dead, instead do something useful with
the scanners atEndOfInput().  This allows a better error message
for early termination of cascading errors.
2016-08-31 13:46:50 -06:00
John Kessenich
81cd764b5f Non-functional: Add some missing const, related to signature selection. 2016-08-26 14:01:43 -06:00
John Kessenich
e3f2c8f98a HLSL: Include shape-changing conversions in overloaded signature selection.
This also enables vecN -> vec1 shape conversions for all places doing shape
conversions.

For signature selection, makes shape changes worse than any other comparison
when deciding what conversions are better than others.
2016-08-25 23:57:39 -06:00
John Kessenich
90dd70f752 HLSL: Allow arbitrary baseType -> baseType conversion of calling arguments.
This also puts a stake in the ground as to which is better when selection
from multiple signatures.
2016-08-25 10:51:29 -06:00
John Kessenich
ab89bbe702 Merge branch 'overloaded-400' of github.com:KhronosGroup/glslang 2016-08-23 18:30:20 -06:00
John Kessenich
32c169dbdf Front-end: Warn for likely missed change in default precisions.
This is part of the change to have desktop shaders respect precision
qualifiers on Vulkan, but since the defaults are all highp, and that's
different from ES fragment shaders, detect likely cases and warn about
them (but being careful to not be too noisy if it's unlikely to be a
problem).
2016-08-23 18:13:08 -06:00
John Kessenich
54571c2519 Vulkan: Have desktop shaders respect precision qualifiers.
Sets highp defaults for the appropriate types, for all stages,
and turns on precision qualifiers for non-ES shaders.  Required
fixing some qualifier orders for desktop built-in declarations
for pre-420 shaders.
2016-08-23 18:12:25 -06:00
John Kessenich
4d535640e4 Front-End: Non-ES trigger for processing precision qualifiers.
Initially non-functional: Have an independent test for whether to obey
precision qualifiers other than whether the profile is es.
2016-08-23 18:12:25 -06:00
steve-lunarg
efe9724795 HLSL: Add EHTokStringConstant, so that string attributes may be parsed 2016-08-22 17:13:17 -06:00
John Kessenich
370234888d Front-end: Add a function selector under implicit type conversion.
Use the new function selector for #version 400 and above,
parameterized for the GLSL #version 400 selection rules.

This can be used for both GLSL and HLSL, and other languages
as well.
2016-08-21 14:26:27 -06:00
John Kessenich
0a04b4df02 Front-end/Non-functional: Add some const/auto, useful for upcoming changes. 2016-08-19 07:27:28 -06:00
John Kessenich
2c6038ecf1 Merge branch 'cpp-headers' 2016-08-11 10:01:13 -06:00
John Kessenich
0a6fb85494 Infrastructure: Make shared symbol-table cache complete, delete work around.
Fixes issue #370 and replaces PR #371.
2016-08-09 19:15:17 -06:00
steve-lunarg
c4a1307403 HLSL: add implicit promotions for assignments and function returns. 2016-08-09 13:48:47 -06:00
John Kessenich
426542ba57 Front-end: Fix issue #443: dvec3 uses only 2 components of second location. 2016-08-08 15:34:00 -06:00
John Kessenich
4583b61e20 HLSL: Smear scalars to match vectors for relational operations.
Yield a vector relational compare and a vector result.
2016-08-07 19:14:22 -06:00
John Kessenich
267590d452 Whitespace: Nonfunctional: fix inconsistent white space, esp. no tabs. 2016-08-05 17:34:34 -06:00
John Kessenich
66ec80e01b Build: C++ headers: Replace PR #366 with a more directed version. 2016-08-05 14:04:23 -06:00
John Kessenich
9f9bad09b4 PP: Add missing atoms that effect -E mode.
When preprocessing only, some tokens were emitted as <bad token>.
This fixes them to preserve their original content.

This supplants PR #182, with a correction and test results.
2016-08-04 22:46:19 -06:00
John Kessenich
f6640761c4 Front-end: Implement 2nd task of issue #400; precision of result and operation.
From the ES spec + Bugzilla 15931 and GL_KHR_vulkan_glsl:
- Update precision qualifiers for all built-in function prototypes.
- Implement the new algorithm used to distinguish built-in function
  operation precisions from result precisions.
Also add tracking of separate result and operation precisions, and
use that in generating SPIR-V.
(SPIR-V cares about precision of operation, while the front-end
cares about precision of result, for propagation.)
2016-08-02 21:48:02 -06:00
John Kessenich
1176530bf5 SPV: Prevent issue #415 with better semantic checking. 2016-07-31 12:39:46 -06:00
John Kessenich
11e1a073f3 PP: Fix issue #407; handle empty identifier.
The sequence

  #define m()
  int m"

creates a token of no length (a string of 0 size).  Protect
against a string of 0 size as well as the existing protect
against a null string.
2016-07-30 13:39:52 -06:00
John Kessenich
7208473c69 PP: Non-functional; fix typos and blank spaces. 2016-07-30 13:25:33 -06:00
John Kessenich
73d4fb5bc5 PP: Fix issue #408: # as last character in macro.
This would look ahead for a second #, for token pasting, and if not
found, backup one token.  This is fine, unless at the end of line,
which would backup the #, rather than the look ahead.
2016-07-30 12:41:35 -06:00
John Kessenich
ff13213547 Front-ends GLSL/HLSL: Fix initializer lists for structs of one member.
Single member structs initialized with an initializer list had
an incorrect argument for constructor emulation.
2016-07-29 18:29:06 -06:00
John Kessenich
fea226ba43 HLSL: Add shape conversions for scalar -> vector assigments.
Also, this allows turning on the error check for a failed assigment
when parsing.

This makes 39 HLSL tests have a working assignment that was previously
silently dropped, due to lack of this functionality.
2016-07-28 18:41:20 -06:00
John Kessenich
a26a5170a3 Non-functional: Rationalize location and use of mapTypeToConstructor(). 2016-07-28 16:56:52 -06:00
John Kessenich
414f735443 Front-end: Fix issue #409, check for implicitly-sized binding arrays. 2016-07-27 14:55:06 -06:00
LoopDawg
6256146ef3 HLSL: enable #line extension by default for HLSL source. 2016-07-23 10:45:00 -06:00
John Kessenich
c8f6903b1d Front-end: Fix default layout(component) widths and correct for doubles.
Replaces PR #372.
2016-07-21 15:43:00 -06:00
John Kessenich
bc9e1d118a Infrastructure: Remove bad characters from file. 2016-07-21 15:41:10 -06:00
John Kessenich
75e547c0f2 Merge pull request #401 from steve-lunarg/samplers
HLSL: Add texture Load method & decomposition
2016-07-21 14:41:47 -06:00
LoopDawg
f245101954 HLSL: Add texture Load method & decomposition 2016-07-21 09:42:35 -06:00
Rex Xu
9d93a2370d Implement 4 AMD-specific extensions.
- Support GL_AMD_shader_ballot (SPV_AMD_shader_ballot).
- Support GL_AMD_shader_trinary_minmax (SPV_AMD_shader_trinary_minmax).
- Support GL_AMD_shader_explicit_vertex_parameter
  (SPV_AMD_shader_explicit_vertex_parameter).
- Support GL_AMD_gcn_shader (SPV_AMD_gcn_shader).
2016-07-20 16:18:05 +08:00
John Kessenich
934855a642 Fix issue #382: Detect implicitly-sized atomic_uint arrays. 2016-07-19 15:50:58 -06:00
John Kessenich
f2cfe27021 Fix issue #388.
Protect more against error recovery of bad built-in variable redeclarations.
2016-07-19 15:13:47 -06:00
John Kessenich
31a51becd2 Fix issue #391: input stack underflow on too few macro arguments. 2016-07-19 14:32:52 -06:00
John Kessenich
d8509b3367 Front-end: Implement compile-time constant folding for any() and all(). 2016-07-12 01:27:48 -06:00
John Kessenich
a86836ede2 Front-end: Fix known crashes by early exit on error (issue #29, issue #34, issue #35).
Added -C option to request cascading errors.  By default, will exit early,
to avoid all error-recovery-based crashes.

This works by simulating end-of-file in input on first error, so no
need for exception handling, or stack unwinding, or any complex error
checking/handling to get out of the stack.
2016-07-09 14:53:11 -06:00
John Kessenich
7f349c73db Build: Remove causes of pedantic warnings. Addresses issue #352 and PR #242. 2016-07-08 22:09:10 -06:00
John Kessenich
91e4aa5900 SPV: Further refine OpenGL vs. Vulkan differences for SPIR-V.
Includes adding test cases to verify the differences.
2016-07-07 19:27:15 -06:00
John Kessenich
2d0cc786f3 SPV: Turn on atomic-storage functionality for SPIR-V.
This is used by OpenGL, but not Vulkan.
Includes:
 - atomicCounter, atomicIncrement, atomicCounterDecrement
 - atomic_uint layout-offset checking
 - AtomicStorage capability
2016-07-07 17:40:35 -06:00
John Kessenich
266b1d3e84 Merge pull request #363 from cdwfs/vertex_instance_id
Suggest gl_[Vertex,Instance]Index when gl_[Vertex,Instance]ID is encountered in Vulkan GLSL
2016-06-30 18:15:24 -06:00
John Kessenich
f5dcdf0123 Front-end: Don't lower precision of bit operations taking highp inputs.
See issue #350 for detail.  This may be an ESSL specification issue.
2016-06-30 16:46:24 -06:00
Cort Stratton
85028e1469 When parsing Vulkan GLSL, suggest gl_VertexIndex/gl_InstanceIndex instead of gl_VertexID/gl_InstanceID. 2016-06-30 12:51:10 -07:00
John Kessenich
187254b92c Merge pull request #357 from steve-lunarg/intrinsics
HLSL: Add asdouble, fma, & mad intrinsics
2016-06-29 10:50:33 -06:00
John Kessenich
bc17b0a96a Front-end: Add missing rule for function-return implicit type conversion check.
Basic (component) type conversion was done, but checking that the composite
was correct was not being done.

Fixes issue #353.
2016-06-28 23:56:31 -06:00
John Kessenich
b10feabc95 Front-end: Non-functional: Move return-value handling from .y to .cpp.
This is to make the real change needed in the next commit easier.
2016-06-28 23:54:36 -06:00
John Kessenich
1011971981 Front-end: Handle simultaneous use of SPIR-V or non-SPIR-V.
Symbol table sharing has different sets of symbols for different
versions. This needs to take into account that SPIR-V has different
symbols than non-SPIR-V.
2016-06-28 16:16:43 -06:00
LoopDawg
1b7fd0f7b7 Add asdouble, fma, & mad intrinsics and change profile to allow doubles when parsing prototypes 2016-06-28 15:38:38 -06:00