Check if the repository contains the known-good commit.

Some texture methods use an addition coordinate dimension for the mip

Initialize the full set of shareable symbol tables; The common (cross-stage) and those shareable per-stage.

Reject certain combinations that are illegal sample methods. For example, 3D arrays.

arg order queries

Determine whether this fnCandidate overload is supported for these arguments. Some of this is based on the BaseFunctions table, but for functions not

This rejects prototypes not normally valid for GLSL and it's way of finding overloaded built-ins under implicit type conversion. It is possible that

LOD calculations don't pass the array level in the coordinate.

Print infinities and NaNs, and numbers in a portable way. Goals: - portable (across IEEE 754 platforms) - shows all possible IEEE values - shows simpl

Write SPIR-V out to a binary file

Write SPIR-V out to a text file with 32-bit hexadecimal words

Set up all the parameterizing descriptions of the opcodes, operands, etc.

Take a single compilation unit, and run the preprocessor on it. Return: True if there were no issues found in preprocessing, False if during preproces

Do all command-line argument parsing. This includes building up the work-items to be processed later, and saving all the command-line options. Does

Process an optional binding base of one the forms: --argname name {uniform|buffer|push_constant}

Parse either a .conf file provided by the user or the default from glslang::DefaultTBuiltInResource

Most processes are recorded when set in the intermediate representation, These are the few that are not.

Entry point.

KHR_vulkan_glsl says "Two arrays sized with specialization constants are the same type only if sized with the same symbol, involving no operations"

.conf => this is a config file that can set limits/resources

To do this on the fly, we want to leave the current state of our thread's pool allocator intact, so: - Switch to a new pool for parsing the built-ins

Do a full compile on the given strings for a single compilation unit forming a complete stage. The result of the machine dependent compilation is lef

Use the SPIRV-Tools disassembler to print SPIR-V using a SPV_ENV_UNIVERSAL_1_3 environment.

Apply the SPIRV-Tools optimizer to strip debug info from SPIR-V. This is implicitly done by SpirvToolsTransform if spvOptions->stripDebugInfo is set,

Apply the SPIRV-Tools optimizer to generated SPIR-V. HLSL SPIR-V is legalized in the process.

Apply the SPIRV-Tools validator to generated SPIR-V.

Translate glslang sampler type to SPIR-V dimensionality.

There are multiple paths in for setting environment stuff. TEnvironment takes precedence, for what it sets, so sort all this out. Ideally, the interna

Translate glslang language (stage) to SPIR-V execution model.

Translate glslang profile to SPIR-V source language.

See if the tabled versioning information allows the current version.

push a dynamic component selection onto the access chain, only applicable with a non-trivial swizzle or no swizzle

Create a for-loop sequence.

Log initial settings and transforms. See comment for class TProcesses.

For an identifier that is already declared, add more qualification to it.

Helper function to add the namespace scope mangling syntax to a string.

Convert the node's shape of type for the given type, as allowed by the operation involved: 'op'. Generally, the AST represents allowed GLSL shapes, s

Turn the top-level node sequence built up of wrapupSwitchSubsequence9) into a switch node.

Turn the top-level node sequence built up of wrapupSwitchSubsequence into a switch node.

Install 'this' as the first parameter. 'this' is reflected in the list of parameters, but not the mangled name.

Accumulate locations used for inputs, outputs, and uniforms, payload, callable data, and tileImageEXT and check for collisions as the accumulation is

Accumulate bindings and offsets, and check for collisions as the accumulation is done. Returns < 0 if no collision, >= 0 if collision and the value r

Accumulate xfb buffer ranges and check for collisions as the accumulation is done. Returns < 0 if no collision, >= 0 if collision and the value retur

Recursively make the implicit array size the explicit array size. Expicit arrays are compile-time or link-time sized, never run-time sized. Sometimes,

Update the level info in symbol's unique ID to current level

Both test, and if necessary spit out an error, to see if the node is really supported as an array index.

See if this qualifier can be an array. Returns true if there is an error.

Do size checking for an array type's size.

Same error message for all places assignments don't work.

Map the given string to an attribute enum from TAttributeType, or EatNone if invalid.

Same error message for all binary operations don't work.