[spirv-reader] Add ConvertFToU, ConvertFToS

Bug: tint:3
Change-Id: I9f3188e0aac64e98da785c4df2e8b2aa42b71cf8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/23402
Reviewed-by: dan sinclair <dsinclair@google.com>
This commit is contained in:
David Neto 2020-06-18 17:57:23 +00:00
parent 206d7401ae
commit d7868e34c2
4 changed files with 556 additions and 26 deletions

View File

@ -50,6 +50,7 @@
#include "src/ast/storage_class.h"
#include "src/ast/switch_statement.h"
#include "src/ast/type/bool_type.h"
#include "src/ast/type/type.h"
#include "src/ast/type/u32_type.h"
#include "src/ast/type/vector_type.h"
#include "src/ast/type_constructor_expression.h"
@ -2473,12 +2474,9 @@ TypedExpression FunctionEmitter::MaybeEmitCombinatorialValue(
auto arg1 = MakeOperand(inst, 1);
auto binary_expr = std::make_unique<ast::BinaryExpression>(
binary_op, std::move(arg0.expr), std::move(arg1.expr));
auto* forced_result_ty = parser_impl_.ForcedResultType(opcode, arg0.type);
if (forced_result_ty && forced_result_ty != ast_type) {
return {ast_type, std::make_unique<ast::AsExpression>(
ast_type, std::move(binary_expr))};
}
return {ast_type, std::move(binary_expr)};
TypedExpression result(ast_type, std::move(binary_expr));
return parser_impl_.RectifyForcedResultType(std::move(result), opcode,
arg0.type);
}
auto unary_op = ast::UnaryOp::kNegation;
@ -2486,12 +2484,9 @@ TypedExpression FunctionEmitter::MaybeEmitCombinatorialValue(
auto arg0 = MakeOperand(inst, 0);
auto unary_expr = std::make_unique<ast::UnaryOpExpression>(
unary_op, std::move(arg0.expr));
auto* forced_result_ty = parser_impl_.ForcedResultType(opcode, arg0.type);
if (forced_result_ty && forced_result_ty != ast_type) {
return {ast_type, std::make_unique<ast::AsExpression>(
ast_type, std::move(unary_expr))};
}
return {ast_type, std::move(unary_expr)};
TypedExpression result(ast_type, std::move(unary_expr));
return parser_impl_.RectifyForcedResultType(std::move(result), opcode,
arg0.type);
}
if (opcode == SpvOpAccessChain || opcode == SpvOpInBoundsAccessChain) {
@ -2539,7 +2534,8 @@ TypedExpression FunctionEmitter::MaybeEmitCombinatorialValue(
if (opcode == SpvOpVectorShuffle) {
return MakeVectorShuffle(inst);
}
if (opcode == SpvOpConvertSToF || opcode == SpvOpConvertUToF) {
if (opcode == SpvOpConvertSToF || opcode == SpvOpConvertUToF ||
opcode == SpvOpConvertFToS || opcode == SpvOpConvertFToU) {
return MakeNumericConversion(inst);
}
@ -2547,13 +2543,9 @@ TypedExpression FunctionEmitter::MaybeEmitCombinatorialValue(
// glsl.std.450 readonly function
// Instructions:
// OpCopyObject
// OpUndef
// OpBitcast
// OpSatConvertSToU
// OpSatConvertUToS
// OpConvertFToS
// OpConvertFToU
// OpSatConvertSToU // Only in Kernel (OpenCL), not in WebGPU
// OpSatConvertUToS // Only in Kernel (OpenCL), not in WebGPU
// OpUConvert // Only needed when multiple widths supported
// OpSConvert // Only needed when multiple widths supported
// OpFConvert // Only needed when multiple widths supported
@ -2902,11 +2894,52 @@ void FunctionEmitter::RegisterValuesNeedingNamedDefinition() {
TypedExpression FunctionEmitter::MakeNumericConversion(
const spvtools::opt::Instruction& inst) {
auto* result_type = parser_impl_.ConvertType(inst.type_id());
const auto opcode = inst.opcode();
auto* requested_type = parser_impl_.ConvertType(inst.type_id());
auto arg_expr = MakeOperand(inst, 0);
if (!arg_expr.expr || !arg_expr.type) {
return {};
}
return {result_type, std::make_unique<ast::CastExpression>(
result_type, std::move(arg_expr.expr))};
ast::type::Type* expr_type = nullptr;
if ((opcode == SpvOpConvertSToF) || (opcode == SpvOpConvertUToF)) {
if (arg_expr.type->is_integer_scalar_or_vector()) {
expr_type = requested_type;
} else {
Fail() << "operand for conversion to floating point must be integral "
"scalar or vector, but got: "
<< arg_expr.type->type_name();
}
} else if (inst.opcode() == SpvOpConvertFToU) {
if (arg_expr.type->is_float_scalar_or_vector()) {
expr_type = parser_impl_.GetUnsignedIntMatchingShape(arg_expr.type);
} else {
Fail() << "operand for conversion to unsigned integer must be floating "
"point scalar or vector, but got: "
<< arg_expr.type->type_name();
}
} else if (inst.opcode() == SpvOpConvertFToS) {
if (arg_expr.type->is_float_scalar_or_vector()) {
expr_type = parser_impl_.GetSignedIntMatchingShape(arg_expr.type);
} else {
Fail() << "operand for conversion to signed integer must be floating "
"point scalar or vector, but got: "
<< arg_expr.type->type_name();
}
}
if (expr_type == nullptr) {
// The diagnostic has already been emitted.
return {};
}
TypedExpression result(expr_type, std::make_unique<ast::CastExpression>(
expr_type, std::move(arg_expr.expr)));
if (requested_type == expr_type) {
return result;
}
return {requested_type, std::make_unique<ast::AsExpression>(
requested_type, std::move(result.expr))};
}
} // namespace spirv

View File

@ -26,6 +26,7 @@ namespace reader {
namespace spirv {
namespace {
using ::testing::Eq;
using ::testing::HasSubstr;
std::string CommonTypes() {
@ -33,10 +34,16 @@ std::string CommonTypes() {
%void = OpTypeVoid
%voidfn = OpTypeFunction %void
%bool = OpTypeBool
%uint = OpTypeInt 32 0
%int = OpTypeInt 32 1
%float = OpTypeFloat 32
%true = OpConstantTrue %bool
%false = OpConstantFalse %bool
%v2bool = OpTypeVector %bool 2
%v2bool_t_f = OpConstantComposite %v2bool %true %false
%uint_10 = OpConstant %uint 10
%uint_20 = OpConstant %uint 20
%int_30 = OpConstant %int 30
@ -119,6 +126,105 @@ TEST_F(SpvUnaryConversionTest, Bitcast_Vector) {
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertSToF_BadArg) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertSToF %float %void
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
HasSubstr("unhandled expression for ID 2\n%2 = OpTypeVoid"));
}
TEST_F(SpvUnaryConversionTest, ConvertUToF_BadArg) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertUToF %float %void
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
HasSubstr("unhandled expression for ID 2\n%2 = OpTypeVoid"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_BadArg) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToS %float %void
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
HasSubstr("unhandled expression for ID 2\n%2 = OpTypeVoid"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_BadArg) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToU %float %void
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
HasSubstr("unhandled expression for ID 2\n%2 = OpTypeVoid"));
}
TEST_F(SpvUnaryConversionTest, ConvertSToF_Scalar_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertSToF %float %false
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
HasSubstr("operand for conversion to floating point must be "
"integral scalar or vector, but got: __bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertSToF_Vector_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertSToF %v2float %v2bool_t_f
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(
p->error(),
HasSubstr("operand for conversion to floating point must be integral "
"scalar or vector, but got: __vec_2__bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertSToF_Scalar_FromSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
@ -227,6 +333,39 @@ TEST_F(SpvUnaryConversionTest, ConvertSToF_Vector_FromUnsigned) {
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertUToF_Scalar_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertUToF %float %false
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(), Eq("operand for conversion to floating point must be "
"integral scalar or vector, but got: __bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertUToF_Vector_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertUToF %v2float %v2bool_t_f
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
Eq("operand for conversion to floating point must be integral "
"scalar or vector, but got: __vec_2__bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertUToF_Scalar_FromSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
@ -335,14 +474,296 @@ TEST_F(SpvUnaryConversionTest, ConvertUToF_Vector_FromUnsigned) {
<< ToString(fe.ast_body());
}
// TODO(dneto): OpConvertFToU
// TODO(dneto): OpConvertFToS
TEST_F(SpvUnaryConversionTest, ConvertFToS_Scalar_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToS %int %uint_10
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
Eq("operand for conversion to signed integer must be floating "
"point scalar or vector, but got: __u32"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_Vector_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToS %v2float %v2bool_t_f
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
Eq("operand for conversion to signed integer must be floating "
"point scalar or vector, but got: __vec_2__bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_Scalar_ToSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %float %float_50
%1 = OpConvertFToS %int %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__i32
{
Cast<__i32>(
Identifier{x_30}
)
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_Scalar_ToUnsigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %float %float_50
%1 = OpConvertFToS %uint %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__u32
{
As<__u32>{
Cast<__i32>(
Identifier{x_30}
)
}
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_Vector_ToSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %v2float %v2float_50_60
%1 = OpConvertFToS %v2int %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__vec_2__i32
{
Cast<__vec_2__i32>(
Identifier{x_30}
)
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToS_Vector_ToUnsigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %v2float %v2float_50_60
%1 = OpConvertFToS %v2uint %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__vec_2__u32
{
As<__vec_2__u32>{
Cast<__vec_2__i32>(
Identifier{x_30}
)
}
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Scalar_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToU %int %uint_10
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
Eq("operand for conversion to unsigned integer must be floating "
"point scalar or vector, but got: __u32"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Vector_BadArgType) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%1 = OpConvertFToU %v2float %v2bool_t_f
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_FALSE(fe.EmitBody());
EXPECT_THAT(p->error(),
Eq("operand for conversion to unsigned integer must be floating "
"point scalar or vector, but got: __vec_2__bool"));
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Scalar_ToSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %float %float_50
%1 = OpConvertFToU %int %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__i32
{
As<__i32>{
Cast<__u32>(
Identifier{x_30}
)
}
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Scalar_ToUnsigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %float %float_50
%1 = OpConvertFToU %uint %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__u32
{
Cast<__u32>(
Identifier{x_30}
)
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Vector_ToSigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %v2float %v2float_50_60
%1 = OpConvertFToU %v2int %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__vec_2__i32
{
As<__vec_2__i32>{
Cast<__vec_2__u32>(
Identifier{x_30}
)
}
}
})"))
<< ToString(fe.ast_body());
}
TEST_F(SpvUnaryConversionTest, ConvertFToU_Vector_ToUnsigned) {
const auto assembly = CommonTypes() + R"(
%100 = OpFunction %void None %voidfn
%entry = OpLabel
%30 = OpCopyObject %v2float %v2float_50_60
%1 = OpConvertFToU %v2uint %30
OpReturn
OpFunctionEnd
)";
auto* p = parser(test::Assemble(assembly));
ASSERT_TRUE(p->BuildAndParseInternalModuleExceptFunctions());
FunctionEmitter fe(p, *spirv_function(100));
EXPECT_TRUE(fe.EmitBody()) << p->error();
EXPECT_THAT(ToString(fe.ast_body()), HasSubstr(R"(Variable{
x_1
none
__vec_2__u32
{
Cast<__vec_2__u32>(
Identifier{x_30}
)
}
})"))
<< ToString(fe.ast_body());
}
// TODO(dneto): OpSConvert // only if multiple widths
// TODO(dneto): OpUConvert // only if multiple widths
// TODO(dneto): OpFConvert // only if multiple widths
// TODO(dneto): OpQuantizeToF16 // only if f16 supported
// TODO(dneto): OpConvertSToU
// TODO(dneto): OpConvertUToS
// TODO(dneto): OpSatConvertSToU // Kernel (OpenCL), not in WebGPU
// TODO(dneto): OpSatConvertUToS // Kernel (OpenCL), not in WebGPU
} // namespace
} // namespace spirv

View File

@ -1087,6 +1087,54 @@ ast::type::Type* ParserImpl::ForcedResultType(
return nullptr;
}
ast::type::Type* ParserImpl::GetSignedIntMatchingShape(ast::type::Type* other) {
if (other == nullptr) {
Fail() << "no type provided";
}
auto* i32 = ctx_.type_mgr().Get(std::make_unique<ast::type::I32Type>());
if (other->IsF32() || other->IsU32() || other->IsI32()) {
return i32;
}
auto* vec_ty = other->AsVector();
if (vec_ty) {
return ctx_.type_mgr().Get(
std::make_unique<ast::type::VectorType>(i32, vec_ty->size()));
}
Fail() << "required numeric scalar or vector, but got " << other->type_name();
return nullptr;
}
ast::type::Type* ParserImpl::GetUnsignedIntMatchingShape(
ast::type::Type* other) {
if (other == nullptr) {
Fail() << "no type provided";
return nullptr;
}
auto* u32 = ctx_.type_mgr().Get(std::make_unique<ast::type::U32Type>());
if (other->IsF32() || other->IsU32() || other->IsI32()) {
return u32;
}
auto* vec_ty = other->AsVector();
if (vec_ty) {
return ctx_.type_mgr().Get(
std::make_unique<ast::type::VectorType>(u32, vec_ty->size()));
}
Fail() << "required numeric scalar or vector, but got " << other->type_name();
return nullptr;
}
TypedExpression ParserImpl::RectifyForcedResultType(
TypedExpression expr,
SpvOp op,
ast::type::Type* first_operand_type) {
auto* forced_result_ty = ForcedResultType(op, first_operand_type);
if ((forced_result_ty == nullptr) || (forced_result_ty == expr.type)) {
return expr;
}
return {expr.type,
std::make_unique<ast::AsExpression>(expr.type, std::move(expr.expr))};
}
bool ParserImpl::EmitFunctions() {
if (!success_) {
return false;

View File

@ -280,6 +280,34 @@ class ParserImpl : Reader {
ast::type::Type* ForcedResultType(SpvOp op,
ast::type::Type* first_operand_type);
/// Returns a signed integer scalar or vector type matching the shape (scalar,
/// vector, and component bit width) of another type, which itself is a
/// numeric scalar or vector. Returns null if the other type does not meet the
/// requirement.
/// @param other the type whose shape must be matched
/// @returns the signed scalar or vector type
ast::type::Type* GetSignedIntMatchingShape(ast::type::Type* other);
/// Returns a signed integer scalar or vector type matching the shape (scalar,
/// vector, and component bit width) of another type, which itself is a
/// numeric scalar or vector. Returns null if the other type does not meet the
/// requirement.
/// @param other the type whose shape must be matched
/// @returns the unsigned scalar or vector type
ast::type::Type* GetUnsignedIntMatchingShape(ast::type::Type* other);
/// Wraps the given expression in an as-cast to the given expression's type,
/// when the underlying operation produces a forced result type different
/// from the expression's result type. Otherwise, returns the given expression
/// unchanged.
/// @param expr the expression to pass through or to wrap
/// @param op the SPIR-V opcode
/// @param first_operand_type the AST type for the first operand.
/// @returns the forced AST result type, or nullptr if no forcing is required.
TypedExpression RectifyForcedResultType(TypedExpression expr,
SpvOp op,
ast::type::Type* first_operand_type);
/// @returns the registered boolean type.
ast::type::Type* BoolType() const { return bool_type_; }