llvm/flang/lib/semantics/expression.h

307 lines
11 KiB
C
Raw Normal View History

// Copyright (c) 2018-2019, NVIDIA CORPORATION. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#ifndef FORTRAN_SEMANTICS_EXPRESSION_H_
#define FORTRAN_SEMANTICS_EXPRESSION_H_
#include "semantics.h"
#include "../common/Fortran.h"
#include "../common/indirection.h"
#include "../evaluate/expression.h"
#include "../evaluate/tools.h"
#include "../evaluate/type.h"
#include "../parser/char-block.h"
#include "../parser/parse-tree-visitor.h"
#include "../parser/parse-tree.h"
#include <map>
#include <optional>
#include <variant>
using namespace Fortran::parser::literals;
namespace Fortran::parser {
struct SourceLocationFindingVisitor {
template<typename A> bool Pre(const A &) { return true; }
template<typename A> void Post(const A &) {}
bool Pre(const Expr &);
template<typename A> bool Pre(const Statement<A> &stmt) {
source = stmt.source;
return false;
}
void Post(const CharBlock &);
CharBlock source;
};
template<typename A> CharBlock FindSourceLocation(const A &x) {
SourceLocationFindingVisitor visitor;
Walk(x, visitor);
return visitor.source;
}
}
using namespace Fortran::parser::literals;
// The expression semantic analysis code has its implementation in
// namespace Fortran::evaluate, but the exposed API to it is in the
// namespace Fortran::semantics (below).
//
// The ExpressionAnalyzer wraps a SemanticsContext reference
// and implements constraint checking on expressions using the
// parse tree node wrappers that mirror the grammar annotations used
// in the Fortran standard (i.e., scalar-, constant-, &c.).
namespace Fortran::evaluate {
class ExpressionAnalyzer {
public:
using MaybeExpr = std::optional<Expr<SomeType>>;
explicit ExpressionAnalyzer(semantics::SemanticsContext &sc) : context_{sc} {}
ExpressionAnalyzer(ExpressionAnalyzer &) = default;
semantics::SemanticsContext &context() const { return context_; }
FoldingContext &GetFoldingContext() const {
return context_.foldingContext();
}
parser::ContextualMessages &GetContextualMessages() {
return GetFoldingContext().messages();
}
template<typename... A> parser::Message *Say(A... args) {
return GetContextualMessages().Say(std::forward<A>(args)...);
}
template<typename T, typename... A>
parser::Message *SayAt(const T &parsed, A... args) {
return Say(parser::FindSourceLocation(parsed), std::forward<A>(args)...);
}
int GetDefaultKind(common::TypeCategory);
DynamicType GetDefaultKindOfType(common::TypeCategory);
// Return false and emit error if these checks fail:
bool CheckIntrinsicKind(TypeCategory, std::int64_t kind);
bool CheckIntrinsicSize(TypeCategory, std::int64_t size);
// Manage a set of active array constructor implied DO loops.
bool AddAcImpliedDo(parser::CharBlock, int);
void RemoveAcImpliedDo(parser::CharBlock);
std::optional<int> IsAcImpliedDo(parser::CharBlock) const;
Expr<SubscriptInteger> AnalyzeKindSelector(common::TypeCategory category,
const std::optional<parser::KindSelector> &);
MaybeExpr Analyze(const parser::Expr &);
MaybeExpr Analyze(const parser::Variable &);
template<typename A> MaybeExpr Analyze(const common::Indirection<A> &x) {
return Analyze(x.value());
}
template<typename A> MaybeExpr Analyze(const std::optional<A> &x) {
if (x.has_value()) {
return Analyze(*x);
} else {
return std::nullopt;
}
}
// Implement constraint-checking wrappers from the Fortran grammar
template<typename A> MaybeExpr Analyze(const parser::Scalar<A> &x) {
auto result{Analyze(x.thing)};
if (result.has_value()) {
if (int rank{result->Rank()}; rank != 0) {
SayAt(x, "Must be a scalar value, but is a rank-%d array"_err_en_US,
rank);
}
}
return result;
}
template<typename A> MaybeExpr Analyze(const parser::Constant<A> &x) {
auto result{Analyze(x.thing)};
if (result.has_value()) {
*result = Fold(GetFoldingContext(), std::move(*result));
if (!IsConstantExpr(*result)) {
SayAt(x, "Must be a constant value"_err_en_US);
}
}
return result;
}
template<typename A> MaybeExpr Analyze(const parser::Integer<A> &x) {
auto result{Analyze(x.thing)};
if (result.has_value()) {
if (!std::holds_alternative<Expr<SomeInteger>>(result->u)) {
SayAt(x, "Must have INTEGER type"_err_en_US);
}
}
return result;
}
template<typename A> MaybeExpr Analyze(const parser::Logical<A> &x) {
auto result{Analyze(x.thing)};
if (result.has_value()) {
if (!std::holds_alternative<Expr<SomeLogical>>(result->u)) {
SayAt(x, "Must have LOGICAL type"_err_en_US);
}
}
return result;
}
template<typename A> MaybeExpr Analyze(const parser::DefaultChar<A> &x) {
auto result{Analyze(x.thing)};
if (result.has_value()) {
if (auto *charExpr{std::get_if<Expr<SomeCharacter>>(&result->u)}) {
if (charExpr->GetKind() ==
context().defaultKinds().GetDefaultKind(TypeCategory::Character)) {
return result;
}
}
SayAt(x, "Must have default CHARACTER type"_err_en_US);
}
return result;
}
protected:
int IntegerTypeSpecKind(const parser::IntegerTypeSpec &);
private:
MaybeExpr Analyze(const parser::Designator &);
MaybeExpr Analyze(const parser::IntLiteralConstant &);
MaybeExpr Analyze(const parser::SignedIntLiteralConstant &);
MaybeExpr Analyze(const parser::RealLiteralConstant &);
MaybeExpr Analyze(const parser::SignedRealLiteralConstant &);
MaybeExpr Analyze(const parser::ComplexPart &);
MaybeExpr Analyze(const parser::ComplexLiteralConstant &);
MaybeExpr Analyze(const parser::LogicalLiteralConstant &);
MaybeExpr Analyze(const parser::CharLiteralConstant &);
MaybeExpr Analyze(const parser::HollerithLiteralConstant &);
MaybeExpr Analyze(const parser::BOZLiteralConstant &);
MaybeExpr Analyze(const parser::Name &);
MaybeExpr Analyze(const parser::NamedConstant &);
MaybeExpr Analyze(const parser::Substring &);
MaybeExpr Analyze(const parser::ArrayElement &);
MaybeExpr Analyze(const parser::StructureComponent &);
MaybeExpr Analyze(const parser::CoindexedNamedObject &);
MaybeExpr Analyze(const parser::CharLiteralConstantSubstring &);
MaybeExpr Analyze(const parser::ArrayConstructor &);
MaybeExpr Analyze(const parser::StructureConstructor &);
MaybeExpr Analyze(const parser::FunctionReference &);
MaybeExpr Analyze(const parser::Expr::Parentheses &);
MaybeExpr Analyze(const parser::Expr::UnaryPlus &);
MaybeExpr Analyze(const parser::Expr::Negate &);
MaybeExpr Analyze(const parser::Expr::NOT &);
MaybeExpr Analyze(const parser::Expr::PercentLoc &);
MaybeExpr Analyze(const parser::Expr::DefinedUnary &);
MaybeExpr Analyze(const parser::Expr::Power &);
MaybeExpr Analyze(const parser::Expr::Multiply &);
MaybeExpr Analyze(const parser::Expr::Divide &);
MaybeExpr Analyze(const parser::Expr::Add &);
MaybeExpr Analyze(const parser::Expr::Subtract &);
MaybeExpr Analyze(const parser::Expr::ComplexConstructor &);
MaybeExpr Analyze(const parser::Expr::Concat &);
MaybeExpr Analyze(const parser::Expr::LT &);
MaybeExpr Analyze(const parser::Expr::LE &);
MaybeExpr Analyze(const parser::Expr::EQ &);
MaybeExpr Analyze(const parser::Expr::NE &);
MaybeExpr Analyze(const parser::Expr::GE &);
MaybeExpr Analyze(const parser::Expr::GT &);
MaybeExpr Analyze(const parser::Expr::AND &);
MaybeExpr Analyze(const parser::Expr::OR &);
MaybeExpr Analyze(const parser::Expr::EQV &);
MaybeExpr Analyze(const parser::Expr::NEQV &);
MaybeExpr Analyze(const parser::Expr::XOR &);
MaybeExpr Analyze(const parser::Expr::DefinedBinary &);
template<typename A> MaybeExpr Analyze(const A &x) {
return Analyze(x.u); // default case
}
template<typename... As> MaybeExpr Analyze(const std::variant<As...> &u) {
return std::visit([&](const auto &x) { return Analyze(x); }, u);
}
// Analysis subroutines
int AnalyzeKindParam(const std::optional<parser::KindParam> &,
int defaultKind, int kanjiKind = -1);
template<typename PARSED> MaybeExpr IntLiteralConstant(const PARSED &);
MaybeExpr AnalyzeString(std::string &&, int kind);
std::optional<Expr<SubscriptInteger>> AsSubscript(MaybeExpr &&);
std::optional<Expr<SubscriptInteger>> TripletPart(
const std::optional<parser::Subscript> &);
std::optional<Subscript> AnalyzeSectionSubscript(
const parser::SectionSubscript &);
std::vector<Subscript> AnalyzeSectionSubscripts(
const std::list<parser::SectionSubscript> &);
MaybeExpr CompleteSubscripts(ArrayRef &&);
MaybeExpr ApplySubscripts(DataRef &&, std::vector<Subscript> &&);
MaybeExpr TopLevelChecks(DataRef &&);
std::optional<Expr<SubscriptInteger>> GetSubstringBound(
const std::optional<parser::ScalarIntExpr> &);
struct CallAndArguments {
ProcedureDesignator procedureDesignator;
ActualArguments arguments;
};
std::optional<CallAndArguments> Procedure(
const parser::ProcedureDesignator &, ActualArguments &);
semantics::SemanticsContext &context_;
std::map<parser::CharBlock, int> acImpliedDos_; // values are INTEGER kinds
};
template<typename L, typename R>
bool AreConformable(const L &left, const R &right) {
int leftRank{left.Rank()};
if (leftRank == 0) {
return true;
}
int rightRank{right.Rank()};
return rightRank == 0 || leftRank == rightRank;
}
template<typename L, typename R>
void ConformabilityCheck(
parser::ContextualMessages &context, const L &left, const R &right) {
if (!AreConformable(left, right)) {
context.Say("left operand has rank %d, right operand has rank %d"_err_en_US,
left.Rank(), right.Rank());
}
}
[flang] Create framework for checking statement semantics Add `SemanticsVisitor` as the visitor class to perform statement semantics checks. Its template parameters are "checker" classes that perform the checks. They have `Enter` and `Leave` functions that are called for the corresponding parse tree nodes (`Enter` before the children, `Leave` after). Unlike `Pre` and `Post` in visitors they cannot prevent the parse tree walker from visiting child nodes. Existing checks have been incorporated into this framework: - `ExprChecker` replaces `AnalyzeExpressions()` - `AssignmentChecker` replaces `AnalyzeAssignments()` - `DoConcurrentChecker` replaces `CheckDoConcurrentConstraints()` Adding a new checker requires: - defining the checker class: - with BaseChecker as virtual base class - constructible from `SemanticsContext` - with Enter/Leave functions for nodes of interest - add the checker class to the template parameters of `StatementSemantics` Because these checkers and also `ResolveNamesVisitor` require tracking the current statement source location, that has been moved into `SemanticsContext`. `ResolveNamesVisitor` and `SemanticsVisitor` update the location when `Statement` nodes are encountered, making it available for error messages. `AnalyzeKindSelector()` now has access to the current statement through the context and so no longer needs to have it passed in. Test `assign01.f90` was added to verify that `AssignmentChecker` is actually doing something. Original-commit: flang-compiler/f18@3a222c36731029fabf026e5301dc60f0587595be Reviewed-on: https://github.com/flang-compiler/f18/pull/315 Tree-same-pre-rewrite: false
2019-03-06 01:52:50 +01:00
} // namespace Fortran::evaluate
namespace Fortran::semantics {
// Semantic analysis of one expression.
template<typename A>
std::optional<evaluate::Expr<evaluate::SomeType>> AnalyzeExpr(
SemanticsContext &context, const A &expr) {
return evaluate::ExpressionAnalyzer{context}.Analyze(expr);
}
// Semantic analysis of an intrinsic type's KIND parameter expression.
evaluate::Expr<evaluate::SubscriptInteger> AnalyzeKindSelector(
[flang] Create framework for checking statement semantics Add `SemanticsVisitor` as the visitor class to perform statement semantics checks. Its template parameters are "checker" classes that perform the checks. They have `Enter` and `Leave` functions that are called for the corresponding parse tree nodes (`Enter` before the children, `Leave` after). Unlike `Pre` and `Post` in visitors they cannot prevent the parse tree walker from visiting child nodes. Existing checks have been incorporated into this framework: - `ExprChecker` replaces `AnalyzeExpressions()` - `AssignmentChecker` replaces `AnalyzeAssignments()` - `DoConcurrentChecker` replaces `CheckDoConcurrentConstraints()` Adding a new checker requires: - defining the checker class: - with BaseChecker as virtual base class - constructible from `SemanticsContext` - with Enter/Leave functions for nodes of interest - add the checker class to the template parameters of `StatementSemantics` Because these checkers and also `ResolveNamesVisitor` require tracking the current statement source location, that has been moved into `SemanticsContext`. `ResolveNamesVisitor` and `SemanticsVisitor` update the location when `Statement` nodes are encountered, making it available for error messages. `AnalyzeKindSelector()` now has access to the current statement through the context and so no longer needs to have it passed in. Test `assign01.f90` was added to verify that `AssignmentChecker` is actually doing something. Original-commit: flang-compiler/f18@3a222c36731029fabf026e5301dc60f0587595be Reviewed-on: https://github.com/flang-compiler/f18/pull/315 Tree-same-pre-rewrite: false
2019-03-06 01:52:50 +01:00
SemanticsContext &, common::TypeCategory,
const std::optional<parser::KindSelector> &);
[flang] Create framework for checking statement semantics Add `SemanticsVisitor` as the visitor class to perform statement semantics checks. Its template parameters are "checker" classes that perform the checks. They have `Enter` and `Leave` functions that are called for the corresponding parse tree nodes (`Enter` before the children, `Leave` after). Unlike `Pre` and `Post` in visitors they cannot prevent the parse tree walker from visiting child nodes. Existing checks have been incorporated into this framework: - `ExprChecker` replaces `AnalyzeExpressions()` - `AssignmentChecker` replaces `AnalyzeAssignments()` - `DoConcurrentChecker` replaces `CheckDoConcurrentConstraints()` Adding a new checker requires: - defining the checker class: - with BaseChecker as virtual base class - constructible from `SemanticsContext` - with Enter/Leave functions for nodes of interest - add the checker class to the template parameters of `StatementSemantics` Because these checkers and also `ResolveNamesVisitor` require tracking the current statement source location, that has been moved into `SemanticsContext`. `ResolveNamesVisitor` and `SemanticsVisitor` update the location when `Statement` nodes are encountered, making it available for error messages. `AnalyzeKindSelector()` now has access to the current statement through the context and so no longer needs to have it passed in. Test `assign01.f90` was added to verify that `AssignmentChecker` is actually doing something. Original-commit: flang-compiler/f18@3a222c36731029fabf026e5301dc60f0587595be Reviewed-on: https://github.com/flang-compiler/f18/pull/315 Tree-same-pre-rewrite: false
2019-03-06 01:52:50 +01:00
// Semantic analysis of all expressions in a parse tree, which becomes
[flang] Create framework for checking statement semantics Add `SemanticsVisitor` as the visitor class to perform statement semantics checks. Its template parameters are "checker" classes that perform the checks. They have `Enter` and `Leave` functions that are called for the corresponding parse tree nodes (`Enter` before the children, `Leave` after). Unlike `Pre` and `Post` in visitors they cannot prevent the parse tree walker from visiting child nodes. Existing checks have been incorporated into this framework: - `ExprChecker` replaces `AnalyzeExpressions()` - `AssignmentChecker` replaces `AnalyzeAssignments()` - `DoConcurrentChecker` replaces `CheckDoConcurrentConstraints()` Adding a new checker requires: - defining the checker class: - with BaseChecker as virtual base class - constructible from `SemanticsContext` - with Enter/Leave functions for nodes of interest - add the checker class to the template parameters of `StatementSemantics` Because these checkers and also `ResolveNamesVisitor` require tracking the current statement source location, that has been moved into `SemanticsContext`. `ResolveNamesVisitor` and `SemanticsVisitor` update the location when `Statement` nodes are encountered, making it available for error messages. `AnalyzeKindSelector()` now has access to the current statement through the context and so no longer needs to have it passed in. Test `assign01.f90` was added to verify that `AssignmentChecker` is actually doing something. Original-commit: flang-compiler/f18@3a222c36731029fabf026e5301dc60f0587595be Reviewed-on: https://github.com/flang-compiler/f18/pull/315 Tree-same-pre-rewrite: false
2019-03-06 01:52:50 +01:00
// decorated with typed representations for top-level expressions.
class ExprChecker : public virtual BaseChecker {
public:
explicit ExprChecker(SemanticsContext &context) : context_{context} {}
void Enter(const parser::Expr &);
void Enter(const parser::Variable &);
[flang] Create framework for checking statement semantics Add `SemanticsVisitor` as the visitor class to perform statement semantics checks. Its template parameters are "checker" classes that perform the checks. They have `Enter` and `Leave` functions that are called for the corresponding parse tree nodes (`Enter` before the children, `Leave` after). Unlike `Pre` and `Post` in visitors they cannot prevent the parse tree walker from visiting child nodes. Existing checks have been incorporated into this framework: - `ExprChecker` replaces `AnalyzeExpressions()` - `AssignmentChecker` replaces `AnalyzeAssignments()` - `DoConcurrentChecker` replaces `CheckDoConcurrentConstraints()` Adding a new checker requires: - defining the checker class: - with BaseChecker as virtual base class - constructible from `SemanticsContext` - with Enter/Leave functions for nodes of interest - add the checker class to the template parameters of `StatementSemantics` Because these checkers and also `ResolveNamesVisitor` require tracking the current statement source location, that has been moved into `SemanticsContext`. `ResolveNamesVisitor` and `SemanticsVisitor` update the location when `Statement` nodes are encountered, making it available for error messages. `AnalyzeKindSelector()` now has access to the current statement through the context and so no longer needs to have it passed in. Test `assign01.f90` was added to verify that `AssignmentChecker` is actually doing something. Original-commit: flang-compiler/f18@3a222c36731029fabf026e5301dc60f0587595be Reviewed-on: https://github.com/flang-compiler/f18/pull/315 Tree-same-pre-rewrite: false
2019-03-06 01:52:50 +01:00
private:
SemanticsContext &context_;
};
} // namespace Fortran::semantics
#endif // FORTRAN_SEMANTICS_EXPRESSION_H_