/
githubmirror
/
roslyn
Обзор
Документация
Войти
/
githubmirror
/
roslyn
Код
Запросы
0
Пакеты
0
Релизы
0
Аналитика
Безопасность
main
src/Compilers/CSharp/Portable/Binder/Semantics/Conversions/UserDefinedExplicitConversions.cs
507 строк
27 KB
Todd Grunke
Reduce work in ConversionsBase.AddUserDefinedConversionsToExplicitCandidateSet (#76835)
23 янв 2025, 19:16
Не верифицирован
23 янв 2025, 19:16
55e8ff7
Код
Авторство
О чём код?
// Licensed to the .NET Foundation under one or more agreements. // The .NET Foundation licenses this file to you under the MIT license. // See the LICENSE file in the project root for more information. #nullable disable using System.Collections.Immutable; using System.Diagnostics; using System.Linq; using Microsoft.CodeAnalysis.CSharp.Symbols; using Microsoft.CodeAnalysis.CSharp.Syntax; using Microsoft.CodeAnalysis.PooledObjects; using Microsoft.CodeAnalysis.Text; using System.Collections.Generic; namespace Microsoft.CodeAnalysis.CSharp { internal partial class ConversionsBase { private UserDefinedConversionResult AnalyzeExplicitUserDefinedConversions( BoundExpression sourceExpression, TypeSymbol source, TypeSymbol target, bool isChecked, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { Debug.Assert(sourceExpression is null || Compilation is not null); Debug.Assert(sourceExpression != null || (object)source != null); Debug.Assert((object)target != null); // SPEC: A user-defined explicit conversion from type S to type T is processed // SPEC: as follows: // SPEC: Find the set of types D from which user-defined conversion operators // SPEC: will be considered... var d = ArrayBuilder<(NamedTypeSymbol ParticipatingType, TypeParameterSymbol ConstrainedToTypeOpt)>.GetInstance(); ComputeUserDefinedExplicitConversionTypeSet(source, target, d, ref useSiteInfo); // SPEC: Find the set of applicable user-defined and lifted conversion operators, U... var ubuild = ArrayBuilder<UserDefinedConversionAnalysis>.GetInstance(); ComputeApplicableUserDefinedExplicitConversionSet(sourceExpression, source, target, isChecked: isChecked, d, ubuild, ref useSiteInfo); d.Free(); ImmutableArray<UserDefinedConversionAnalysis> u = ubuild.ToImmutableAndFree(); // SPEC: If U is empty, the conversion is undefined and a compile-time error occurs. if (u.Length == 0) { return UserDefinedConversionResult.NoApplicableOperators(u); } // SPEC: Find the most specific source type SX of the operators in U... TypeSymbol sx = MostSpecificSourceTypeForExplicitUserDefinedConversion(u, sourceExpression, source, ref useSiteInfo); if ((object)sx == null) { return UserDefinedConversionResult.NoBestSourceType(u); } // SPEC: Find the most specific target type TX of the operators in U... TypeSymbol tx = MostSpecificTargetTypeForExplicitUserDefinedConversion(u, target, ref useSiteInfo); if ((object)tx == null) { return UserDefinedConversionResult.NoBestTargetType(u); } int? best = MostSpecificConversionOperator(sx, tx, u); if (best == null) { return UserDefinedConversionResult.Ambiguous(u); } return UserDefinedConversionResult.Valid(u, best.Value); } private static void ComputeUserDefinedExplicitConversionTypeSet(TypeSymbol source, TypeSymbol target, ArrayBuilder<(NamedTypeSymbol ParticipatingType, TypeParameterSymbol ConstrainedToTypeOpt)> d, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { // Spec 6.4.5: User-defined explicit conversions // Find the set of types, D, from which user-defined conversion operators will be considered. // This set consists of S0 (if S0 is a class or struct), the base classes of S0 (if S0 is a class), // T0 (if T0 is a class or struct), and the base classes of T0 (if T0 is a class). AddTypesParticipatingInUserDefinedConversion(d, source, includeBaseTypes: true, useSiteInfo: ref useSiteInfo); AddTypesParticipatingInUserDefinedConversion(d, target, includeBaseTypes: true, useSiteInfo: ref useSiteInfo); } private void ComputeApplicableUserDefinedExplicitConversionSet( BoundExpression sourceExpression, TypeSymbol source, TypeSymbol target, bool isChecked, ArrayBuilder<(NamedTypeSymbol ParticipatingType, TypeParameterSymbol ConstrainedToTypeOpt)> d, ArrayBuilder<UserDefinedConversionAnalysis> u, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { Debug.Assert(sourceExpression is null || Compilation is not null); Debug.Assert(sourceExpression != null || (object)source != null); Debug.Assert((object)target != null); Debug.Assert(d != null); Debug.Assert(u != null); bool haveInterfaces = false; foreach ((NamedTypeSymbol declaringType, TypeParameterSymbol constrainedToTypeOpt) in d) { if (declaringType.IsInterface) { Debug.Assert(constrainedToTypeOpt is not null); haveInterfaces = true; } else { addCandidatesFromType(constrainedToTypeOpt: null, declaringType, sourceExpression, source, target, isChecked: isChecked, u, ref useSiteInfo); } } if (u.Count == 0 && haveInterfaces) { foreach ((NamedTypeSymbol declaringType, TypeParameterSymbol constrainedToTypeOpt) in d) { if (declaringType.IsInterface) { addCandidatesFromType(constrainedToTypeOpt: constrainedToTypeOpt, declaringType, sourceExpression, source, target, isChecked: isChecked, u, ref useSiteInfo); } } } void addCandidatesFromType( TypeParameterSymbol constrainedToTypeOpt, NamedTypeSymbol declaringType, BoundExpression sourceExpression, TypeSymbol source, TypeSymbol target, bool isChecked, ArrayBuilder<UserDefinedConversionAnalysis> u, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { AddUserDefinedConversionsToExplicitCandidateSet(sourceExpression, source, target, u, constrainedToTypeOpt, declaringType, isExplicit: true, isChecked: isChecked, ref useSiteInfo); AddUserDefinedConversionsToExplicitCandidateSet(sourceExpression, source, target, u, constrainedToTypeOpt, declaringType, isExplicit: false, isChecked: isChecked, ref useSiteInfo); } } private void AddUserDefinedConversionsToExplicitCandidateSet( BoundExpression sourceExpression, TypeSymbol source, TypeSymbol target, ArrayBuilder<UserDefinedConversionAnalysis> u, TypeParameterSymbol constrainedToTypeOpt, NamedTypeSymbol declaringType, bool isExplicit, bool isChecked, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { Debug.Assert(sourceExpression is null || Compilation is not null); Debug.Assert(sourceExpression != null || (object)source != null); Debug.Assert((object)target != null); Debug.Assert(u != null); Debug.Assert((object)declaringType != null); // SPEC: Find the set of applicable user-defined and lifted conversion operators, U. // SPEC: The set consists of the user-defined and lifted implicit or explicit // SPEC: conversion operators declared by the classes and structs in D that convert // SPEC: from a type encompassing E or encompassed by S (if it exists) to a type // SPEC: encompassing or encompassed by T. // DELIBERATE SPEC VIOLATION: // // The spec here essentially says that we add an applicable "regular" conversion and // an applicable lifted conversion, if there is one, to the candidate set, and then // let them duke it out to determine which one is "best". // // This is not at all what the native compiler does, and attempting to implement // the specification, or slight variations on it, produces too many backwards-compatibility // breaking changes. // // The native compiler deviates from the specification in two major ways here. // First, it does not add *both* the regular and lifted forms to the candidate set. // Second, the way it characterizes a "lifted" form is very, very different from // how the specification characterizes a lifted form. // // An operation, in this case, X-->Y, is properly said to be "lifted" to X?-->Y? via // the rule that X?-->Y? matches the behavior of X-->Y for non-null X, and converts // null X to null Y otherwise. // // The native compiler, by contrast, takes the existing operator and "lifts" either // the operator's parameter type or the operator's return type to nullable. For // example, a conversion from X?-->Y would be "lifted" to X?-->Y? by making the // conversion from X? to Y, and then from Y to Y?. No "lifting" semantics // are imposed; we do not check to see if the X? is null. This operator is not // actually "lifted" at all; rather, an implicit conversion is applied to the // output. **The native compiler considers the result type Y? of that standard implicit // conversion to be the result type of the "lifted" conversion**, rather than // properly considering Y to be the result type of the conversion for the purposes // of computing the best output type. // // Moreover: the native compiler actually *does* implement nullable lifting semantics // in the case where the input type of the user-defined conversion is a non-nullable // value type and the output type is a nullable value type **or pointer type, or // reference type**. This is an enormous departure from the specification; the // native compiler will take a user-defined conversion from X-->Y? or X-->C and "lift" // it to a conversion from X?-->Y? or X?-->C that has nullable semantics. // // This is quite confusing. In this code we will classify the conversion as either // "normal" or "lifted" on the basis of *whether or not special lifting semantics // are to be applied*. That is, whether or not a later rewriting pass is going to // need to insert a check to see if the source expression is null, and decide // whether or not to call the underlying unlifted conversion or produce a null // value without calling the unlifted conversion. // DELIBERATE SPEC VIOLATION: See the comment regarding bug 17021 in // UserDefinedImplicitConversions.cs. if ((object)source != null && source.IsInterfaceType() || target.IsInterfaceType()) { return; } if (IgnoreUserDefinedSpanConversions(source, target)) { return; } var operators = ArrayBuilder<MethodSymbol>.GetInstance(); declaringType.AddOperators( isExplicit ? (isChecked ? WellKnownMemberNames.CheckedExplicitConversionName : WellKnownMemberNames.ExplicitConversionName) : WellKnownMemberNames.ImplicitConversionName, operators); if (isExplicit && isChecked) { // Add regular operators as well. if (operators.IsEmpty) { declaringType.AddOperators(WellKnownMemberNames.ExplicitConversionName, operators); } else { var originalOperatorCount = operators.Count; var operators2 = ArrayBuilder<MethodSymbol>.GetInstance(); declaringType.AddOperators(WellKnownMemberNames.ExplicitConversionName, operators2); foreach (MethodSymbol op2 in operators2) { // Drop operators that have a match among the checked ones. bool add = true; for (var i = 0; i < originalOperatorCount; i++) { if (SourceMemberContainerTypeSymbol.DoOperatorsPair(operators[i], op2)) { add = false; break; } } if (add) { operators.Add(op2); } } operators2.Free(); } } foreach (MethodSymbol op in operators) { // We might have a bad operator and be in an error recovery situation. Ignore it. if (op.ReturnsVoid || op.ParameterCount != 1 || op.ReturnType.TypeKind == TypeKind.Error) { continue; } TypeSymbol convertsFrom = op.GetParameterType(0); TypeSymbol convertsTo = op.ReturnType; Conversion fromConversion = EncompassingExplicitConversion(sourceExpression, source, convertsFrom, ref useSiteInfo); Conversion toConversion = EncompassingExplicitConversion(convertsTo, target, ref useSiteInfo); // We accept candidates for which the parameter type encompasses the *underlying* source type. if (!fromConversion.Exists && (object)source != null && source.IsNullableType() && EncompassingExplicitConversion(source.GetNullableUnderlyingType(), convertsFrom, ref useSiteInfo).Exists) { fromConversion = ClassifyBuiltInConversion(source, convertsFrom, isChecked: isChecked, ref useSiteInfo); } // As in dev11 (and the revised spec), we also accept candidates for which the return type is encompassed by the *stripped* target type. if (!toConversion.Exists && (object)target != null && target.IsNullableType() && EncompassingExplicitConversion(convertsTo, target.GetNullableUnderlyingType(), ref useSiteInfo).Exists) { toConversion = ClassifyBuiltInConversion(convertsTo, target, isChecked: isChecked, ref useSiteInfo); } // In the corresponding implicit conversion code we can get away with first // checking to see if standard implicit conversions exist from the source type // to the parameter type, and from the return type to the target type. If not, // then we can check for a lifted operator. // // That's not going to cut it in the explicit conversion code. Suppose we have // a conversion X-->Y and have source type X? and target type Y?. There *are* // standard explicit conversions from X?-->X and Y?-->Y, but we do not want // to bind this as an *unlifted* conversion from X? to Y?; we want such a thing // to be a *lifted* conversion from X? to Y?, that checks for null on the source // and decides to not call the underlying user-defined conversion if it is null. // // We therefore cannot do what we do in the implicit conversions, where we check // to see if the unlifted conversion works, and if it does, then don't add the lifted // conversion at all. Rather, we have to see if what we're building here is a // lifted conversion or not. // // Under what circumstances is this conversion a lifted conversion? (In the // "spec" sense of a lifted conversion; that is, that we check for null // and skip the user-defined conversion if necessary). // // * The source type must be a nullable value type. // * The parameter type must be a non-nullable value type. // * The target type must be able to take on a null value. if (fromConversion.Exists && toConversion.Exists) { if ((object)source != null && source.IsNullableType() && convertsFrom.IsValidNullableTypeArgument() && target.CanBeAssignedNull()) { TypeSymbol nullableFrom = MakeNullableType(convertsFrom); TypeSymbol nullableTo = convertsTo.IsValidNullableTypeArgument() ? MakeNullableType(convertsTo) : convertsTo; Conversion liftedFromConversion = EncompassingExplicitConversion(sourceExpression, source, nullableFrom, ref useSiteInfo); Conversion liftedToConversion = EncompassingExplicitConversion(nullableTo, target, ref useSiteInfo); Debug.Assert(liftedFromConversion.Exists); Debug.Assert(liftedToConversion.Exists); u.Add(UserDefinedConversionAnalysis.Lifted(constrainedToTypeOpt, op, liftedFromConversion, liftedToConversion, nullableFrom, nullableTo)); } else { // There is an additional spec violation in the native compiler. Suppose // we have a conversion from X-->Y and are asked to do "Y? y = new X();" Clearly // the intention is to convert from X-->Y via the implicit conversion, and then // stick a standard implicit conversion from Y-->Y? on the back end. **In this // situation, the native compiler treats the conversion as though it were // actually X-->Y? in source for the purposes of determining the best target // type of a set of operators. // // Similarly, if we have a conversion from X-->Y and are asked to do // an explicit conversion from X? to Y then we treat the conversion as // though it really were X?-->Y for the purposes of determining the best // source type of a set of operators. // // We perpetuate these fictions here, except when X or Y is not a valid // type argument to `Nullable<T>`. if (target.IsNullableType() && convertsTo.IsValidNullableTypeArgument()) { convertsTo = MakeNullableType(convertsTo); toConversion = EncompassingExplicitConversion(convertsTo, target, ref useSiteInfo); } if ((object)source != null && source.IsNullableType() && convertsFrom.IsValidNullableTypeArgument()) { convertsFrom = MakeNullableType(convertsFrom); fromConversion = EncompassingExplicitConversion(convertsFrom, source, ref useSiteInfo); } u.Add(UserDefinedConversionAnalysis.Normal(constrainedToTypeOpt, op, fromConversion, toConversion, convertsFrom, convertsTo)); } } } operators.Free(); } private TypeSymbol MostSpecificSourceTypeForExplicitUserDefinedConversion( ImmutableArray<UserDefinedConversionAnalysis> u, BoundExpression sourceExpression, TypeSymbol source, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { // SPEC: If any of the operators in U convert from S then SX is S. // SPEC: Otherwise, if any of the operators in U convert from types // SPEC: that encompass E, then SX is the most encompassed type // SPEC: in the combined set of the source types of these operators. // SPEC: If no most encompassed type can be found then the // SPEC: conversion is ambiguous and a compile-time error occurs. // SPEC: Otherwise, SX is the most encompassing type in the combined // SPEC: set of the source types of these operators. If exactly one // SPEC: most encompassing type cannot be found then the conversion // SPEC: is ambiguous and a compile-time error occurs. // DELIBERATE SPEC VIOLATION: // The native compiler deviates from the specification in the way it // determines what the "converts from" type is. The specification is pretty // clear that the "converts from" type is the actual parameter type of the // conversion operator, or, in the case of a lifted operator, the lifted-to- // nullable type. That is, if we have X-->Y then the converts-to type of // the operator in its normal form is Y, and the converts-to type of the // operator in its lifted form is Y?. // // The native compiler does not do this. Suppose we have a user-defined // conversion X-->Y, and the cast (Y)(some_nullable_x). The native // compiler will consider the converts-from type of X-->Y to be X?, surprisingly // enough. // // We have already written the "FromType" into the conversion analysis to // perpetuate this fiction. if ((object)source != null) { if (u.Any(static (conv, source) => TypeSymbol.Equals(conv.FromType, source, TypeCompareKind.ConsiderEverything2), source)) { return source; } CompoundUseSiteInfo<AssemblySymbol> inLambdaUseSiteInfo = useSiteInfo; System.Func<UserDefinedConversionAnalysis, bool> isValid = conv => IsEncompassedBy(sourceExpression, source, conv.FromType, ref inLambdaUseSiteInfo); if (u.Any(isValid)) { var result = MostEncompassedType(u, isValid, conv => conv.FromType, ref inLambdaUseSiteInfo); useSiteInfo = inLambdaUseSiteInfo; return result; } useSiteInfo = inLambdaUseSiteInfo; } return MostEncompassingType(u, conv => conv.FromType, ref useSiteInfo); } private TypeSymbol MostSpecificTargetTypeForExplicitUserDefinedConversion( ImmutableArray<UserDefinedConversionAnalysis> u, TypeSymbol target, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { // SPEC: If any of the operators in U convert to T then TX is T. // SPEC: Otherwise, if any of the operators in U convert to types that are // SPEC: encompassed by T then TX is the most encompassing type in the combined // SPEC: set of target types of those operators. If exactly one most encompassing // SPEC: type cannot be found then the conversion is ambiguous and a compile-time // SPEC: error occurs. // SPEC: Otherwise, Tx is the most encompassed type in the combined set of target // SPEC: types of the operators in U. If no most encompassed type can be found, // SPEC: then the conversion is ambiguous and a compile-time error occurs. // DELIBERATE SPEC VIOLATION: // The native compiler deviates from the specification in the way it // determines what the "converts to" type is. The specification is pretty // clear that the "converts to" type is the actual return type of the // conversion operator, or, in the case of a lifted operator, the lifted-to- // nullable type. That is, if we have X-->Y then the converts-to type of // the operator in its normal form is Y, and the converts-to type of the // operator in its lifted form is Y?. // // The native compiler does not do this. Suppose we have a user-defined // conversion X-->Y, and the assignment Y? y = new X(); -- the native // compiler will consider the converts-to type of X-->Y to be Y?, surprisingly // enough. // // We have already written the "ToType" into the conversion analysis to // perpetuate this fiction. if (u.Any(static (conv, target) => TypeSymbol.Equals(conv.ToType, target, TypeCompareKind.ConsiderEverything2), target)) { return target; } CompoundUseSiteInfo<AssemblySymbol> inLambdaUseSiteInfo = useSiteInfo; System.Func<UserDefinedConversionAnalysis, bool> isValid = conv => IsEncompassedBy(conv.ToType, target, ref inLambdaUseSiteInfo); if (u.Any(isValid)) { var result = MostEncompassingType(u, isValid, conv => conv.ToType, ref inLambdaUseSiteInfo); useSiteInfo = inLambdaUseSiteInfo; return result; } useSiteInfo = inLambdaUseSiteInfo; return MostEncompassedType(u, conv => conv.ToType, ref useSiteInfo); } private Conversion EncompassingExplicitConversion(BoundExpression expr, TypeSymbol a, TypeSymbol b, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { Debug.Assert(expr is null || Compilation is not null); Debug.Assert(expr != null || (object)a != null); Debug.Assert((object)b != null); // SPEC: If a standard implicit conversion exists from a type A to a type B // SPEC: and if neither A nor B is an interface type then A is said to be // SPEC: encompassed by B, and B is said to encompass A. // DELIBERATE SPEC VIOLATION: We should be checking to see if A and B are // interface types here. See the comment regarding bug 17021 in // UserDefinedImplicitConversions.cs // DELIBERATE SPEC VIOLATION: // We do not support an encompassing implicit conversion from a zero constant // to an enum type, because the native compiler did not. It would be a breaking // change. var result = ClassifyStandardConversion(expr, a, b, ref useSiteInfo); return result.IsEnumeration ? Conversion.NoConversion : result; } private Conversion EncompassingExplicitConversion(TypeSymbol a, TypeSymbol b, ref CompoundUseSiteInfo<AssemblySymbol> useSiteInfo) { return EncompassingExplicitConversion(expr: null, a, b, ref useSiteInfo); } } }