
This is a major change on how we represent nested name qualifications in the AST. * The nested name specifier itself and how it's stored is changed. The prefixes for types are handled within the type hierarchy, which makes canonicalization for them super cheap, no memory allocation required. Also translating a type into nested name specifier form becomes a no-op. An identifier is stored as a DependentNameType. The nested name specifier gains a lightweight handle class, to be used instead of passing around pointers, which is similar to what is implemented for TemplateName. There is still one free bit available, and this handle can be used within a PointerUnion and PointerIntPair, which should keep bit-packing aficionados happy. * The ElaboratedType node is removed, all type nodes in which it could previously apply to can now store the elaborated keyword and name qualifier, tail allocating when present. * TagTypes can now point to the exact declaration found when producing these, as opposed to the previous situation of there only existing one TagType per entity. This increases the amount of type sugar retained, and can have several applications, for example in tracking module ownership, and other tools which care about source file origins, such as IWYU. These TagTypes are lazily allocated, in order to limit the increase in AST size. This patch offers a great performance benefit. It greatly improves compilation time for [stdexec](https://github.com/NVIDIA/stdexec). For one datapoint, for `test_on2.cpp` in that project, which is the slowest compiling test, this patch improves `-c` compilation time by about 7.2%, with the `-fsyntax-only` improvement being at ~12%. This has great results on compile-time-tracker as well:  This patch also further enables other optimziations in the future, and will reduce the performance impact of template specialization resugaring when that lands. It has some other miscelaneous drive-by fixes. About the review: Yes the patch is huge, sorry about that. Part of the reason is that I started by the nested name specifier part, before the ElaboratedType part, but that had a huge performance downside, as ElaboratedType is a big performance hog. I didn't have the steam to go back and change the patch after the fact. There is also a lot of internal API changes, and it made sense to remove ElaboratedType in one go, versus removing it from one type at a time, as that would present much more churn to the users. Also, the nested name specifier having a different API avoids missing changes related to how prefixes work now, which could make existing code compile but not work. How to review: The important changes are all in `clang/include/clang/AST` and `clang/lib/AST`, with also important changes in `clang/lib/Sema/TreeTransform.h`. The rest and bulk of the changes are mostly consequences of the changes in API. PS: TagType::getDecl is renamed to `getOriginalDecl` in this patch, just for easier to rebasing. I plan to rename it back after this lands. Fixes #136624 Fixes https://github.com/llvm/llvm-project/issues/43179 Fixes https://github.com/llvm/llvm-project/issues/68670 Fixes https://github.com/llvm/llvm-project/issues/92757
2321 lines
89 KiB
C++
2321 lines
89 KiB
C++
//===--- EasilySwappableParametersCheck.cpp - clang-tidy ------------------===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "EasilySwappableParametersCheck.h"
|
|
#include "../utils/OptionsUtils.h"
|
|
#include "clang/AST/ASTContext.h"
|
|
#include "clang/AST/RecursiveASTVisitor.h"
|
|
#include "clang/ASTMatchers/ASTMatchFinder.h"
|
|
#include "clang/Lex/Lexer.h"
|
|
#include "llvm/ADT/SmallSet.h"
|
|
|
|
#define DEBUG_TYPE "EasilySwappableParametersCheck"
|
|
#include "llvm/Support/Debug.h"
|
|
#include <optional>
|
|
|
|
namespace optutils = clang::tidy::utils::options;
|
|
|
|
/// The default value for the MinimumLength check option.
|
|
static constexpr std::size_t DefaultMinimumLength = 2;
|
|
|
|
/// The default value for ignored parameter names.
|
|
static constexpr llvm::StringLiteral DefaultIgnoredParameterNames = "\"\";"
|
|
"iterator;"
|
|
"Iterator;"
|
|
"begin;"
|
|
"Begin;"
|
|
"end;"
|
|
"End;"
|
|
"first;"
|
|
"First;"
|
|
"last;"
|
|
"Last;"
|
|
"lhs;"
|
|
"LHS;"
|
|
"rhs;"
|
|
"RHS";
|
|
|
|
/// The default value for ignored parameter type suffixes.
|
|
static constexpr llvm::StringLiteral DefaultIgnoredParameterTypeSuffixes =
|
|
"bool;"
|
|
"Bool;"
|
|
"_Bool;"
|
|
"it;"
|
|
"It;"
|
|
"iterator;"
|
|
"Iterator;"
|
|
"inputit;"
|
|
"InputIt;"
|
|
"forwardit;"
|
|
"ForwardIt;"
|
|
"bidirit;"
|
|
"BidirIt;"
|
|
"constiterator;"
|
|
"const_iterator;"
|
|
"Const_Iterator;"
|
|
"Constiterator;"
|
|
"ConstIterator;"
|
|
"RandomIt;"
|
|
"randomit;"
|
|
"random_iterator;"
|
|
"ReverseIt;"
|
|
"reverse_iterator;"
|
|
"reverse_const_iterator;"
|
|
"ConstReverseIterator;"
|
|
"Const_Reverse_Iterator;"
|
|
"const_reverse_iterator;"
|
|
"Constreverseiterator;"
|
|
"constreverseiterator";
|
|
|
|
/// The default value for the QualifiersMix check option.
|
|
static constexpr bool DefaultQualifiersMix = false;
|
|
|
|
/// The default value for the ModelImplicitConversions check option.
|
|
static constexpr bool DefaultModelImplicitConversions = true;
|
|
|
|
/// The default value for suppressing diagnostics about parameters that are
|
|
/// used together.
|
|
static constexpr bool DefaultSuppressParametersUsedTogether = true;
|
|
|
|
/// The default value for the NamePrefixSuffixSilenceDissimilarityTreshold
|
|
/// check option.
|
|
static constexpr std::size_t
|
|
DefaultNamePrefixSuffixSilenceDissimilarityTreshold = 1;
|
|
|
|
using namespace clang::ast_matchers;
|
|
|
|
namespace clang::tidy::bugprone {
|
|
|
|
using TheCheck = EasilySwappableParametersCheck;
|
|
|
|
namespace filter {
|
|
class SimilarlyUsedParameterPairSuppressor;
|
|
|
|
static bool isIgnoredParameter(const TheCheck &Check, const ParmVarDecl *Node);
|
|
static inline bool
|
|
isSimilarlyUsedParameter(const SimilarlyUsedParameterPairSuppressor &Suppressor,
|
|
const ParmVarDecl *Param1, const ParmVarDecl *Param2);
|
|
static bool prefixSuffixCoverUnderThreshold(std::size_t Threshold,
|
|
StringRef Str1, StringRef Str2);
|
|
} // namespace filter
|
|
|
|
namespace model {
|
|
|
|
/// The language features involved in allowing the mix between two parameters.
|
|
enum class MixFlags : unsigned char {
|
|
Invalid = 0, ///< Sentinel bit pattern. DO NOT USE!
|
|
|
|
/// Certain constructs (such as pointers to noexcept/non-noexcept functions)
|
|
/// have the same CanonicalType, which would result in false positives.
|
|
/// During the recursive modelling call, this flag is set if a later diagnosed
|
|
/// canonical type equivalence should be thrown away.
|
|
WorkaroundDisableCanonicalEquivalence = 1,
|
|
|
|
None = 2, ///< Mix between the two parameters is not possible.
|
|
Trivial = 4, ///< The two mix trivially, and are the exact same type.
|
|
Canonical = 8, ///< The two mix because the types refer to the same
|
|
/// CanonicalType, but we do not elaborate as to how.
|
|
TypeAlias = 16, ///< The path from one type to the other involves
|
|
/// desugaring type aliases.
|
|
ReferenceBind = 32, ///< The mix involves the binding power of "const &".
|
|
Qualifiers = 64, ///< The mix involves change in the qualifiers.
|
|
ImplicitConversion = 128, ///< The mixing of the parameters is possible
|
|
/// through implicit conversions between the types.
|
|
|
|
LLVM_MARK_AS_BITMASK_ENUM(/* LargestValue =*/ImplicitConversion)
|
|
};
|
|
LLVM_ENABLE_BITMASK_ENUMS_IN_NAMESPACE();
|
|
|
|
/// Returns whether the SearchedFlag is turned on in the Data.
|
|
static inline bool hasFlag(MixFlags Data, MixFlags SearchedFlag) {
|
|
assert(SearchedFlag != MixFlags::Invalid &&
|
|
"can't be used to detect lack of all bits!");
|
|
|
|
// "Data & SearchedFlag" would need static_cast<bool>() in conditions.
|
|
return (Data & SearchedFlag) == SearchedFlag;
|
|
}
|
|
|
|
#ifndef NDEBUG
|
|
|
|
// The modelling logic of this check is more complex than usual, and
|
|
// potentially hard to understand without the ability to see into the
|
|
// representation during the recursive descent. This debug code is only
|
|
// compiled in 'Debug' mode, or if LLVM_ENABLE_ASSERTIONS config is turned on.
|
|
|
|
/// Formats the MixFlags enum into a useful, user-readable representation.
|
|
static inline std::string formatMixFlags(MixFlags F) {
|
|
if (F == MixFlags::Invalid)
|
|
return "#Inv!";
|
|
|
|
SmallString<8> Str{"-------"};
|
|
|
|
if (hasFlag(F, MixFlags::None))
|
|
// Shows the None bit explicitly, as it can be applied in the recursion
|
|
// even if other bits are set.
|
|
Str[0] = '!';
|
|
if (hasFlag(F, MixFlags::Trivial))
|
|
Str[1] = 'T';
|
|
if (hasFlag(F, MixFlags::Canonical))
|
|
Str[2] = 'C';
|
|
if (hasFlag(F, MixFlags::TypeAlias))
|
|
Str[3] = 't';
|
|
if (hasFlag(F, MixFlags::ReferenceBind))
|
|
Str[4] = '&';
|
|
if (hasFlag(F, MixFlags::Qualifiers))
|
|
Str[5] = 'Q';
|
|
if (hasFlag(F, MixFlags::ImplicitConversion))
|
|
Str[6] = 'i';
|
|
|
|
if (hasFlag(F, MixFlags::WorkaroundDisableCanonicalEquivalence))
|
|
Str.append("(~C)");
|
|
|
|
return Str.str().str();
|
|
}
|
|
|
|
#endif // NDEBUG
|
|
|
|
/// The results of the steps of an Implicit Conversion Sequence is saved in
|
|
/// an instance of this record.
|
|
///
|
|
/// A ConversionSequence maps the steps of the conversion with a member for
|
|
/// each type involved in the conversion. Imagine going from a hypothetical
|
|
/// Complex class to projecting it to the real part as a const double.
|
|
///
|
|
/// I.e., given:
|
|
///
|
|
/// struct Complex {
|
|
/// operator double() const;
|
|
/// };
|
|
///
|
|
/// void functionBeingAnalysed(Complex C, const double R);
|
|
///
|
|
/// we will get the following sequence:
|
|
///
|
|
/// (Begin=) Complex
|
|
///
|
|
/// The first standard conversion is a qualification adjustment.
|
|
/// (AfterFirstStandard=) const Complex
|
|
///
|
|
/// Then the user-defined conversion is executed.
|
|
/// (UDConvOp.ConversionOperatorResultType=) double
|
|
///
|
|
/// Then this 'double' is qualifier-adjusted to 'const double'.
|
|
/// (AfterSecondStandard=) double
|
|
///
|
|
/// The conversion's result has now been calculated, so it ends here.
|
|
/// (End=) double.
|
|
///
|
|
/// Explicit storing of Begin and End in this record is needed, because
|
|
/// getting to what Begin and End here are needs further resolution of types,
|
|
/// e.g. in the case of typedefs:
|
|
///
|
|
/// using Comp = Complex;
|
|
/// using CD = const double;
|
|
/// void functionBeingAnalysed2(Comp C, CD R);
|
|
///
|
|
/// In this case, the user will be diagnosed with a potential conversion
|
|
/// between the two typedefs as written in the code, but to elaborate the
|
|
/// reasoning behind this conversion, we also need to show what the typedefs
|
|
/// mean. See FormattedConversionSequence towards the bottom of this file!
|
|
struct ConversionSequence {
|
|
enum UserDefinedConversionKind { UDCK_None, UDCK_Ctor, UDCK_Oper };
|
|
|
|
struct UserDefinedConvertingConstructor {
|
|
const CXXConstructorDecl *Fun;
|
|
QualType ConstructorParameterType;
|
|
QualType UserDefinedType;
|
|
};
|
|
|
|
struct UserDefinedConversionOperator {
|
|
const CXXConversionDecl *Fun;
|
|
QualType UserDefinedType;
|
|
QualType ConversionOperatorResultType;
|
|
};
|
|
|
|
/// The type the conversion stared from.
|
|
QualType Begin;
|
|
|
|
/// The intermediate type after the first Standard Conversion Sequence.
|
|
QualType AfterFirstStandard;
|
|
|
|
/// The details of the user-defined conversion involved, as a tagged union.
|
|
union {
|
|
char None;
|
|
UserDefinedConvertingConstructor UDConvCtor;
|
|
UserDefinedConversionOperator UDConvOp;
|
|
};
|
|
UserDefinedConversionKind UDConvKind;
|
|
|
|
/// The intermediate type after performing the second Standard Conversion
|
|
/// Sequence.
|
|
QualType AfterSecondStandard;
|
|
|
|
/// The result type the conversion targeted.
|
|
QualType End;
|
|
|
|
ConversionSequence() : None(0), UDConvKind(UDCK_None) {}
|
|
ConversionSequence(QualType From, QualType To)
|
|
: Begin(From), None(0), UDConvKind(UDCK_None), End(To) {}
|
|
|
|
explicit operator bool() const {
|
|
return !AfterFirstStandard.isNull() || UDConvKind != UDCK_None ||
|
|
!AfterSecondStandard.isNull();
|
|
}
|
|
|
|
/// Returns all the "steps" (non-unique and non-similar) types involved in
|
|
/// the conversion sequence. This method does **NOT** return Begin and End.
|
|
SmallVector<QualType, 4> getInvolvedTypesInSequence() const {
|
|
SmallVector<QualType, 4> Ret;
|
|
auto EmplaceIfDifferent = [&Ret](QualType QT) {
|
|
if (QT.isNull())
|
|
return;
|
|
if (Ret.empty())
|
|
Ret.emplace_back(QT);
|
|
else if (Ret.back() != QT)
|
|
Ret.emplace_back(QT);
|
|
};
|
|
|
|
EmplaceIfDifferent(AfterFirstStandard);
|
|
switch (UDConvKind) {
|
|
case UDCK_Ctor:
|
|
EmplaceIfDifferent(UDConvCtor.ConstructorParameterType);
|
|
EmplaceIfDifferent(UDConvCtor.UserDefinedType);
|
|
break;
|
|
case UDCK_Oper:
|
|
EmplaceIfDifferent(UDConvOp.UserDefinedType);
|
|
EmplaceIfDifferent(UDConvOp.ConversionOperatorResultType);
|
|
break;
|
|
case UDCK_None:
|
|
break;
|
|
}
|
|
EmplaceIfDifferent(AfterSecondStandard);
|
|
|
|
return Ret;
|
|
}
|
|
|
|
/// Updates the steps of the conversion sequence with the steps from the
|
|
/// other instance.
|
|
///
|
|
/// \note This method does not check if the resulting conversion sequence is
|
|
/// sensible!
|
|
ConversionSequence &update(const ConversionSequence &RHS) {
|
|
if (!RHS.AfterFirstStandard.isNull())
|
|
AfterFirstStandard = RHS.AfterFirstStandard;
|
|
switch (RHS.UDConvKind) {
|
|
case UDCK_Ctor:
|
|
UDConvKind = UDCK_Ctor;
|
|
UDConvCtor = RHS.UDConvCtor;
|
|
break;
|
|
case UDCK_Oper:
|
|
UDConvKind = UDCK_Oper;
|
|
UDConvOp = RHS.UDConvOp;
|
|
break;
|
|
case UDCK_None:
|
|
break;
|
|
}
|
|
if (!RHS.AfterSecondStandard.isNull())
|
|
AfterSecondStandard = RHS.AfterSecondStandard;
|
|
|
|
return *this;
|
|
}
|
|
|
|
/// Sets the user-defined conversion to the given constructor.
|
|
void setConversion(const UserDefinedConvertingConstructor &UDCC) {
|
|
UDConvKind = UDCK_Ctor;
|
|
UDConvCtor = UDCC;
|
|
}
|
|
|
|
/// Sets the user-defined conversion to the given operator.
|
|
void setConversion(const UserDefinedConversionOperator &UDCO) {
|
|
UDConvKind = UDCK_Oper;
|
|
UDConvOp = UDCO;
|
|
}
|
|
|
|
/// Returns the type in the conversion that's formally "in our hands" once
|
|
/// the user-defined conversion is executed.
|
|
QualType getTypeAfterUserDefinedConversion() const {
|
|
switch (UDConvKind) {
|
|
case UDCK_Ctor:
|
|
return UDConvCtor.UserDefinedType;
|
|
case UDCK_Oper:
|
|
return UDConvOp.ConversionOperatorResultType;
|
|
case UDCK_None:
|
|
return {};
|
|
}
|
|
llvm_unreachable("Invalid UDConv kind.");
|
|
}
|
|
|
|
const CXXMethodDecl *getUserDefinedConversionFunction() const {
|
|
switch (UDConvKind) {
|
|
case UDCK_Ctor:
|
|
return UDConvCtor.Fun;
|
|
case UDCK_Oper:
|
|
return UDConvOp.Fun;
|
|
case UDCK_None:
|
|
return {};
|
|
}
|
|
llvm_unreachable("Invalid UDConv kind.");
|
|
}
|
|
|
|
/// Returns the SourceRange in the text that corresponds to the interesting
|
|
/// part of the user-defined conversion. This is either the parameter type
|
|
/// in a converting constructor, or the conversion result type in a conversion
|
|
/// operator.
|
|
SourceRange getUserDefinedConversionHighlight() const {
|
|
switch (UDConvKind) {
|
|
case UDCK_Ctor:
|
|
return UDConvCtor.Fun->getParamDecl(0)->getSourceRange();
|
|
case UDCK_Oper:
|
|
// getReturnTypeSourceRange() does not work for CXXConversionDecls as the
|
|
// returned type is physically behind the declaration's name ("operator").
|
|
if (const FunctionTypeLoc FTL = UDConvOp.Fun->getFunctionTypeLoc())
|
|
if (const TypeLoc RetLoc = FTL.getReturnLoc())
|
|
return RetLoc.getSourceRange();
|
|
return {};
|
|
case UDCK_None:
|
|
return {};
|
|
}
|
|
llvm_unreachable("Invalid UDConv kind.");
|
|
}
|
|
};
|
|
|
|
/// Contains the metadata for the mixability result between two types,
|
|
/// independently of which parameters they were calculated from.
|
|
struct MixData {
|
|
/// The flag bits of the mix indicating what language features allow for it.
|
|
MixFlags Flags = MixFlags::Invalid;
|
|
|
|
/// A potentially calculated common underlying type after desugaring, that
|
|
/// both sides of the mix can originate from.
|
|
QualType CommonType;
|
|
|
|
/// The steps an implicit conversion performs to get from one type to the
|
|
/// other.
|
|
ConversionSequence Conversion, ConversionRTL;
|
|
|
|
/// True if the MixData was specifically created with only a one-way
|
|
/// conversion modelled.
|
|
bool CreatedFromOneWayConversion = false;
|
|
|
|
MixData(MixFlags Flags) : Flags(Flags) {}
|
|
MixData(MixFlags Flags, QualType CommonType)
|
|
: Flags(Flags), CommonType(CommonType) {}
|
|
MixData(MixFlags Flags, ConversionSequence Conv)
|
|
: Flags(Flags), Conversion(Conv), CreatedFromOneWayConversion(true) {}
|
|
MixData(MixFlags Flags, ConversionSequence LTR, ConversionSequence RTL)
|
|
: Flags(Flags), Conversion(LTR), ConversionRTL(RTL) {}
|
|
MixData(MixFlags Flags, QualType CommonType, ConversionSequence LTR,
|
|
ConversionSequence RTL)
|
|
: Flags(Flags), CommonType(CommonType), Conversion(LTR),
|
|
ConversionRTL(RTL) {}
|
|
|
|
void sanitize() {
|
|
assert(Flags != MixFlags::Invalid && "sanitize() called on invalid bitvec");
|
|
|
|
MixFlags CanonicalAndWorkaround =
|
|
MixFlags::Canonical | MixFlags::WorkaroundDisableCanonicalEquivalence;
|
|
if ((Flags & CanonicalAndWorkaround) == CanonicalAndWorkaround) {
|
|
// A workaround for too eagerly equivalent canonical types was requested,
|
|
// and a canonical equivalence was proven. Fulfill the request and throw
|
|
// this result away.
|
|
Flags = MixFlags::None;
|
|
return;
|
|
}
|
|
|
|
if (hasFlag(Flags, MixFlags::None)) {
|
|
// If anywhere down the recursion a potential mix "path" is deemed
|
|
// impossible, throw away all the other bits because the mix is not
|
|
// possible.
|
|
Flags = MixFlags::None;
|
|
return;
|
|
}
|
|
|
|
if (Flags == MixFlags::Trivial)
|
|
return;
|
|
|
|
if (static_cast<bool>(Flags ^ MixFlags::Trivial))
|
|
// If the mix involves somewhere trivial equivalence but down the
|
|
// recursion other bit(s) were set, remove the trivial bit, as it is not
|
|
// trivial.
|
|
Flags &= ~MixFlags::Trivial;
|
|
|
|
bool ShouldHaveImplicitConvFlag = false;
|
|
if (CreatedFromOneWayConversion && Conversion)
|
|
ShouldHaveImplicitConvFlag = true;
|
|
else if (!CreatedFromOneWayConversion && Conversion && ConversionRTL)
|
|
// Only say that we have implicit conversion mix possibility if it is
|
|
// bidirectional. Otherwise, the compiler would report an *actual* swap
|
|
// at a call site...
|
|
ShouldHaveImplicitConvFlag = true;
|
|
|
|
if (ShouldHaveImplicitConvFlag)
|
|
Flags |= MixFlags::ImplicitConversion;
|
|
else
|
|
Flags &= ~MixFlags::ImplicitConversion;
|
|
}
|
|
|
|
bool isValid() const { return Flags >= MixFlags::None; }
|
|
|
|
bool indicatesMixability() const { return Flags > MixFlags::None; }
|
|
|
|
/// Add the specified flag bits to the flags.
|
|
MixData operator|(MixFlags EnableFlags) const {
|
|
if (CreatedFromOneWayConversion) {
|
|
MixData M{Flags | EnableFlags, Conversion};
|
|
M.CommonType = CommonType;
|
|
return M;
|
|
}
|
|
return {Flags | EnableFlags, CommonType, Conversion, ConversionRTL};
|
|
}
|
|
|
|
/// Add the specified flag bits to the flags.
|
|
MixData &operator|=(MixFlags EnableFlags) {
|
|
Flags |= EnableFlags;
|
|
return *this;
|
|
}
|
|
|
|
template <typename F> MixData withCommonTypeTransformed(const F &Func) const {
|
|
if (CommonType.isNull())
|
|
return *this;
|
|
|
|
QualType NewCommonType = Func(CommonType);
|
|
|
|
if (CreatedFromOneWayConversion) {
|
|
MixData M{Flags, Conversion};
|
|
M.CommonType = NewCommonType;
|
|
return M;
|
|
}
|
|
|
|
return {Flags, NewCommonType, Conversion, ConversionRTL};
|
|
}
|
|
};
|
|
|
|
/// A named tuple that contains the information for a mix between two concrete
|
|
/// parameters.
|
|
struct Mix {
|
|
const ParmVarDecl *First, *Second;
|
|
MixData Data;
|
|
|
|
Mix(const ParmVarDecl *F, const ParmVarDecl *S, MixData Data)
|
|
: First(F), Second(S), Data(std::move(Data)) {}
|
|
|
|
void sanitize() { Data.sanitize(); }
|
|
MixFlags flags() const { return Data.Flags; }
|
|
bool flagsValid() const { return Data.isValid(); }
|
|
bool mixable() const { return Data.indicatesMixability(); }
|
|
QualType commonUnderlyingType() const { return Data.CommonType; }
|
|
const ConversionSequence &leftToRightConversionSequence() const {
|
|
return Data.Conversion;
|
|
}
|
|
const ConversionSequence &rightToLeftConversionSequence() const {
|
|
return Data.ConversionRTL;
|
|
}
|
|
};
|
|
|
|
// NOLINTNEXTLINE(misc-redundant-expression): Seems to be a bogus warning.
|
|
static_assert(std::is_trivially_copyable_v<Mix> &&
|
|
std::is_trivially_move_constructible_v<Mix> &&
|
|
std::is_trivially_move_assignable_v<Mix>,
|
|
"Keep frequently used data simple!");
|
|
|
|
struct MixableParameterRange {
|
|
/// A container for Mixes.
|
|
using MixVector = SmallVector<Mix, 8>;
|
|
|
|
/// The number of parameters iterated to build the instance.
|
|
std::size_t NumParamsChecked = 0;
|
|
|
|
/// The individual flags and supporting information for the mixes.
|
|
MixVector Mixes;
|
|
|
|
/// Gets the leftmost parameter of the range.
|
|
const ParmVarDecl *getFirstParam() const {
|
|
// The first element is the LHS of the very first mix in the range.
|
|
assert(!Mixes.empty());
|
|
return Mixes.front().First;
|
|
}
|
|
|
|
/// Gets the rightmost parameter of the range.
|
|
const ParmVarDecl *getLastParam() const {
|
|
// The builder function breaks building an instance of this type if it
|
|
// finds something that can not be mixed with the rest, by going *forward*
|
|
// in the list of parameters. So at any moment of break, the RHS of the last
|
|
// element of the mix vector is also the last element of the mixing range.
|
|
assert(!Mixes.empty());
|
|
return Mixes.back().Second;
|
|
}
|
|
};
|
|
|
|
/// Helper enum for the recursive calls in the modelling that toggle what kinds
|
|
/// of implicit conversions are to be modelled.
|
|
enum class ImplicitConversionModellingMode : unsigned char {
|
|
///< No implicit conversions are modelled.
|
|
None,
|
|
|
|
///< The full implicit conversion sequence is modelled.
|
|
All,
|
|
|
|
///< Only model a unidirectional implicit conversion and within it only one
|
|
/// standard conversion sequence.
|
|
OneWaySingleStandardOnly
|
|
};
|
|
|
|
static MixData
|
|
isLRefEquallyBindingToType(const TheCheck &Check,
|
|
const LValueReferenceType *LRef, QualType Ty,
|
|
const ASTContext &Ctx, bool IsRefRHS,
|
|
ImplicitConversionModellingMode ImplicitMode);
|
|
|
|
static MixData
|
|
approximateImplicitConversion(const TheCheck &Check, QualType LType,
|
|
QualType RType, const ASTContext &Ctx,
|
|
ImplicitConversionModellingMode ImplicitMode);
|
|
|
|
static inline bool isUselessSugar(const Type *T) {
|
|
return isa<AttributedType, DecayedType, ParenType>(T);
|
|
}
|
|
|
|
namespace {
|
|
|
|
struct NonCVRQualifiersResult {
|
|
/// True if the types are qualified in a way that even after equating or
|
|
/// removing local CVR qualification, even if the unqualified types
|
|
/// themselves would mix, the qualified ones don't, because there are some
|
|
/// other local qualifiers that are not equal.
|
|
bool HasMixabilityBreakingQualifiers;
|
|
|
|
/// The set of equal qualifiers between the two types.
|
|
Qualifiers CommonQualifiers;
|
|
};
|
|
|
|
} // namespace
|
|
|
|
/// Returns if the two types are qualified in a way that ever after equating or
|
|
/// removing local CVR qualification, even if the unqualified types would mix,
|
|
/// the qualified ones don't, because there are some other local qualifiers
|
|
/// that aren't equal.
|
|
static NonCVRQualifiersResult
|
|
getNonCVRQualifiers(const ASTContext &Ctx, QualType LType, QualType RType) {
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> getNonCVRQualifiers for LType:\n";
|
|
LType.dump(llvm::dbgs(), Ctx); llvm::dbgs() << "\nand RType:\n";
|
|
RType.dump(llvm::dbgs(), Ctx); llvm::dbgs() << '\n';);
|
|
Qualifiers LQual = LType.getLocalQualifiers(),
|
|
RQual = RType.getLocalQualifiers();
|
|
|
|
// Strip potential CVR. That is handled by the check option QualifiersMix.
|
|
LQual.removeCVRQualifiers();
|
|
RQual.removeCVRQualifiers();
|
|
|
|
NonCVRQualifiersResult Ret;
|
|
Ret.CommonQualifiers = Qualifiers::removeCommonQualifiers(LQual, RQual);
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "--- hasNonCVRMixabilityBreakingQualifiers. "
|
|
"Removed common qualifiers: ";
|
|
Ret.CommonQualifiers.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << "\n\tremaining on LType: ";
|
|
LQual.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << "\n\tremaining on RType: ";
|
|
RQual.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << '\n';);
|
|
|
|
// If there are no other non-cvr non-common qualifiers left, we can deduce
|
|
// that mixability isn't broken.
|
|
Ret.HasMixabilityBreakingQualifiers =
|
|
LQual.hasQualifiers() || RQual.hasQualifiers();
|
|
|
|
return Ret;
|
|
}
|
|
|
|
/// Approximate the way how LType and RType might refer to "essentially the
|
|
/// same" type, in a sense that at a particular call site, an expression of
|
|
/// type LType and RType might be successfully passed to a variable (in our
|
|
/// specific case, a parameter) of type RType and LType, respectively.
|
|
/// Note the swapped order!
|
|
///
|
|
/// The returned data structure is not guaranteed to be properly set, as this
|
|
/// function is potentially recursive. It is the caller's responsibility to
|
|
/// call sanitize() on the result once the recursion is over.
|
|
static MixData
|
|
calculateMixability(const TheCheck &Check, QualType LType, QualType RType,
|
|
const ASTContext &Ctx,
|
|
ImplicitConversionModellingMode ImplicitMode) {
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> calculateMixability for LType:\n";
|
|
LType.dump(llvm::dbgs(), Ctx); llvm::dbgs() << "\nand RType:\n";
|
|
RType.dump(llvm::dbgs(), Ctx); llvm::dbgs() << '\n';);
|
|
if (LType == RType) {
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< calculateMixability. Trivial equality.\n");
|
|
return {MixFlags::Trivial, LType};
|
|
}
|
|
|
|
// Dissolve certain type sugars that do not affect the mixability of one type
|
|
// with the other, and also do not require any sort of elaboration for the
|
|
// user to understand.
|
|
if (isUselessSugar(LType.getTypePtr())) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- calculateMixability. LHS is useless sugar.\n");
|
|
return calculateMixability(Check, LType.getSingleStepDesugaredType(Ctx),
|
|
RType, Ctx, ImplicitMode);
|
|
}
|
|
if (isUselessSugar(RType.getTypePtr())) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- calculateMixability. RHS is useless sugar.\n");
|
|
return calculateMixability(
|
|
Check, LType, RType.getSingleStepDesugaredType(Ctx), Ctx, ImplicitMode);
|
|
}
|
|
|
|
const auto *LLRef = LType->getAs<LValueReferenceType>();
|
|
const auto *RLRef = RType->getAs<LValueReferenceType>();
|
|
if (LLRef && RLRef) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. LHS and RHS are &.\n");
|
|
|
|
return calculateMixability(Check, LLRef->getPointeeType(),
|
|
RLRef->getPointeeType(), Ctx, ImplicitMode)
|
|
.withCommonTypeTransformed(
|
|
[&Ctx](QualType QT) { return Ctx.getLValueReferenceType(QT); });
|
|
}
|
|
// At a particular call site, what could be passed to a 'T' or 'const T' might
|
|
// also be passed to a 'const T &' without the call site putting a direct
|
|
// side effect on the passed expressions.
|
|
if (LLRef) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. LHS is &.\n");
|
|
return isLRefEquallyBindingToType(Check, LLRef, RType, Ctx, false,
|
|
ImplicitMode) |
|
|
MixFlags::ReferenceBind;
|
|
}
|
|
if (RLRef) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. RHS is &.\n");
|
|
return isLRefEquallyBindingToType(Check, RLRef, LType, Ctx, true,
|
|
ImplicitMode) |
|
|
MixFlags::ReferenceBind;
|
|
}
|
|
|
|
if (LType->getAs<TypedefType>()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. LHS is typedef.\n");
|
|
return calculateMixability(Check, LType.getSingleStepDesugaredType(Ctx),
|
|
RType, Ctx, ImplicitMode) |
|
|
MixFlags::TypeAlias;
|
|
}
|
|
if (RType->getAs<TypedefType>()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. RHS is typedef.\n");
|
|
return calculateMixability(Check, LType,
|
|
RType.getSingleStepDesugaredType(Ctx), Ctx,
|
|
ImplicitMode) |
|
|
MixFlags::TypeAlias;
|
|
}
|
|
|
|
// A parameter of type 'cvr1 T' and another of potentially differently
|
|
// qualified 'cvr2 T' may bind with the same power, if the user so requested.
|
|
//
|
|
// Whether to do this check for the inner unqualified types.
|
|
bool CompareUnqualifiedTypes = false;
|
|
if (LType.getLocalCVRQualifiers() != RType.getLocalCVRQualifiers()) {
|
|
LLVM_DEBUG(if (LType.getLocalCVRQualifiers()) {
|
|
llvm::dbgs() << "--- calculateMixability. LHS has CVR-Qualifiers: ";
|
|
Qualifiers::fromCVRMask(LType.getLocalCVRQualifiers())
|
|
.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << '\n';
|
|
});
|
|
LLVM_DEBUG(if (RType.getLocalCVRQualifiers()) {
|
|
llvm::dbgs() << "--- calculateMixability. RHS has CVR-Qualifiers: ";
|
|
Qualifiers::fromCVRMask(RType.getLocalCVRQualifiers())
|
|
.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << '\n';
|
|
});
|
|
|
|
if (!Check.QualifiersMix) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< calculateMixability. QualifiersMix turned off - not "
|
|
"mixable.\n");
|
|
return {MixFlags::None};
|
|
}
|
|
|
|
CompareUnqualifiedTypes = true;
|
|
}
|
|
// Whether the two types had the same CVR qualifiers.
|
|
bool OriginallySameQualifiers = false;
|
|
if (LType.getLocalCVRQualifiers() == RType.getLocalCVRQualifiers() &&
|
|
LType.getLocalCVRQualifiers() != 0) {
|
|
LLVM_DEBUG(if (LType.getLocalCVRQualifiers()) {
|
|
llvm::dbgs()
|
|
<< "--- calculateMixability. LHS and RHS have same CVR-Qualifiers: ";
|
|
Qualifiers::fromCVRMask(LType.getLocalCVRQualifiers())
|
|
.print(llvm::dbgs(), Ctx.getPrintingPolicy());
|
|
llvm::dbgs() << '\n';
|
|
});
|
|
|
|
CompareUnqualifiedTypes = true;
|
|
OriginallySameQualifiers = true;
|
|
}
|
|
|
|
if (CompareUnqualifiedTypes) {
|
|
NonCVRQualifiersResult AdditionalQuals =
|
|
getNonCVRQualifiers(Ctx, LType, RType);
|
|
if (AdditionalQuals.HasMixabilityBreakingQualifiers) {
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< calculateMixability. Additional "
|
|
"non-equal incompatible qualifiers.\n");
|
|
return {MixFlags::None};
|
|
}
|
|
|
|
MixData UnqualifiedMixability =
|
|
calculateMixability(Check, LType.getLocalUnqualifiedType(),
|
|
RType.getLocalUnqualifiedType(), Ctx, ImplicitMode)
|
|
.withCommonTypeTransformed([&AdditionalQuals, &Ctx](QualType QT) {
|
|
// Once the mixability was deduced, apply the qualifiers common
|
|
// to the two type back onto the diagnostic printout.
|
|
return Ctx.getQualifiedType(QT, AdditionalQuals.CommonQualifiers);
|
|
});
|
|
|
|
if (!OriginallySameQualifiers)
|
|
// User-enabled qualifier change modelled for the mix.
|
|
return UnqualifiedMixability | MixFlags::Qualifiers;
|
|
|
|
// Apply the same qualifier back into the found common type if they were
|
|
// the same.
|
|
return UnqualifiedMixability.withCommonTypeTransformed(
|
|
[&Ctx, LType](QualType QT) {
|
|
return Ctx.getQualifiedType(QT, LType.getLocalQualifiers());
|
|
});
|
|
}
|
|
|
|
// Certain constructs match on the last catch-all getCanonicalType() equality,
|
|
// which is perhaps something not what we want. If this variable is true,
|
|
// the canonical type equality will be ignored.
|
|
bool RecursiveReturnDiscardingCanonicalType = false;
|
|
|
|
if (LType->isPointerType() && RType->isPointerType()) {
|
|
// If both types are pointers, and pointed to the exact same type,
|
|
// LType == RType took care of that. Try to see if the pointee type has
|
|
// some other match. However, this must not consider implicit conversions.
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- calculateMixability. LHS and RHS are Ptrs.\n");
|
|
MixData MixOfPointee =
|
|
calculateMixability(Check, LType->getPointeeType(),
|
|
RType->getPointeeType(), Ctx,
|
|
ImplicitConversionModellingMode::None)
|
|
.withCommonTypeTransformed(
|
|
[&Ctx](QualType QT) { return Ctx.getPointerType(QT); });
|
|
if (hasFlag(MixOfPointee.Flags,
|
|
MixFlags::WorkaroundDisableCanonicalEquivalence))
|
|
RecursiveReturnDiscardingCanonicalType = true;
|
|
|
|
MixOfPointee.sanitize();
|
|
if (MixOfPointee.indicatesMixability()) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< calculateMixability. Pointees are mixable.\n");
|
|
return MixOfPointee;
|
|
}
|
|
}
|
|
|
|
if (ImplicitMode > ImplicitConversionModellingMode::None) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. Start implicit...\n");
|
|
MixData MixLTR =
|
|
approximateImplicitConversion(Check, LType, RType, Ctx, ImplicitMode);
|
|
LLVM_DEBUG(
|
|
if (hasFlag(MixLTR.Flags, MixFlags::ImplicitConversion)) llvm::dbgs()
|
|
<< "--- calculateMixability. Implicit Left -> Right found.\n";);
|
|
|
|
if (ImplicitMode ==
|
|
ImplicitConversionModellingMode::OneWaySingleStandardOnly &&
|
|
MixLTR.Conversion && !MixLTR.Conversion.AfterFirstStandard.isNull() &&
|
|
MixLTR.Conversion.UDConvKind == ConversionSequence::UDCK_None &&
|
|
MixLTR.Conversion.AfterSecondStandard.isNull()) {
|
|
// The invoker of the method requested only modelling a single standard
|
|
// conversion, in only the forward direction, and they got just that.
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< calculateMixability. Implicit "
|
|
"conversion, one-way, standard-only.\n");
|
|
return {MixFlags::ImplicitConversion, MixLTR.Conversion};
|
|
}
|
|
|
|
// Otherwise if the invoker requested a full modelling, do the other
|
|
// direction as well.
|
|
MixData MixRTL =
|
|
approximateImplicitConversion(Check, RType, LType, Ctx, ImplicitMode);
|
|
LLVM_DEBUG(
|
|
if (hasFlag(MixRTL.Flags, MixFlags::ImplicitConversion)) llvm::dbgs()
|
|
<< "--- calculateMixability. Implicit Right -> Left found.\n";);
|
|
|
|
if (MixLTR.Conversion && MixRTL.Conversion) {
|
|
LLVM_DEBUG(
|
|
llvm::dbgs()
|
|
<< "<<< calculateMixability. Implicit conversion, bidirectional.\n");
|
|
return {MixFlags::ImplicitConversion, MixLTR.Conversion,
|
|
MixRTL.Conversion};
|
|
}
|
|
}
|
|
|
|
if (RecursiveReturnDiscardingCanonicalType)
|
|
LLVM_DEBUG(llvm::dbgs() << "--- calculateMixability. Before CanonicalType, "
|
|
"Discard was enabled.\n");
|
|
|
|
// Certain kinds unfortunately need to be side-stepped for canonical type
|
|
// matching.
|
|
if (LType->getAs<FunctionProtoType>() || RType->getAs<FunctionProtoType>()) {
|
|
// Unfortunately, the canonical type of a function pointer becomes the
|
|
// same even if exactly one is "noexcept" and the other isn't, making us
|
|
// give a false positive report irrespective of implicit conversions.
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- calculateMixability. Discarding potential canonical "
|
|
"equivalence on FunctionProtoTypes.\n");
|
|
RecursiveReturnDiscardingCanonicalType = true;
|
|
}
|
|
|
|
MixData MixToReturn{MixFlags::None};
|
|
|
|
// If none of the previous logic found a match, try if Clang otherwise
|
|
// believes the types to be the same.
|
|
QualType LCanonical = LType.getCanonicalType();
|
|
if (LCanonical == RType.getCanonicalType()) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< calculateMixability. Same CanonicalType.\n");
|
|
MixToReturn = {MixFlags::Canonical, LCanonical};
|
|
}
|
|
|
|
if (RecursiveReturnDiscardingCanonicalType)
|
|
MixToReturn |= MixFlags::WorkaroundDisableCanonicalEquivalence;
|
|
|
|
LLVM_DEBUG(if (MixToReturn.Flags == MixFlags::None) llvm::dbgs()
|
|
<< "<<< calculateMixability. No match found.\n");
|
|
return MixToReturn;
|
|
}
|
|
|
|
/// Calculates if the reference binds an expression of the given type. This is
|
|
/// true iff 'LRef' is some 'const T &' type, and the 'Ty' is 'T' or 'const T'.
|
|
///
|
|
/// \param ImplicitMode is forwarded in the possible recursive call to
|
|
/// calculateMixability.
|
|
static MixData
|
|
isLRefEquallyBindingToType(const TheCheck &Check,
|
|
const LValueReferenceType *LRef, QualType Ty,
|
|
const ASTContext &Ctx, bool IsRefRHS,
|
|
ImplicitConversionModellingMode ImplicitMode) {
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> isLRefEquallyBindingToType for LRef:\n";
|
|
LRef->dump(llvm::dbgs(), Ctx); llvm::dbgs() << "\nand Type:\n";
|
|
Ty.dump(llvm::dbgs(), Ctx); llvm::dbgs() << '\n';);
|
|
|
|
QualType ReferredType = LRef->getPointeeType();
|
|
if (!ReferredType.isLocalConstQualified() &&
|
|
ReferredType->getAs<TypedefType>()) {
|
|
LLVM_DEBUG(
|
|
llvm::dbgs()
|
|
<< "--- isLRefEquallyBindingToType. Non-const LRef to Typedef.\n");
|
|
ReferredType = ReferredType.getDesugaredType(Ctx);
|
|
if (!ReferredType.isLocalConstQualified()) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< isLRefEquallyBindingToType. Typedef is not const.\n");
|
|
return {MixFlags::None};
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "--- isLRefEquallyBindingToType. Typedef is "
|
|
"const, considering as const LRef.\n");
|
|
} else if (!ReferredType.isLocalConstQualified()) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< isLRefEquallyBindingToType. Not const LRef.\n");
|
|
return {MixFlags::None};
|
|
};
|
|
|
|
assert(ReferredType.isLocalConstQualified() &&
|
|
"Reaching this point means we are sure LRef is effectively a const&.");
|
|
|
|
if (ReferredType == Ty) {
|
|
LLVM_DEBUG(
|
|
llvm::dbgs()
|
|
<< "<<< isLRefEquallyBindingToType. Type of referred matches.\n");
|
|
return {MixFlags::Trivial, ReferredType};
|
|
}
|
|
|
|
QualType NonConstReferredType = ReferredType;
|
|
NonConstReferredType.removeLocalConst();
|
|
if (NonConstReferredType == Ty) {
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< isLRefEquallyBindingToType. Type of "
|
|
"referred matches to non-const qualified.\n");
|
|
return {MixFlags::Trivial, NonConstReferredType};
|
|
}
|
|
|
|
LLVM_DEBUG(
|
|
llvm::dbgs()
|
|
<< "--- isLRefEquallyBindingToType. Checking mix for underlying type.\n");
|
|
return IsRefRHS ? calculateMixability(Check, Ty, NonConstReferredType, Ctx,
|
|
ImplicitMode)
|
|
: calculateMixability(Check, NonConstReferredType, Ty, Ctx,
|
|
ImplicitMode);
|
|
}
|
|
|
|
static inline bool isDerivedToBase(const CXXRecordDecl *Derived,
|
|
const CXXRecordDecl *Base) {
|
|
return Derived && Base && Derived->isCompleteDefinition() &&
|
|
Base->isCompleteDefinition() && Derived->isDerivedFrom(Base);
|
|
}
|
|
|
|
static std::optional<QualType>
|
|
approximateStandardConversionSequence(const TheCheck &Check, QualType From,
|
|
QualType To, const ASTContext &Ctx) {
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> approximateStdConv for LType:\n";
|
|
From.dump(llvm::dbgs(), Ctx); llvm::dbgs() << "\nand RType:\n";
|
|
To.dump(llvm::dbgs(), Ctx); llvm::dbgs() << '\n';);
|
|
|
|
// A standard conversion sequence consists of the following, in order:
|
|
// * Maybe either LValue->RValue conv., Array->Ptr conv., Function->Ptr conv.
|
|
// * Maybe Numeric promotion or conversion.
|
|
// * Maybe function pointer conversion.
|
|
// * Maybe qualifier adjustments.
|
|
QualType WorkType = From;
|
|
// Get out the qualifiers of the original type. This will always be
|
|
// re-applied to the WorkType to ensure it is the same qualification as the
|
|
// original From was.
|
|
auto FastQualifiersToApply = static_cast<unsigned>(
|
|
From.split().Quals.getAsOpaqueValue() & Qualifiers::FastMask);
|
|
|
|
// LValue->RValue is irrelevant for the check, because it is a thing to be
|
|
// done at a call site, and will be performed if need be performed.
|
|
|
|
// Array->Pointer decay is handled by the main method in desugaring
|
|
// the parameter's DecayedType as "useless sugar".
|
|
|
|
// Function->Pointer conversions are also irrelevant, because a
|
|
// "FunctionType" cannot be the type of a parameter variable, so this
|
|
// conversion is only meaningful at call sites.
|
|
|
|
// Numeric promotions and conversions.
|
|
const auto *FromBuiltin = WorkType->getAs<BuiltinType>();
|
|
const auto *ToBuiltin = To->getAs<BuiltinType>();
|
|
bool FromNumeric = FromBuiltin && (FromBuiltin->isIntegerType() ||
|
|
FromBuiltin->isFloatingType());
|
|
bool ToNumeric =
|
|
ToBuiltin && (ToBuiltin->isIntegerType() || ToBuiltin->isFloatingType());
|
|
if (FromNumeric && ToNumeric) {
|
|
// If both are integral types, the numeric conversion is performed.
|
|
// Reapply the qualifiers of the original type, however, so
|
|
// "const int -> double" in this case moves over to
|
|
// "const double -> double".
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- approximateStdConv. Conversion between numerics.\n");
|
|
WorkType = QualType{ToBuiltin, FastQualifiersToApply};
|
|
}
|
|
|
|
const auto *FromEnum = WorkType->getAs<EnumType>();
|
|
const auto *ToEnum = To->getAs<EnumType>();
|
|
if (FromEnum && ToNumeric && FromEnum->isUnscopedEnumerationType()) {
|
|
// Unscoped enumerations (or enumerations in C) convert to numerics.
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- approximateStdConv. Unscoped enum to numeric.\n");
|
|
WorkType = QualType{ToBuiltin, FastQualifiersToApply};
|
|
} else if (FromNumeric && ToEnum && ToEnum->isUnscopedEnumerationType()) {
|
|
// Numeric types convert to enumerations only in C.
|
|
if (Ctx.getLangOpts().CPlusPlus) {
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< approximateStdConv. Numeric to unscoped "
|
|
"enum, not possible in C++!\n");
|
|
return {};
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- approximateStdConv. Numeric to unscoped enum.\n");
|
|
WorkType = QualType{ToEnum, FastQualifiersToApply};
|
|
}
|
|
|
|
// Check for pointer conversions.
|
|
const auto *FromPtr = WorkType->getAs<PointerType>();
|
|
const auto *ToPtr = To->getAs<PointerType>();
|
|
if (FromPtr && ToPtr) {
|
|
if (ToPtr->isVoidPointerType()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateStdConv. To void pointer.\n");
|
|
WorkType = QualType{ToPtr, FastQualifiersToApply};
|
|
}
|
|
|
|
const auto *FromRecordPtr = FromPtr->getPointeeCXXRecordDecl();
|
|
const auto *ToRecordPtr = ToPtr->getPointeeCXXRecordDecl();
|
|
if (isDerivedToBase(FromRecordPtr, ToRecordPtr)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateStdConv. Derived* to Base*\n");
|
|
WorkType = QualType{ToPtr, FastQualifiersToApply};
|
|
}
|
|
}
|
|
|
|
// Model the slicing Derived-to-Base too, as "BaseT temporary = derived;"
|
|
// can also be compiled.
|
|
const auto *FromRecord = WorkType->getAsCXXRecordDecl();
|
|
const auto *ToRecord = To->getAsCXXRecordDecl();
|
|
if (isDerivedToBase(FromRecord, ToRecord)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateStdConv. Derived To Base.\n");
|
|
WorkType = QualType{
|
|
ToRecord->getASTContext().getCanonicalTagType(ToRecord)->getTypePtr(),
|
|
FastQualifiersToApply};
|
|
}
|
|
|
|
if (Ctx.getLangOpts().CPlusPlus17 && FromPtr && ToPtr) {
|
|
// Function pointer conversion: A noexcept function pointer can be passed
|
|
// to a non-noexcept one.
|
|
const auto *FromFunctionPtr =
|
|
FromPtr->getPointeeType()->getAs<FunctionProtoType>();
|
|
const auto *ToFunctionPtr =
|
|
ToPtr->getPointeeType()->getAs<FunctionProtoType>();
|
|
if (FromFunctionPtr && ToFunctionPtr &&
|
|
FromFunctionPtr->hasNoexceptExceptionSpec() &&
|
|
!ToFunctionPtr->hasNoexceptExceptionSpec()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateStdConv. noexcept function "
|
|
"pointer to non-noexcept.\n");
|
|
WorkType = QualType{ToPtr, FastQualifiersToApply};
|
|
}
|
|
}
|
|
|
|
// Qualifier adjustments are modelled according to the user's request in
|
|
// the QualifiersMix check config.
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- approximateStdConv. Trying qualifier adjustment...\n");
|
|
MixData QualConv = calculateMixability(Check, WorkType, To, Ctx,
|
|
ImplicitConversionModellingMode::None);
|
|
QualConv.sanitize();
|
|
if (hasFlag(QualConv.Flags, MixFlags::Qualifiers)) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< approximateStdConv. Qualifiers adjusted.\n");
|
|
WorkType = To;
|
|
}
|
|
|
|
if (Ctx.hasSameType(WorkType, To)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< approximateStdConv. Reached 'To' type.\n");
|
|
return {Ctx.getCommonSugaredType(WorkType, To)};
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< approximateStdConv. Did not reach 'To'.\n");
|
|
return {};
|
|
}
|
|
|
|
namespace {
|
|
|
|
/// Helper class for storing possible user-defined conversion calls that
|
|
/// *could* take place in an implicit conversion, and selecting the one that
|
|
/// most likely *does*, if any.
|
|
class UserDefinedConversionSelector {
|
|
public:
|
|
/// The conversion associated with a conversion function, together with the
|
|
/// mixability flags of the conversion function's parameter or return type
|
|
/// to the rest of the sequence the selector is used in, and the sequence
|
|
/// that applied through the conversion itself.
|
|
struct PreparedConversion {
|
|
const CXXMethodDecl *ConversionFun;
|
|
MixFlags Flags;
|
|
ConversionSequence Seq;
|
|
|
|
PreparedConversion(const CXXMethodDecl *CMD, MixFlags F,
|
|
ConversionSequence S)
|
|
: ConversionFun(CMD), Flags(F), Seq(S) {}
|
|
};
|
|
|
|
UserDefinedConversionSelector(const TheCheck &Check) : Check(Check) {}
|
|
|
|
/// Adds the conversion between the two types for the given function into
|
|
/// the possible implicit conversion set. FromType and ToType is either:
|
|
/// * the result of a standard sequence and a converting ctor parameter
|
|
/// * the return type of a conversion operator and the expected target of
|
|
/// an implicit conversion.
|
|
void addConversion(const CXXMethodDecl *ConvFun, QualType FromType,
|
|
QualType ToType) {
|
|
// Try to go from the FromType to the ToType with only a single implicit
|
|
// conversion, to see if the conversion function is applicable.
|
|
MixData Mix = calculateMixability(
|
|
Check, FromType, ToType, ConvFun->getASTContext(),
|
|
ImplicitConversionModellingMode::OneWaySingleStandardOnly);
|
|
Mix.sanitize();
|
|
if (!Mix.indicatesMixability())
|
|
return;
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "--- tryConversion. Found viable with flags: "
|
|
<< formatMixFlags(Mix.Flags) << '\n');
|
|
FlaggedConversions.emplace_back(ConvFun, Mix.Flags, Mix.Conversion);
|
|
}
|
|
|
|
/// Selects the best conversion function that is applicable from the
|
|
/// prepared set of potential conversion functions taken.
|
|
std::optional<PreparedConversion> operator()() const {
|
|
if (FlaggedConversions.empty()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- selectUserDefinedConv. Empty.\n");
|
|
return {};
|
|
}
|
|
if (FlaggedConversions.size() == 1) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- selectUserDefinedConv. Single.\n");
|
|
return FlaggedConversions.front();
|
|
}
|
|
|
|
std::optional<PreparedConversion> BestConversion;
|
|
unsigned short HowManyGoodConversions = 0;
|
|
for (const auto &Prepared : FlaggedConversions) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- selectUserDefinedConv. Candidate flags: "
|
|
<< formatMixFlags(Prepared.Flags) << '\n');
|
|
if (!BestConversion) {
|
|
BestConversion = Prepared;
|
|
++HowManyGoodConversions;
|
|
continue;
|
|
}
|
|
|
|
bool BestConversionHasImplicit =
|
|
hasFlag(BestConversion->Flags, MixFlags::ImplicitConversion);
|
|
bool ThisConversionHasImplicit =
|
|
hasFlag(Prepared.Flags, MixFlags::ImplicitConversion);
|
|
if (!BestConversionHasImplicit && ThisConversionHasImplicit)
|
|
// This is a worse conversion, because a better one was found earlier.
|
|
continue;
|
|
|
|
if (BestConversionHasImplicit && !ThisConversionHasImplicit) {
|
|
// If the so far best selected conversion needs a previous implicit
|
|
// conversion to match the user-defined converting function, but this
|
|
// conversion does not, this is a better conversion, and we can throw
|
|
// away the previously selected conversion(s).
|
|
BestConversion = Prepared;
|
|
HowManyGoodConversions = 1;
|
|
continue;
|
|
}
|
|
|
|
if (BestConversionHasImplicit == ThisConversionHasImplicit)
|
|
// The current conversion is the same in term of goodness than the
|
|
// already selected one.
|
|
++HowManyGoodConversions;
|
|
}
|
|
|
|
if (HowManyGoodConversions == 1) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- selectUserDefinedConv. Unique result. Flags: "
|
|
<< formatMixFlags(BestConversion->Flags) << '\n');
|
|
return BestConversion;
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "--- selectUserDefinedConv. No, or ambiguous.\n");
|
|
return {};
|
|
}
|
|
|
|
private:
|
|
llvm::SmallVector<PreparedConversion, 2> FlaggedConversions;
|
|
const TheCheck &Check;
|
|
};
|
|
|
|
} // namespace
|
|
|
|
static std::optional<ConversionSequence>
|
|
tryConversionOperators(const TheCheck &Check, const CXXRecordDecl *RD,
|
|
QualType ToType) {
|
|
if (!RD || !RD->isCompleteDefinition())
|
|
return {};
|
|
RD = RD->getDefinition();
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> tryConversionOperators: " << RD->getName()
|
|
<< " to:\n";
|
|
ToType.dump(llvm::dbgs(), RD->getASTContext());
|
|
llvm::dbgs() << '\n';);
|
|
|
|
UserDefinedConversionSelector ConversionSet{Check};
|
|
|
|
for (const NamedDecl *Method : RD->getVisibleConversionFunctions()) {
|
|
const auto *Con = dyn_cast<CXXConversionDecl>(Method);
|
|
if (!Con || Con->isExplicit())
|
|
continue;
|
|
LLVM_DEBUG(llvm::dbgs() << "--- tryConversionOperators. Trying:\n";
|
|
Con->dump(llvm::dbgs()); llvm::dbgs() << '\n';);
|
|
|
|
// Try to go from the result of conversion operator to the expected type,
|
|
// without calculating another user-defined conversion.
|
|
ConversionSet.addConversion(Con, Con->getConversionType(), ToType);
|
|
}
|
|
|
|
if (std::optional<UserDefinedConversionSelector::PreparedConversion>
|
|
SelectedConversion = ConversionSet()) {
|
|
CanQualType RecordType = RD->getASTContext().getCanonicalTagType(RD);
|
|
|
|
ConversionSequence Result{RecordType, ToType};
|
|
// The conversion from the operator call's return type to ToType was
|
|
// modelled as a "pre-conversion" in the operator call, but it is the
|
|
// "post-conversion" from the point of view of the original conversion
|
|
// we are modelling.
|
|
Result.AfterSecondStandard = SelectedConversion->Seq.AfterFirstStandard;
|
|
|
|
ConversionSequence::UserDefinedConversionOperator ConvOp;
|
|
ConvOp.Fun = cast<CXXConversionDecl>(SelectedConversion->ConversionFun);
|
|
ConvOp.UserDefinedType = RecordType;
|
|
ConvOp.ConversionOperatorResultType = ConvOp.Fun->getConversionType();
|
|
Result.setConversion(ConvOp);
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< tryConversionOperators. Found result.\n");
|
|
return Result;
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< tryConversionOperators. No conversion.\n");
|
|
return {};
|
|
}
|
|
|
|
static std::optional<ConversionSequence>
|
|
tryConvertingConstructors(const TheCheck &Check, QualType FromType,
|
|
const CXXRecordDecl *RD) {
|
|
if (!RD || !RD->isCompleteDefinition())
|
|
return {};
|
|
RD = RD->getDefinition();
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> tryConveringConstructors: " << RD->getName()
|
|
<< " from:\n";
|
|
FromType.dump(llvm::dbgs(), RD->getASTContext());
|
|
llvm::dbgs() << '\n';);
|
|
|
|
UserDefinedConversionSelector ConversionSet{Check};
|
|
|
|
for (const CXXConstructorDecl *Con : RD->ctors()) {
|
|
if (Con->isCopyOrMoveConstructor() ||
|
|
!Con->isConvertingConstructor(/* AllowExplicit =*/false))
|
|
continue;
|
|
LLVM_DEBUG(llvm::dbgs() << "--- tryConvertingConstructors. Trying:\n";
|
|
Con->dump(llvm::dbgs()); llvm::dbgs() << '\n';);
|
|
|
|
// Try to go from the original FromType to the converting constructor's
|
|
// parameter type without another user-defined conversion.
|
|
ConversionSet.addConversion(Con, FromType, Con->getParamDecl(0)->getType());
|
|
}
|
|
|
|
if (std::optional<UserDefinedConversionSelector::PreparedConversion>
|
|
SelectedConversion = ConversionSet()) {
|
|
CanQualType RecordType = RD->getASTContext().getCanonicalTagType(RD);
|
|
|
|
ConversionSequence Result{FromType, RecordType};
|
|
Result.AfterFirstStandard = SelectedConversion->Seq.AfterFirstStandard;
|
|
|
|
ConversionSequence::UserDefinedConvertingConstructor Ctor;
|
|
Ctor.Fun = cast<CXXConstructorDecl>(SelectedConversion->ConversionFun);
|
|
Ctor.ConstructorParameterType = Ctor.Fun->getParamDecl(0)->getType();
|
|
Ctor.UserDefinedType = RecordType;
|
|
Result.setConversion(Ctor);
|
|
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< tryConvertingConstructors. Found result.\n");
|
|
return Result;
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "<<< tryConvertingConstructors. No conversion.\n");
|
|
return {};
|
|
}
|
|
|
|
/// Returns whether an expression of LType can be used in an RType context, as
|
|
/// per the implicit conversion rules.
|
|
///
|
|
/// Note: the result of this operation, unlike that of calculateMixability, is
|
|
/// **NOT** symmetric.
|
|
static MixData
|
|
approximateImplicitConversion(const TheCheck &Check, QualType LType,
|
|
QualType RType, const ASTContext &Ctx,
|
|
ImplicitConversionModellingMode ImplicitMode) {
|
|
LLVM_DEBUG(llvm::dbgs() << ">>> approximateImplicitConversion for LType:\n";
|
|
LType.dump(llvm::dbgs(), Ctx); llvm::dbgs() << "\nand RType:\n";
|
|
RType.dump(llvm::dbgs(), Ctx);
|
|
llvm::dbgs() << "\nimplicit mode: "; switch (ImplicitMode) {
|
|
case ImplicitConversionModellingMode::None:
|
|
llvm::dbgs() << "None";
|
|
break;
|
|
case ImplicitConversionModellingMode::All:
|
|
llvm::dbgs() << "All";
|
|
break;
|
|
case ImplicitConversionModellingMode::OneWaySingleStandardOnly:
|
|
llvm::dbgs() << "OneWay, Single, STD Only";
|
|
break;
|
|
} llvm::dbgs() << '\n';);
|
|
if (LType == RType)
|
|
return {MixFlags::Trivial, LType};
|
|
|
|
// An implicit conversion sequence consists of the following, in order:
|
|
// * Maybe standard conversion sequence.
|
|
// * Maybe user-defined conversion.
|
|
// * Maybe standard conversion sequence.
|
|
ConversionSequence ImplicitSeq{LType, RType};
|
|
QualType WorkType = LType;
|
|
|
|
std::optional<QualType> AfterFirstStdConv =
|
|
approximateStandardConversionSequence(Check, LType, RType, Ctx);
|
|
if (AfterFirstStdConv) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateImplicitConversion. Standard "
|
|
"Pre-Conversion found!\n");
|
|
ImplicitSeq.AfterFirstStandard = *AfterFirstStdConv;
|
|
WorkType = ImplicitSeq.AfterFirstStandard;
|
|
}
|
|
|
|
if (ImplicitMode == ImplicitConversionModellingMode::OneWaySingleStandardOnly)
|
|
// If the caller only requested modelling of a standard conversion, bail.
|
|
return {ImplicitSeq.AfterFirstStandard.isNull()
|
|
? MixFlags::None
|
|
: MixFlags::ImplicitConversion,
|
|
ImplicitSeq};
|
|
|
|
if (Ctx.getLangOpts().CPlusPlus) {
|
|
bool FoundConversionOperator = false, FoundConvertingCtor = false;
|
|
|
|
if (const auto *LRD = WorkType->getAsCXXRecordDecl()) {
|
|
std::optional<ConversionSequence> ConversionOperatorResult =
|
|
tryConversionOperators(Check, LRD, RType);
|
|
if (ConversionOperatorResult) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateImplicitConversion. Found "
|
|
"conversion operator.\n");
|
|
ImplicitSeq.update(*ConversionOperatorResult);
|
|
WorkType = ImplicitSeq.getTypeAfterUserDefinedConversion();
|
|
FoundConversionOperator = true;
|
|
}
|
|
}
|
|
|
|
if (const auto *RRD = RType->getAsCXXRecordDecl()) {
|
|
// Use the original "LType" here, and not WorkType, because the
|
|
// conversion to the converting constructors' parameters will be
|
|
// modelled in the recursive call.
|
|
std::optional<ConversionSequence> ConvCtorResult =
|
|
tryConvertingConstructors(Check, LType, RRD);
|
|
if (ConvCtorResult) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateImplicitConversion. Found "
|
|
"converting constructor.\n");
|
|
ImplicitSeq.update(*ConvCtorResult);
|
|
WorkType = ImplicitSeq.getTypeAfterUserDefinedConversion();
|
|
FoundConvertingCtor = true;
|
|
}
|
|
}
|
|
|
|
if (FoundConversionOperator && FoundConvertingCtor) {
|
|
// If both an operator and a ctor matches, the sequence is ambiguous.
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< approximateImplicitConversion. Found both "
|
|
"user-defined conversion kinds in the same sequence!\n");
|
|
return {MixFlags::None};
|
|
}
|
|
}
|
|
|
|
// After the potential user-defined conversion, another standard conversion
|
|
// sequence might exist.
|
|
LLVM_DEBUG(
|
|
llvm::dbgs()
|
|
<< "--- approximateImplicitConversion. Try to find post-conversion.\n");
|
|
MixData SecondStdConv = approximateImplicitConversion(
|
|
Check, WorkType, RType, Ctx,
|
|
ImplicitConversionModellingMode::OneWaySingleStandardOnly);
|
|
if (SecondStdConv.indicatesMixability()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "--- approximateImplicitConversion. Standard "
|
|
"Post-Conversion found!\n");
|
|
|
|
// The single-step modelling puts the modelled conversion into the "PreStd"
|
|
// variable in the recursive call, but from the PoV of this function, it is
|
|
// the post-conversion.
|
|
ImplicitSeq.AfterSecondStandard =
|
|
SecondStdConv.Conversion.AfterFirstStandard;
|
|
WorkType = ImplicitSeq.AfterSecondStandard;
|
|
}
|
|
|
|
if (ImplicitSeq) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "<<< approximateImplicitConversion. Found a conversion.\n");
|
|
return {MixFlags::ImplicitConversion, ImplicitSeq};
|
|
}
|
|
|
|
LLVM_DEBUG(
|
|
llvm::dbgs() << "<<< approximateImplicitConversion. No match found.\n");
|
|
return {MixFlags::None};
|
|
}
|
|
|
|
static MixableParameterRange modelMixingRange(
|
|
const TheCheck &Check, const FunctionDecl *FD, std::size_t StartIndex,
|
|
const filter::SimilarlyUsedParameterPairSuppressor &UsageBasedSuppressor) {
|
|
std::size_t NumParams = FD->getNumParams();
|
|
assert(StartIndex < NumParams && "out of bounds for start");
|
|
const ASTContext &Ctx = FD->getASTContext();
|
|
|
|
MixableParameterRange Ret;
|
|
// A parameter at index 'StartIndex' had been trivially "checked".
|
|
Ret.NumParamsChecked = 1;
|
|
|
|
for (std::size_t I = StartIndex + 1; I < NumParams; ++I) {
|
|
const ParmVarDecl *Ith = FD->getParamDecl(I);
|
|
StringRef ParamName = Ith->getName();
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "Check param #" << I << " '" << ParamName << "'...\n");
|
|
if (filter::isIgnoredParameter(Check, Ith)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Param #" << I << " is ignored. Break!\n");
|
|
break;
|
|
}
|
|
|
|
StringRef PrevParamName = FD->getParamDecl(I - 1)->getName();
|
|
if (!ParamName.empty() && !PrevParamName.empty() &&
|
|
filter::prefixSuffixCoverUnderThreshold(
|
|
Check.NamePrefixSuffixSilenceDissimilarityTreshold, PrevParamName,
|
|
ParamName)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Parameter '" << ParamName
|
|
<< "' follows a pattern with previous parameter '"
|
|
<< PrevParamName << "'. Break!\n");
|
|
break;
|
|
}
|
|
|
|
// Now try to go forward and build the range of [Start, ..., I, I + 1, ...]
|
|
// parameters that can be messed up at a call site.
|
|
MixableParameterRange::MixVector MixesOfIth;
|
|
for (std::size_t J = StartIndex; J < I; ++J) {
|
|
const ParmVarDecl *Jth = FD->getParamDecl(J);
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "Check mix of #" << J << " against #" << I << "...\n");
|
|
|
|
if (isSimilarlyUsedParameter(UsageBasedSuppressor, Ith, Jth)) {
|
|
// Consider the two similarly used parameters to not be possible in a
|
|
// mix-up at the user's request, if they enabled this heuristic.
|
|
LLVM_DEBUG(llvm::dbgs() << "Parameters #" << I << " and #" << J
|
|
<< " deemed related, ignoring...\n");
|
|
|
|
// If the parameter #I and #J mixes, then I is mixable with something
|
|
// in the current range, so the range has to be broken and I not
|
|
// included.
|
|
MixesOfIth.clear();
|
|
break;
|
|
}
|
|
|
|
Mix M{Jth, Ith,
|
|
calculateMixability(Check, Jth->getType(), Ith->getType(), Ctx,
|
|
Check.ModelImplicitConversions
|
|
? ImplicitConversionModellingMode::All
|
|
: ImplicitConversionModellingMode::None)};
|
|
LLVM_DEBUG(llvm::dbgs() << "Mix flags (raw) : "
|
|
<< formatMixFlags(M.flags()) << '\n');
|
|
M.sanitize();
|
|
LLVM_DEBUG(llvm::dbgs() << "Mix flags (after sanitize): "
|
|
<< formatMixFlags(M.flags()) << '\n');
|
|
|
|
assert(M.flagsValid() && "All flags decayed!");
|
|
|
|
if (M.mixable())
|
|
MixesOfIth.emplace_back(std::move(M));
|
|
}
|
|
|
|
if (MixesOfIth.empty()) {
|
|
// If there weren't any new mixes stored for Ith, the range is
|
|
// [Start, ..., I].
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "Param #" << I
|
|
<< " does not mix with any in the current range. Break!\n");
|
|
break;
|
|
}
|
|
|
|
Ret.Mixes.insert(Ret.Mixes.end(), MixesOfIth.begin(), MixesOfIth.end());
|
|
++Ret.NumParamsChecked; // Otherwise a new param was iterated.
|
|
}
|
|
|
|
return Ret;
|
|
}
|
|
|
|
} // namespace model
|
|
|
|
namespace {
|
|
/// Matches DeclRefExprs and their ignorable wrappers to ParmVarDecls.
|
|
AST_MATCHER_FUNCTION(ast_matchers::internal::Matcher<Stmt>, paramRefExpr) {
|
|
return expr(ignoringParenImpCasts(ignoringElidableConstructorCall(
|
|
declRefExpr(to(parmVarDecl().bind("param"))))));
|
|
}
|
|
} // namespace
|
|
|
|
namespace filter {
|
|
|
|
/// Returns whether the parameter's name or the parameter's type's name is
|
|
/// configured by the user to be ignored from analysis and diagnostic.
|
|
static bool isIgnoredParameter(const TheCheck &Check, const ParmVarDecl *Node) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Checking if '" << Node->getName()
|
|
<< "' is ignored.\n");
|
|
|
|
if (!Node->getIdentifier())
|
|
return llvm::is_contained(Check.IgnoredParameterNames, "\"\"");
|
|
|
|
StringRef NodeName = Node->getName();
|
|
if (llvm::is_contained(Check.IgnoredParameterNames, NodeName)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "\tName ignored.\n");
|
|
return true;
|
|
}
|
|
|
|
StringRef NodeTypeName = [Node] {
|
|
const ASTContext &Ctx = Node->getASTContext();
|
|
const SourceManager &SM = Ctx.getSourceManager();
|
|
SourceLocation B = Node->getTypeSpecStartLoc();
|
|
SourceLocation E = Node->getTypeSpecEndLoc();
|
|
LangOptions LO;
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "\tType name code is '"
|
|
<< Lexer::getSourceText(
|
|
CharSourceRange::getTokenRange(B, E), SM, LO)
|
|
<< "'...\n");
|
|
if (B.isMacroID()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "\t\tBeginning is macro.\n");
|
|
B = SM.getTopMacroCallerLoc(B);
|
|
}
|
|
if (E.isMacroID()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "\t\tEnding is macro.\n");
|
|
E = Lexer::getLocForEndOfToken(SM.getTopMacroCallerLoc(E), 0, SM, LO);
|
|
}
|
|
LLVM_DEBUG(llvm::dbgs() << "\tType name code is '"
|
|
<< Lexer::getSourceText(
|
|
CharSourceRange::getTokenRange(B, E), SM, LO)
|
|
<< "'...\n");
|
|
|
|
return Lexer::getSourceText(CharSourceRange::getTokenRange(B, E), SM, LO);
|
|
}();
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "\tType name is '" << NodeTypeName << "'\n");
|
|
if (!NodeTypeName.empty()) {
|
|
if (llvm::any_of(Check.IgnoredParameterTypeSuffixes,
|
|
[NodeTypeName](StringRef E) {
|
|
return !E.empty() && NodeTypeName.ends_with(E);
|
|
})) {
|
|
LLVM_DEBUG(llvm::dbgs() << "\tType suffix ignored.\n");
|
|
return true;
|
|
}
|
|
}
|
|
|
|
return false;
|
|
}
|
|
|
|
/// This namespace contains the implementations for the suppression of
|
|
/// diagnostics from similarly-used ("related") parameters.
|
|
namespace relatedness_heuristic {
|
|
|
|
static constexpr std::size_t SmallDataStructureSize = 4;
|
|
|
|
template <typename T, std::size_t N = SmallDataStructureSize>
|
|
using ParamToSmallSetMap =
|
|
llvm::DenseMap<const ParmVarDecl *, llvm::SmallSet<T, N>>;
|
|
|
|
/// Returns whether the sets mapped to the two elements in the map have at
|
|
/// least one element in common.
|
|
template <typename MapTy, typename ElemTy>
|
|
static bool lazyMapOfSetsIntersectionExists(const MapTy &Map, const ElemTy &E1,
|
|
const ElemTy &E2) {
|
|
auto E1Iterator = Map.find(E1);
|
|
auto E2Iterator = Map.find(E2);
|
|
if (E1Iterator == Map.end() || E2Iterator == Map.end())
|
|
return false;
|
|
|
|
for (const auto &E1SetElem : E1Iterator->second)
|
|
if (E2Iterator->second.contains(E1SetElem))
|
|
return true;
|
|
|
|
return false;
|
|
}
|
|
|
|
/// Implements the heuristic that marks two parameters related if there is
|
|
/// a usage for both in the same strict expression subtree. A strict
|
|
/// expression subtree is a tree which only includes Expr nodes, i.e. no
|
|
/// Stmts and no Decls.
|
|
class AppearsInSameExpr : public RecursiveASTVisitor<AppearsInSameExpr> {
|
|
using Base = RecursiveASTVisitor<AppearsInSameExpr>;
|
|
|
|
const FunctionDecl *FD;
|
|
const Expr *CurrentExprOnlyTreeRoot = nullptr;
|
|
llvm::DenseMap<const ParmVarDecl *,
|
|
llvm::SmallPtrSet<const Expr *, SmallDataStructureSize>>
|
|
ParentExprsForParamRefs;
|
|
|
|
public:
|
|
void setup(const FunctionDecl *FD) {
|
|
this->FD = FD;
|
|
TraverseFunctionDecl(const_cast<FunctionDecl *>(FD));
|
|
}
|
|
|
|
bool operator()(const ParmVarDecl *Param1, const ParmVarDecl *Param2) const {
|
|
return lazyMapOfSetsIntersectionExists(ParentExprsForParamRefs, Param1,
|
|
Param2);
|
|
}
|
|
|
|
bool TraverseDecl(Decl *D) {
|
|
CurrentExprOnlyTreeRoot = nullptr;
|
|
return Base::TraverseDecl(D);
|
|
}
|
|
|
|
bool TraverseStmt(Stmt *S, DataRecursionQueue *Queue = nullptr) {
|
|
if (auto *E = dyn_cast_or_null<Expr>(S)) {
|
|
bool RootSetInCurrentStackFrame = false;
|
|
if (!CurrentExprOnlyTreeRoot) {
|
|
CurrentExprOnlyTreeRoot = E;
|
|
RootSetInCurrentStackFrame = true;
|
|
}
|
|
|
|
bool Ret = Base::TraverseStmt(S);
|
|
|
|
if (RootSetInCurrentStackFrame)
|
|
CurrentExprOnlyTreeRoot = nullptr;
|
|
|
|
return Ret;
|
|
}
|
|
|
|
// A Stmt breaks the strictly Expr subtree.
|
|
CurrentExprOnlyTreeRoot = nullptr;
|
|
return Base::TraverseStmt(S);
|
|
}
|
|
|
|
bool VisitDeclRefExpr(DeclRefExpr *DRE) {
|
|
if (!CurrentExprOnlyTreeRoot)
|
|
return true;
|
|
|
|
if (auto *PVD = dyn_cast<ParmVarDecl>(DRE->getDecl()))
|
|
if (llvm::find(FD->parameters(), PVD))
|
|
ParentExprsForParamRefs[PVD].insert(CurrentExprOnlyTreeRoot);
|
|
|
|
return true;
|
|
}
|
|
};
|
|
|
|
/// Implements the heuristic that marks two parameters related if there are
|
|
/// two separate calls to the same function (overload) and the parameters are
|
|
/// passed to the same index in both calls, i.e f(a, b) and f(a, c) passes
|
|
/// b and c to the same index (2) of f(), marking them related.
|
|
class PassedToSameFunction {
|
|
ParamToSmallSetMap<std::pair<const FunctionDecl *, unsigned>> TargetParams;
|
|
|
|
public:
|
|
void setup(const FunctionDecl *FD) {
|
|
auto ParamsAsArgsInFnCalls =
|
|
match(functionDecl(forEachDescendant(
|
|
callExpr(forEachArgumentWithParam(
|
|
paramRefExpr(), parmVarDecl().bind("passed-to")))
|
|
.bind("call-expr"))),
|
|
*FD, FD->getASTContext());
|
|
for (const auto &Match : ParamsAsArgsInFnCalls) {
|
|
const auto *PassedParamOfThisFn = Match.getNodeAs<ParmVarDecl>("param");
|
|
const auto *CE = Match.getNodeAs<CallExpr>("call-expr");
|
|
const auto *PassedToParam = Match.getNodeAs<ParmVarDecl>("passed-to");
|
|
assert(PassedParamOfThisFn && CE && PassedToParam);
|
|
|
|
const FunctionDecl *CalledFn = CE->getDirectCallee();
|
|
if (!CalledFn)
|
|
continue;
|
|
|
|
std::optional<unsigned> TargetIdx;
|
|
unsigned NumFnParams = CalledFn->getNumParams();
|
|
for (unsigned Idx = 0; Idx < NumFnParams; ++Idx)
|
|
if (CalledFn->getParamDecl(Idx) == PassedToParam)
|
|
TargetIdx.emplace(Idx);
|
|
|
|
assert(TargetIdx && "Matched, but didn't find index?");
|
|
TargetParams[PassedParamOfThisFn].insert(
|
|
{CalledFn->getCanonicalDecl(), *TargetIdx});
|
|
}
|
|
}
|
|
|
|
bool operator()(const ParmVarDecl *Param1, const ParmVarDecl *Param2) const {
|
|
return lazyMapOfSetsIntersectionExists(TargetParams, Param1, Param2);
|
|
}
|
|
};
|
|
|
|
/// Implements the heuristic that marks two parameters related if the same
|
|
/// member is accessed (referred to) inside the current function's body.
|
|
class AccessedSameMemberOf {
|
|
ParamToSmallSetMap<const Decl *> AccessedMembers;
|
|
|
|
public:
|
|
void setup(const FunctionDecl *FD) {
|
|
auto MembersCalledOnParams = match(
|
|
functionDecl(forEachDescendant(
|
|
memberExpr(hasObjectExpression(paramRefExpr())).bind("mem-expr"))),
|
|
*FD, FD->getASTContext());
|
|
|
|
for (const auto &Match : MembersCalledOnParams) {
|
|
const auto *AccessedParam = Match.getNodeAs<ParmVarDecl>("param");
|
|
const auto *ME = Match.getNodeAs<MemberExpr>("mem-expr");
|
|
assert(AccessedParam && ME);
|
|
AccessedMembers[AccessedParam].insert(
|
|
ME->getMemberDecl()->getCanonicalDecl());
|
|
}
|
|
}
|
|
|
|
bool operator()(const ParmVarDecl *Param1, const ParmVarDecl *Param2) const {
|
|
return lazyMapOfSetsIntersectionExists(AccessedMembers, Param1, Param2);
|
|
}
|
|
};
|
|
|
|
/// Implements the heuristic that marks two parameters related if different
|
|
/// ReturnStmts return them from the function.
|
|
class Returned {
|
|
llvm::SmallVector<const ParmVarDecl *, SmallDataStructureSize> ReturnedParams;
|
|
|
|
public:
|
|
void setup(const FunctionDecl *FD) {
|
|
// TODO: Handle co_return.
|
|
auto ParamReturns = match(functionDecl(forEachDescendant(
|
|
returnStmt(hasReturnValue(paramRefExpr())))),
|
|
*FD, FD->getASTContext());
|
|
for (const auto &Match : ParamReturns) {
|
|
const auto *ReturnedParam = Match.getNodeAs<ParmVarDecl>("param");
|
|
assert(ReturnedParam);
|
|
|
|
if (find(FD->parameters(), ReturnedParam) == FD->param_end())
|
|
// Inside the subtree of a FunctionDecl there might be ReturnStmts of
|
|
// a parameter that isn't the parameter of the function, e.g. in the
|
|
// case of lambdas.
|
|
continue;
|
|
|
|
ReturnedParams.emplace_back(ReturnedParam);
|
|
}
|
|
}
|
|
|
|
bool operator()(const ParmVarDecl *Param1, const ParmVarDecl *Param2) const {
|
|
return llvm::is_contained(ReturnedParams, Param1) &&
|
|
llvm::is_contained(ReturnedParams, Param2);
|
|
}
|
|
};
|
|
|
|
} // namespace relatedness_heuristic
|
|
|
|
/// Helper class that is used to detect if two parameters of the same function
|
|
/// are used in a similar fashion, to suppress the result.
|
|
class SimilarlyUsedParameterPairSuppressor {
|
|
const bool Enabled;
|
|
relatedness_heuristic::AppearsInSameExpr SameExpr;
|
|
relatedness_heuristic::PassedToSameFunction PassToFun;
|
|
relatedness_heuristic::AccessedSameMemberOf SameMember;
|
|
relatedness_heuristic::Returned Returns;
|
|
|
|
public:
|
|
SimilarlyUsedParameterPairSuppressor(const FunctionDecl *FD, bool Enable)
|
|
: Enabled(Enable) {
|
|
if (!Enable)
|
|
return;
|
|
|
|
SameExpr.setup(FD);
|
|
PassToFun.setup(FD);
|
|
SameMember.setup(FD);
|
|
Returns.setup(FD);
|
|
}
|
|
|
|
/// Returns whether the specified two parameters are deemed similarly used
|
|
/// or related by the heuristics.
|
|
bool operator()(const ParmVarDecl *Param1, const ParmVarDecl *Param2) const {
|
|
if (!Enabled)
|
|
return false;
|
|
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "::: Matching similar usage / relatedness heuristic...\n");
|
|
|
|
if (SameExpr(Param1, Param2)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "::: Used in the same expression.\n");
|
|
return true;
|
|
}
|
|
|
|
if (PassToFun(Param1, Param2)) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "::: Passed to same function in different calls.\n");
|
|
return true;
|
|
}
|
|
|
|
if (SameMember(Param1, Param2)) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "::: Same member field access or method called.\n");
|
|
return true;
|
|
}
|
|
|
|
if (Returns(Param1, Param2)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "::: Both parameter returned.\n");
|
|
return true;
|
|
}
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "::: None.\n");
|
|
return false;
|
|
}
|
|
};
|
|
|
|
// (This function hoists the call to operator() of the wrapper, so we do not
|
|
// need to define the previous class at the top of the file.)
|
|
static inline bool
|
|
isSimilarlyUsedParameter(const SimilarlyUsedParameterPairSuppressor &Suppressor,
|
|
const ParmVarDecl *Param1, const ParmVarDecl *Param2) {
|
|
return Suppressor(Param1, Param2);
|
|
}
|
|
|
|
static void padStringAtEnd(SmallVectorImpl<char> &Str, std::size_t ToLen) {
|
|
while (Str.size() < ToLen)
|
|
Str.emplace_back('\0');
|
|
}
|
|
|
|
static void padStringAtBegin(SmallVectorImpl<char> &Str, std::size_t ToLen) {
|
|
while (Str.size() < ToLen)
|
|
Str.insert(Str.begin(), '\0');
|
|
}
|
|
|
|
static bool isCommonPrefixWithoutSomeCharacters(std::size_t N, StringRef S1,
|
|
StringRef S2) {
|
|
assert(S1.size() >= N && S2.size() >= N);
|
|
StringRef S1Prefix = S1.take_front(S1.size() - N),
|
|
S2Prefix = S2.take_front(S2.size() - N);
|
|
return S1Prefix == S2Prefix && !S1Prefix.empty();
|
|
}
|
|
|
|
static bool isCommonSuffixWithoutSomeCharacters(std::size_t N, StringRef S1,
|
|
StringRef S2) {
|
|
assert(S1.size() >= N && S2.size() >= N);
|
|
StringRef S1Suffix = S1.take_back(S1.size() - N),
|
|
S2Suffix = S2.take_back(S2.size() - N);
|
|
return S1Suffix == S2Suffix && !S1Suffix.empty();
|
|
}
|
|
|
|
/// Returns whether the two strings are prefixes or suffixes of each other with
|
|
/// at most Threshold characters differing on the non-common end.
|
|
static bool prefixSuffixCoverUnderThreshold(std::size_t Threshold,
|
|
StringRef Str1, StringRef Str2) {
|
|
if (Threshold == 0)
|
|
return false;
|
|
|
|
// Pad the two strings to the longer length.
|
|
std::size_t BiggerLength = std::max(Str1.size(), Str2.size());
|
|
|
|
if (BiggerLength <= Threshold)
|
|
// If the length of the strings is still smaller than the threshold, they
|
|
// would be covered by an empty prefix/suffix with the rest differing.
|
|
// (E.g. "A" and "X" with Threshold = 1 would mean we think they are
|
|
// similar and do not warn about them, which is a too eager assumption.)
|
|
return false;
|
|
|
|
SmallString<32> S1PadE{Str1}, S2PadE{Str2};
|
|
padStringAtEnd(S1PadE, BiggerLength);
|
|
padStringAtEnd(S2PadE, BiggerLength);
|
|
|
|
if (isCommonPrefixWithoutSomeCharacters(
|
|
Threshold, StringRef{S1PadE.begin(), BiggerLength},
|
|
StringRef{S2PadE.begin(), BiggerLength}))
|
|
return true;
|
|
|
|
SmallString<32> S1PadB{Str1}, S2PadB{Str2};
|
|
padStringAtBegin(S1PadB, BiggerLength);
|
|
padStringAtBegin(S2PadB, BiggerLength);
|
|
|
|
if (isCommonSuffixWithoutSomeCharacters(
|
|
Threshold, StringRef{S1PadB.begin(), BiggerLength},
|
|
StringRef{S2PadB.begin(), BiggerLength}))
|
|
return true;
|
|
|
|
return false;
|
|
}
|
|
|
|
} // namespace filter
|
|
|
|
namespace {
|
|
|
|
/// Matches functions that have at least the specified amount of parameters.
|
|
AST_MATCHER_P(FunctionDecl, parameterCountGE, unsigned, N) {
|
|
return Node.getNumParams() >= N;
|
|
}
|
|
|
|
/// Matches *any* overloaded unary and binary operators.
|
|
AST_MATCHER(FunctionDecl, isOverloadedUnaryOrBinaryOperator) {
|
|
switch (Node.getOverloadedOperator()) {
|
|
case OO_None:
|
|
case OO_New:
|
|
case OO_Delete:
|
|
case OO_Array_New:
|
|
case OO_Array_Delete:
|
|
case OO_Conditional:
|
|
case OO_Coawait:
|
|
return false;
|
|
|
|
default:
|
|
return Node.getNumParams() <= 2;
|
|
}
|
|
}
|
|
|
|
} // namespace
|
|
|
|
/// Returns the DefaultMinimumLength if the Value of requested minimum length
|
|
/// is less than 2. Minimum lengths of 0 or 1 are not accepted.
|
|
static inline unsigned clampMinimumLength(const unsigned Value) {
|
|
return Value < 2 ? DefaultMinimumLength : Value;
|
|
}
|
|
|
|
// FIXME: Maybe unneeded, getNameForDiagnostic() is expected to change to return
|
|
// a crafted location when the node itself is unnamed. (See D84658, D85033.)
|
|
/// Returns the diagnostic-friendly name of the node, or empty string.
|
|
static SmallString<64> getName(const NamedDecl *ND) {
|
|
SmallString<64> Name;
|
|
llvm::raw_svector_ostream OS{Name};
|
|
ND->getNameForDiagnostic(OS, ND->getASTContext().getPrintingPolicy(), false);
|
|
return Name;
|
|
}
|
|
|
|
/// Returns the diagnostic-friendly name of the node, or a constant value.
|
|
static SmallString<64> getNameOrUnnamed(const NamedDecl *ND) {
|
|
auto Name = getName(ND);
|
|
if (Name.empty())
|
|
Name = "<unnamed>";
|
|
return Name;
|
|
}
|
|
|
|
/// Returns whether a particular Mix between two parameters should have the
|
|
/// types involved diagnosed to the user. This is only a flag check.
|
|
static inline bool needsToPrintTypeInDiagnostic(const model::Mix &M) {
|
|
using namespace model;
|
|
return static_cast<bool>(
|
|
M.flags() &
|
|
(MixFlags::TypeAlias | MixFlags::ReferenceBind | MixFlags::Qualifiers));
|
|
}
|
|
|
|
/// Returns whether a particular Mix between the two parameters should have
|
|
/// implicit conversions elaborated.
|
|
static inline bool needsToElaborateImplicitConversion(const model::Mix &M) {
|
|
return hasFlag(M.flags(), model::MixFlags::ImplicitConversion);
|
|
}
|
|
|
|
namespace {
|
|
|
|
/// This class formats a conversion sequence into a "Ty1 -> Ty2 -> Ty3" line
|
|
/// that can be used in diagnostics.
|
|
struct FormattedConversionSequence {
|
|
std::string DiagnosticText;
|
|
|
|
/// The formatted sequence is trivial if it is "Ty1 -> Ty2", but Ty1 and
|
|
/// Ty2 are the types that are shown in the code. A trivial diagnostic
|
|
/// does not need to be printed.
|
|
bool Trivial = true;
|
|
|
|
FormattedConversionSequence(const PrintingPolicy &PP,
|
|
StringRef StartTypeAsDiagnosed,
|
|
const model::ConversionSequence &Conv,
|
|
StringRef DestinationTypeAsDiagnosed) {
|
|
llvm::raw_string_ostream OS{DiagnosticText};
|
|
|
|
// Print the type name as it is printed in other places in the diagnostic.
|
|
OS << '\'' << StartTypeAsDiagnosed << '\'';
|
|
std::string LastAddedType = StartTypeAsDiagnosed.str();
|
|
std::size_t NumElementsAdded = 1;
|
|
|
|
// However, the parameter's defined type might not be what the implicit
|
|
// conversion started with, e.g. if a typedef is found to convert.
|
|
std::string SeqBeginTypeStr = Conv.Begin.getAsString(PP);
|
|
std::string SeqEndTypeStr = Conv.End.getAsString(PP);
|
|
if (StartTypeAsDiagnosed != SeqBeginTypeStr) {
|
|
OS << " (as '" << SeqBeginTypeStr << "')";
|
|
LastAddedType = SeqBeginTypeStr;
|
|
Trivial = false;
|
|
}
|
|
|
|
auto AddType = [&](StringRef ToAdd) {
|
|
if (LastAddedType != ToAdd && ToAdd != SeqEndTypeStr) {
|
|
OS << " -> '" << ToAdd << "'";
|
|
LastAddedType = ToAdd.str();
|
|
++NumElementsAdded;
|
|
}
|
|
};
|
|
for (QualType InvolvedType : Conv.getInvolvedTypesInSequence())
|
|
// Print every type that's unique in the sequence into the diagnosis.
|
|
AddType(InvolvedType.getAsString(PP));
|
|
|
|
if (LastAddedType != DestinationTypeAsDiagnosed) {
|
|
OS << " -> '" << DestinationTypeAsDiagnosed << "'";
|
|
LastAddedType = DestinationTypeAsDiagnosed.str();
|
|
++NumElementsAdded;
|
|
}
|
|
|
|
// Same reasoning as with the Begin, e.g. if the converted-to type is a
|
|
// typedef, it will not be the same inside the conversion sequence (where
|
|
// the model already tore off typedefs) as in the code.
|
|
if (DestinationTypeAsDiagnosed != SeqEndTypeStr) {
|
|
OS << " (as '" << SeqEndTypeStr << "')";
|
|
LastAddedType = SeqEndTypeStr;
|
|
Trivial = false;
|
|
}
|
|
|
|
if (Trivial && NumElementsAdded > 2)
|
|
// If the thing is still marked trivial but we have more than the
|
|
// from and to types added, it should not be trivial, and elaborated
|
|
// when printing the diagnostic.
|
|
Trivial = false;
|
|
}
|
|
};
|
|
|
|
/// Retains the elements called with and returns whether the call is done with
|
|
/// a new element.
|
|
template <typename E, std::size_t N> class InsertOnce {
|
|
llvm::SmallSet<E, N> CalledWith;
|
|
|
|
public:
|
|
bool operator()(E El) { return CalledWith.insert(std::move(El)).second; }
|
|
|
|
bool calledWith(const E &El) const { return CalledWith.contains(El); }
|
|
};
|
|
|
|
struct SwappedEqualQualTypePair {
|
|
QualType LHSType, RHSType;
|
|
|
|
bool operator==(const SwappedEqualQualTypePair &Other) const {
|
|
return (LHSType == Other.LHSType && RHSType == Other.RHSType) ||
|
|
(LHSType == Other.RHSType && RHSType == Other.LHSType);
|
|
}
|
|
|
|
bool operator<(const SwappedEqualQualTypePair &Other) const {
|
|
return LHSType < Other.LHSType && RHSType < Other.RHSType;
|
|
}
|
|
};
|
|
|
|
struct TypeAliasDiagnosticTuple {
|
|
QualType LHSType, RHSType, CommonType;
|
|
|
|
bool operator==(const TypeAliasDiagnosticTuple &Other) const {
|
|
return CommonType == Other.CommonType &&
|
|
((LHSType == Other.LHSType && RHSType == Other.RHSType) ||
|
|
(LHSType == Other.RHSType && RHSType == Other.LHSType));
|
|
}
|
|
|
|
bool operator<(const TypeAliasDiagnosticTuple &Other) const {
|
|
return CommonType < Other.CommonType && LHSType < Other.LHSType &&
|
|
RHSType < Other.RHSType;
|
|
}
|
|
};
|
|
|
|
/// Helper class to only emit a diagnostic related to MixFlags::TypeAlias once.
|
|
class UniqueTypeAliasDiagnosticHelper
|
|
: public InsertOnce<TypeAliasDiagnosticTuple, 8> {
|
|
using Base = InsertOnce<TypeAliasDiagnosticTuple, 8>;
|
|
|
|
public:
|
|
/// Returns whether the diagnostic for LHSType and RHSType which are both
|
|
/// referring to CommonType being the same has not been emitted already.
|
|
bool operator()(QualType LHSType, QualType RHSType, QualType CommonType) {
|
|
if (CommonType.isNull() || CommonType == LHSType || CommonType == RHSType)
|
|
return Base::operator()({LHSType, RHSType, {}});
|
|
|
|
TypeAliasDiagnosticTuple ThreeTuple{LHSType, RHSType, CommonType};
|
|
if (!Base::operator()(ThreeTuple))
|
|
return false;
|
|
|
|
bool AlreadySaidLHSAndCommonIsSame = calledWith({LHSType, CommonType, {}});
|
|
bool AlreadySaidRHSAndCommonIsSame = calledWith({RHSType, CommonType, {}});
|
|
if (AlreadySaidLHSAndCommonIsSame && AlreadySaidRHSAndCommonIsSame) {
|
|
// "SomeInt == int" && "SomeOtherInt == int" => "Common(SomeInt,
|
|
// SomeOtherInt) == int", no need to diagnose it. Save the 3-tuple only
|
|
// for shortcut if it ever appears again.
|
|
return false;
|
|
}
|
|
|
|
return true;
|
|
}
|
|
};
|
|
|
|
} // namespace
|
|
|
|
EasilySwappableParametersCheck::EasilySwappableParametersCheck(
|
|
StringRef Name, ClangTidyContext *Context)
|
|
: ClangTidyCheck(Name, Context),
|
|
MinimumLength(clampMinimumLength(
|
|
Options.get("MinimumLength", DefaultMinimumLength))),
|
|
IgnoredParameterNames(optutils::parseStringList(
|
|
Options.get("IgnoredParameterNames", DefaultIgnoredParameterNames))),
|
|
IgnoredParameterTypeSuffixes(optutils::parseStringList(
|
|
Options.get("IgnoredParameterTypeSuffixes",
|
|
DefaultIgnoredParameterTypeSuffixes))),
|
|
QualifiersMix(Options.get("QualifiersMix", DefaultQualifiersMix)),
|
|
ModelImplicitConversions(Options.get("ModelImplicitConversions",
|
|
DefaultModelImplicitConversions)),
|
|
SuppressParametersUsedTogether(
|
|
Options.get("SuppressParametersUsedTogether",
|
|
DefaultSuppressParametersUsedTogether)),
|
|
NamePrefixSuffixSilenceDissimilarityTreshold(
|
|
Options.get("NamePrefixSuffixSilenceDissimilarityTreshold",
|
|
DefaultNamePrefixSuffixSilenceDissimilarityTreshold)) {}
|
|
|
|
void EasilySwappableParametersCheck::storeOptions(
|
|
ClangTidyOptions::OptionMap &Opts) {
|
|
Options.store(Opts, "MinimumLength", MinimumLength);
|
|
Options.store(Opts, "IgnoredParameterNames",
|
|
optutils::serializeStringList(IgnoredParameterNames));
|
|
Options.store(Opts, "IgnoredParameterTypeSuffixes",
|
|
optutils::serializeStringList(IgnoredParameterTypeSuffixes));
|
|
Options.store(Opts, "QualifiersMix", QualifiersMix);
|
|
Options.store(Opts, "ModelImplicitConversions", ModelImplicitConversions);
|
|
Options.store(Opts, "SuppressParametersUsedTogether",
|
|
SuppressParametersUsedTogether);
|
|
Options.store(Opts, "NamePrefixSuffixSilenceDissimilarityTreshold",
|
|
NamePrefixSuffixSilenceDissimilarityTreshold);
|
|
}
|
|
|
|
void EasilySwappableParametersCheck::registerMatchers(MatchFinder *Finder) {
|
|
const auto BaseConstraints = functionDecl(
|
|
// Only report for definition nodes, as fixing the issues reported
|
|
// requires the user to be able to change code.
|
|
isDefinition(), parameterCountGE(MinimumLength),
|
|
unless(isOverloadedUnaryOrBinaryOperator()));
|
|
|
|
Finder->addMatcher(
|
|
functionDecl(BaseConstraints,
|
|
unless(ast_matchers::isTemplateInstantiation()))
|
|
.bind("func"),
|
|
this);
|
|
Finder->addMatcher(
|
|
functionDecl(BaseConstraints, isExplicitTemplateSpecialization())
|
|
.bind("func"),
|
|
this);
|
|
}
|
|
|
|
void EasilySwappableParametersCheck::check(
|
|
const MatchFinder::MatchResult &Result) {
|
|
using namespace model;
|
|
using namespace filter;
|
|
|
|
const auto *FD = Result.Nodes.getNodeAs<FunctionDecl>("func");
|
|
assert(FD);
|
|
|
|
const PrintingPolicy &PP = FD->getASTContext().getPrintingPolicy();
|
|
std::size_t NumParams = FD->getNumParams();
|
|
std::size_t MixableRangeStartIndex = 0;
|
|
|
|
// Spawn one suppressor and if the user requested, gather information from
|
|
// the AST for the parameters' usages.
|
|
filter::SimilarlyUsedParameterPairSuppressor UsageBasedSuppressor{
|
|
FD, SuppressParametersUsedTogether};
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "Begin analysis of " << getName(FD) << " with "
|
|
<< NumParams << " parameters...\n");
|
|
while (MixableRangeStartIndex < NumParams) {
|
|
if (isIgnoredParameter(*this, FD->getParamDecl(MixableRangeStartIndex))) {
|
|
LLVM_DEBUG(llvm::dbgs()
|
|
<< "Parameter #" << MixableRangeStartIndex << " ignored.\n");
|
|
++MixableRangeStartIndex;
|
|
continue;
|
|
}
|
|
|
|
MixableParameterRange R = modelMixingRange(
|
|
*this, FD, MixableRangeStartIndex, UsageBasedSuppressor);
|
|
assert(R.NumParamsChecked > 0 && "Ensure forward progress!");
|
|
MixableRangeStartIndex += R.NumParamsChecked;
|
|
if (R.NumParamsChecked < MinimumLength) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Ignoring range of " << R.NumParamsChecked
|
|
<< " lower than limit.\n");
|
|
continue;
|
|
}
|
|
|
|
bool NeedsAnyTypeNote = llvm::any_of(R.Mixes, needsToPrintTypeInDiagnostic);
|
|
bool HasAnyImplicits =
|
|
llvm::any_of(R.Mixes, needsToElaborateImplicitConversion);
|
|
const ParmVarDecl *First = R.getFirstParam(), *Last = R.getLastParam();
|
|
std::string FirstParamTypeAsWritten = First->getType().getAsString(PP);
|
|
{
|
|
StringRef DiagText;
|
|
|
|
if (HasAnyImplicits)
|
|
DiagText = "%0 adjacent parameters of %1 of convertible types are "
|
|
"easily swapped by mistake";
|
|
else if (NeedsAnyTypeNote)
|
|
DiagText = "%0 adjacent parameters of %1 of similar type are easily "
|
|
"swapped by mistake";
|
|
else
|
|
DiagText = "%0 adjacent parameters of %1 of similar type ('%2') are "
|
|
"easily swapped by mistake";
|
|
|
|
auto Diag = diag(First->getOuterLocStart(), DiagText)
|
|
<< static_cast<unsigned>(R.NumParamsChecked) << FD;
|
|
if (!NeedsAnyTypeNote)
|
|
Diag << FirstParamTypeAsWritten;
|
|
|
|
CharSourceRange HighlightRange = CharSourceRange::getTokenRange(
|
|
First->getBeginLoc(), Last->getEndLoc());
|
|
Diag << HighlightRange;
|
|
}
|
|
|
|
// There is a chance that the previous highlight did not succeed, e.g. when
|
|
// the two parameters are on different lines. For clarity, show the user
|
|
// the involved variable explicitly.
|
|
diag(First->getLocation(), "the first parameter in the range is '%0'",
|
|
DiagnosticIDs::Note)
|
|
<< getNameOrUnnamed(First)
|
|
<< CharSourceRange::getTokenRange(First->getLocation(),
|
|
First->getLocation());
|
|
diag(Last->getLocation(), "the last parameter in the range is '%0'",
|
|
DiagnosticIDs::Note)
|
|
<< getNameOrUnnamed(Last)
|
|
<< CharSourceRange::getTokenRange(Last->getLocation(),
|
|
Last->getLocation());
|
|
|
|
// Helper classes to silence elaborative diagnostic notes that would be
|
|
// too verbose.
|
|
UniqueTypeAliasDiagnosticHelper UniqueTypeAlias;
|
|
InsertOnce<SwappedEqualQualTypePair, 8> UniqueBindPower;
|
|
InsertOnce<SwappedEqualQualTypePair, 8> UniqueImplicitConversion;
|
|
|
|
for (const model::Mix &M : R.Mixes) {
|
|
assert(M.mixable() && "Sentinel or false mix in result.");
|
|
if (!needsToPrintTypeInDiagnostic(M) &&
|
|
!needsToElaborateImplicitConversion(M))
|
|
continue;
|
|
|
|
// Typedefs might result in the type of the variable needing to be
|
|
// emitted to a note diagnostic, so prepare it.
|
|
const ParmVarDecl *LVar = M.First;
|
|
const ParmVarDecl *RVar = M.Second;
|
|
QualType LType = LVar->getType();
|
|
QualType RType = RVar->getType();
|
|
QualType CommonType = M.commonUnderlyingType();
|
|
std::string LTypeStr = LType.getAsString(PP);
|
|
std::string RTypeStr = RType.getAsString(PP);
|
|
std::string CommonTypeStr = CommonType.getAsString(PP);
|
|
|
|
if (hasFlag(M.flags(), MixFlags::TypeAlias) &&
|
|
UniqueTypeAlias(LType, RType, CommonType)) {
|
|
StringRef DiagText;
|
|
bool ExplicitlyPrintCommonType = false;
|
|
if (LTypeStr == CommonTypeStr || RTypeStr == CommonTypeStr) {
|
|
if (hasFlag(M.flags(), MixFlags::Qualifiers))
|
|
DiagText = "after resolving type aliases, '%0' and '%1' share a "
|
|
"common type";
|
|
else
|
|
DiagText =
|
|
"after resolving type aliases, '%0' and '%1' are the same";
|
|
} else if (!CommonType.isNull()) {
|
|
DiagText = "after resolving type aliases, the common type of '%0' "
|
|
"and '%1' is '%2'";
|
|
ExplicitlyPrintCommonType = true;
|
|
}
|
|
|
|
auto Diag =
|
|
diag(LVar->getOuterLocStart(), DiagText, DiagnosticIDs::Note)
|
|
<< LTypeStr << RTypeStr;
|
|
if (ExplicitlyPrintCommonType)
|
|
Diag << CommonTypeStr;
|
|
}
|
|
|
|
if ((hasFlag(M.flags(), MixFlags::ReferenceBind) ||
|
|
hasFlag(M.flags(), MixFlags::Qualifiers)) &&
|
|
UniqueBindPower({LType, RType})) {
|
|
StringRef DiagText = "'%0' and '%1' parameters accept and bind the "
|
|
"same kind of values";
|
|
diag(RVar->getOuterLocStart(), DiagText, DiagnosticIDs::Note)
|
|
<< LTypeStr << RTypeStr;
|
|
}
|
|
|
|
if (needsToElaborateImplicitConversion(M) &&
|
|
UniqueImplicitConversion({LType, RType})) {
|
|
const model::ConversionSequence <R =
|
|
M.leftToRightConversionSequence();
|
|
const model::ConversionSequence &RTL =
|
|
M.rightToLeftConversionSequence();
|
|
FormattedConversionSequence LTRFmt{PP, LTypeStr, LTR, RTypeStr};
|
|
FormattedConversionSequence RTLFmt{PP, RTypeStr, RTL, LTypeStr};
|
|
|
|
StringRef DiagText = "'%0' and '%1' may be implicitly converted";
|
|
if (!LTRFmt.Trivial || !RTLFmt.Trivial)
|
|
DiagText = "'%0' and '%1' may be implicitly converted: %2, %3";
|
|
|
|
{
|
|
auto Diag =
|
|
diag(RVar->getOuterLocStart(), DiagText, DiagnosticIDs::Note)
|
|
<< LTypeStr << RTypeStr;
|
|
|
|
if (!LTRFmt.Trivial || !RTLFmt.Trivial)
|
|
Diag << LTRFmt.DiagnosticText << RTLFmt.DiagnosticText;
|
|
}
|
|
|
|
StringRef ConversionFunctionDiagText =
|
|
"the implicit conversion involves the "
|
|
"%select{|converting constructor|conversion operator}0 "
|
|
"declared here";
|
|
if (const FunctionDecl *LFD = LTR.getUserDefinedConversionFunction())
|
|
diag(LFD->getLocation(), ConversionFunctionDiagText,
|
|
DiagnosticIDs::Note)
|
|
<< static_cast<unsigned>(LTR.UDConvKind)
|
|
<< LTR.getUserDefinedConversionHighlight();
|
|
if (const FunctionDecl *RFD = RTL.getUserDefinedConversionFunction())
|
|
diag(RFD->getLocation(), ConversionFunctionDiagText,
|
|
DiagnosticIDs::Note)
|
|
<< static_cast<unsigned>(RTL.UDConvKind)
|
|
<< RTL.getUserDefinedConversionHighlight();
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
} // namespace clang::tidy::bugprone
|