| 1 |
//===--- ConfusableIdentifierCheck.cpp - |
| 2 |
// clang-tidy--------------------------===// |
| 3 |
// |
| 4 |
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 5 |
// See https://llvm.org/LICENSE.txt for license information. |
| 6 |
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 7 |
// |
| 8 |
//===----------------------------------------------------------------------===// |
| 9 |
|
| 10 |
#include "ConfusableIdentifierCheck.h" |
| 11 |
|
| 12 |
#include "clang/Frontend/CompilerInstance.h" |
| 13 |
#include "clang/Lex/Preprocessor.h" |
| 14 |
#include "llvm/Support/ConvertUTF.h" |
| 15 |
|
| 16 |
namespace { |
| 17 |
// Preprocessed version of |
| 18 |
// https://www.unicode.org/Public/security/latest/confusables.txt |
| 19 |
// |
| 20 |
// This contains a sorted array of { UTF32 codepoint; UTF32 values[N];} |
| 21 |
#include "Confusables.inc" |
| 22 |
} // namespace |
| 23 |
|
| 24 |
namespace clang::tidy::misc { |
| 25 |
|
| 26 |
ConfusableIdentifierCheck::ConfusableIdentifierCheck(StringRef Name, |
| 27 |
ClangTidyContext *Context) |
| 28 |
: ClangTidyCheck(Name, Context) {} |
| 29 |
|
| 30 |
ConfusableIdentifierCheck::~ConfusableIdentifierCheck() = default; |
| 31 |
|
| 32 |
// Build a skeleton out of the Original identifier, inspired by the algorithm |
| 33 |
// described in http://www.unicode.org/reports/tr39/#def-skeleton |
| 34 |
// |
| 35 |
// FIXME: TR39 mandates: |
| 36 |
// |
| 37 |
// For an input string X, define skeleton(X) to be the following transformation |
| 38 |
// on the string: |
| 39 |
// |
| 40 |
// 1. Convert X to NFD format, as described in [UAX15]. |
| 41 |
// 2. Concatenate the prototypes for each character in X according to the |
| 42 |
// specified data, producing a string of exemplar characters. |
| 43 |
// 3. Reapply NFD. |
| 44 |
// |
| 45 |
// We're skipping 1. and 3. for the sake of simplicity, but this can lead to |
| 46 |
// false positive. |
| 47 |
|
| 48 |
std::string ConfusableIdentifierCheck::skeleton(StringRef Name) { |
| 49 |
using namespace llvm; |
| 50 |
std::string SName = Name.str(); |
| 51 |
std::string Skeleton; |
| 52 |
Skeleton.reserve(1 + Name.size()); |
| 53 |
|
| 54 |
const char *Curr = SName.c_str(); |
| 55 |
const char *End = Curr + SName.size(); |
| 56 |
while (Curr < End) { |
| 57 |
|
| 58 |
const char *Prev = Curr; |
| 59 |
UTF32 CodePoint; |
| 60 |
ConversionResult Result = convertUTF8Sequence( |
| 61 |
reinterpret_cast<const UTF8 **>(&Curr), |
| 62 |
reinterpret_cast<const UTF8 *>(End), &CodePoint, strictConversion); |
| 63 |
if (Result != conversionOK) { |
| 64 |
errs() << "Unicode conversion issue\n"; |
| 65 |
break; |
| 66 |
} |
| 67 |
|
| 68 |
StringRef Key(Prev, Curr - Prev); |
| 69 |
auto Where = llvm::lower_bound(ConfusableEntries, CodePoint, |
| 70 |
[](decltype(ConfusableEntries[0]) x, |
| 71 |
UTF32 y) { return x.codepoint < y; }); |
| 72 |
if (Where == std::end(ConfusableEntries) || CodePoint != Where->codepoint) { |
| 73 |
Skeleton.append(Prev, Curr); |
| 74 |
} else { |
| 75 |
UTF8 Buffer[32]; |
| 76 |
UTF8 *BufferStart = std::begin(Buffer); |
| 77 |
UTF8 *IBuffer = BufferStart; |
| 78 |
const UTF32 *ValuesStart = std::begin(Where->values); |
| 79 |
const UTF32 *ValuesEnd = llvm::find(Where->values, '\0'); |
| 80 |
if (ConvertUTF32toUTF8(&ValuesStart, ValuesEnd, &IBuffer, |
| 81 |
std::end(Buffer), |
| 82 |
strictConversion) != conversionOK) { |
| 83 |
errs() << "Unicode conversion issue\n"; |
| 84 |
break; |
| 85 |
} |
| 86 |
Skeleton.append((char *)BufferStart, (char *)IBuffer); |
| 87 |
} |
| 88 |
} |
| 89 |
return Skeleton; |
| 90 |
} |
| 91 |
|
| 92 |
static bool mayShadowImpl(const NamedDecl *ND0, const NamedDecl *ND1) { |
| 93 |
const DeclContext *DC0 = ND0->getDeclContext()->getPrimaryContext(); |
| 94 |
const DeclContext *DC1 = ND1->getDeclContext()->getPrimaryContext(); |
| 95 |
|
| 96 |
if (isa<TemplateTypeParmDecl>(ND0) || isa<TemplateTypeParmDecl>(ND0)) |
| 97 |
return true; |
| 98 |
|
| 99 |
while (DC0->isTransparentContext()) |
| 100 |
DC0 = DC0->getParent(); |
| 101 |
while (DC1->isTransparentContext()) |
| 102 |
DC1 = DC1->getParent(); |
| 103 |
|
| 104 |
if (DC0->Equals(DC1)) |
| 105 |
return true; |
| 106 |
|
| 107 |
return false; |
| 108 |
} |
| 109 |
|
| 110 |
static bool isMemberOf(const NamedDecl *ND, const CXXRecordDecl *RD) { |
| 111 |
const DeclContext *NDParent = ND->getDeclContext(); |
| 112 |
if (!NDParent || !isa<CXXRecordDecl>(NDParent)) |
| 113 |
return false; |
| 114 |
if (NDParent == RD) |
| 115 |
return true; |
| 116 |
return !RD->forallBases( |
| 117 |
[NDParent](const CXXRecordDecl *Base) { return NDParent != Base; }); |
| 118 |
} |
| 119 |
|
| 120 |
static bool mayShadow(const NamedDecl *ND0, const NamedDecl *ND1) { |
| 121 |
|
| 122 |
const DeclContext *DC0 = ND0->getDeclContext()->getPrimaryContext(); |
| 123 |
const DeclContext *DC1 = ND1->getDeclContext()->getPrimaryContext(); |
| 124 |
|
| 125 |
if (const CXXRecordDecl *RD0 = dyn_cast<CXXRecordDecl>(DC0)) { |
| 126 |
RD0 = RD0->getDefinition(); |
| 127 |
if (RD0 && ND1->getAccess() != AS_private && isMemberOf(ND1, RD0)) |
| 128 |
return true; |
| 129 |
} |
| 130 |
if (const CXXRecordDecl *RD1 = dyn_cast<CXXRecordDecl>(DC1)) { |
| 131 |
RD1 = RD1->getDefinition(); |
| 132 |
if (RD1 && ND0->getAccess() != AS_private && isMemberOf(ND0, RD1)) |
| 133 |
return true; |
| 134 |
} |
| 135 |
|
| 136 |
if (DC0->Encloses(DC1)) |
| 137 |
return mayShadowImpl(ND0, ND1); |
| 138 |
if (DC1->Encloses(DC0)) |
| 139 |
return mayShadowImpl(ND1, ND0); |
| 140 |
return false; |
| 141 |
} |
| 142 |
|
| 143 |
void ConfusableIdentifierCheck::check( |
| 144 |
const ast_matchers::MatchFinder::MatchResult &Result) { |
| 145 |
if (const auto *ND = Result.Nodes.getNodeAs<NamedDecl>("nameddecl")) { |
| 146 |
if (IdentifierInfo *NDII = ND->getIdentifier()) { |
| 147 |
StringRef NDName = NDII->getName(); |
| 148 |
llvm::SmallVector<const NamedDecl *> &Mapped = Mapper[skeleton(NDName)]; |
| 149 |
for (const NamedDecl *OND : Mapped) { |
| 150 |
const IdentifierInfo *ONDII = OND->getIdentifier(); |
| 151 |
if (mayShadow(ND, OND)) { |
| 152 |
StringRef ONDName = ONDII->getName(); |
| 153 |
if (ONDName != NDName) { |
| 154 |
diag(ND->getLocation(), "%0 is confusable with %1") << ND << OND; |
| 155 |
diag(OND->getLocation(), "other declaration found here", |
| 156 |
DiagnosticIDs::Note); |
| 157 |
} |
| 158 |
} |
| 159 |
} |
| 160 |
Mapped.push_back(ND); |
| 161 |
} |
| 162 |
} |
| 163 |
} |
| 164 |
|
| 165 |
void ConfusableIdentifierCheck::registerMatchers( |
| 166 |
ast_matchers::MatchFinder *Finder) { |
| 167 |
Finder->addMatcher(ast_matchers::namedDecl().bind("nameddecl"), this); |
| 168 |
} |
| 169 |
|
| 170 |
} // namespace clang::tidy::misc |