1 | //===---------- IssueHash.cpp - Generate identification hashes --*- C++ -*-===// |
2 | // |
3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
4 | // See https://llvm.org/LICENSE.txt for license information. |
5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
6 | // |
7 | //===----------------------------------------------------------------------===// |
8 | |
9 | #include "clang/Analysis/IssueHash.h" |
10 | #include "clang/AST/ASTContext.h" |
11 | #include "clang/AST/Decl.h" |
12 | #include "clang/AST/DeclCXX.h" |
13 | #include "clang/Basic/SourceManager.h" |
14 | #include "clang/Lex/Lexer.h" |
15 | #include "llvm/ADT/StringRef.h" |
16 | #include "llvm/ADT/Twine.h" |
17 | #include "llvm/Support/LineIterator.h" |
18 | #include "llvm/Support/MD5.h" |
19 | |
20 | #include <optional> |
21 | #include <sstream> |
22 | #include <string> |
23 | |
24 | using namespace clang; |
25 | |
26 | // Get a string representation of the parts of the signature that can be |
27 | // overloaded on. |
28 | static std::string GetSignature(const FunctionDecl *Target) { |
29 | if (!Target) |
30 | return "" ; |
31 | std::string Signature; |
32 | |
33 | // When a flow sensitive bug happens in templated code we should not generate |
34 | // distinct hash value for every instantiation. Use the signature from the |
35 | // primary template. |
36 | if (const FunctionDecl *InstantiatedFrom = |
37 | Target->getTemplateInstantiationPattern()) |
38 | Target = InstantiatedFrom; |
39 | |
40 | if (!isa<CXXConstructorDecl>(Val: Target) && !isa<CXXDestructorDecl>(Val: Target) && |
41 | !isa<CXXConversionDecl>(Val: Target)) |
42 | Signature.append(str: Target->getReturnType().getAsString()).append(s: " " ); |
43 | Signature.append(Target->getQualifiedNameAsString()).append("(" ); |
44 | |
45 | for (int i = 0, paramsCount = Target->getNumParams(); i < paramsCount; ++i) { |
46 | if (i) |
47 | Signature.append(s: ", " ); |
48 | Signature.append(Target->getParamDecl(i)->getType().getAsString()); |
49 | } |
50 | |
51 | if (Target->isVariadic()) |
52 | Signature.append(s: ", ..." ); |
53 | Signature.append(s: ")" ); |
54 | |
55 | const auto *TargetT = |
56 | llvm::dyn_cast_or_null<FunctionType>(Target->getType().getTypePtr()); |
57 | |
58 | if (!TargetT || !isa<CXXMethodDecl>(Val: Target)) |
59 | return Signature; |
60 | |
61 | if (TargetT->isConst()) |
62 | Signature.append(s: " const" ); |
63 | if (TargetT->isVolatile()) |
64 | Signature.append(s: " volatile" ); |
65 | if (TargetT->isRestrict()) |
66 | Signature.append(s: " restrict" ); |
67 | |
68 | if (const auto *TargetPT = |
69 | dyn_cast_or_null<FunctionProtoType>(Target->getType().getTypePtr())) { |
70 | switch (TargetPT->getRefQualifier()) { |
71 | case RQ_LValue: |
72 | Signature.append(s: " &" ); |
73 | break; |
74 | case RQ_RValue: |
75 | Signature.append(s: " &&" ); |
76 | break; |
77 | default: |
78 | break; |
79 | } |
80 | } |
81 | |
82 | return Signature; |
83 | } |
84 | |
85 | static std::string GetEnclosingDeclContextSignature(const Decl *D) { |
86 | if (!D) |
87 | return "" ; |
88 | |
89 | if (const auto *ND = dyn_cast<NamedDecl>(Val: D)) { |
90 | std::string DeclName; |
91 | |
92 | switch (ND->getKind()) { |
93 | case Decl::Namespace: |
94 | case Decl::Record: |
95 | case Decl::CXXRecord: |
96 | case Decl::Enum: |
97 | DeclName = ND->getQualifiedNameAsString(); |
98 | break; |
99 | case Decl::CXXConstructor: |
100 | case Decl::CXXDestructor: |
101 | case Decl::CXXConversion: |
102 | case Decl::CXXMethod: |
103 | case Decl::Function: |
104 | DeclName = GetSignature(Target: dyn_cast_or_null<FunctionDecl>(Val: ND)); |
105 | break; |
106 | case Decl::ObjCMethod: |
107 | // ObjC Methods can not be overloaded, qualified name uniquely identifies |
108 | // the method. |
109 | DeclName = ND->getQualifiedNameAsString(); |
110 | break; |
111 | default: |
112 | break; |
113 | } |
114 | |
115 | return DeclName; |
116 | } |
117 | |
118 | return "" ; |
119 | } |
120 | |
121 | static StringRef GetNthLineOfFile(std::optional<llvm::MemoryBufferRef> Buffer, |
122 | int Line) { |
123 | if (!Buffer) |
124 | return "" ; |
125 | |
126 | llvm::line_iterator LI(*Buffer, false); |
127 | for (; !LI.is_at_eof() && LI.line_number() != Line; ++LI) |
128 | ; |
129 | |
130 | return *LI; |
131 | } |
132 | |
133 | static std::string NormalizeLine(const SourceManager &SM, const FullSourceLoc &L, |
134 | const LangOptions &LangOpts) { |
135 | static StringRef Whitespaces = " \t\n" ; |
136 | |
137 | StringRef Str = GetNthLineOfFile(Buffer: SM.getBufferOrNone(FID: L.getFileID(), Loc: L), |
138 | Line: L.getExpansionLineNumber()); |
139 | StringRef::size_type col = Str.find_first_not_of(Chars: Whitespaces); |
140 | if (col == StringRef::npos) |
141 | col = 1; // The line only contains whitespace. |
142 | else |
143 | col++; |
144 | SourceLocation StartOfLine = |
145 | SM.translateLineCol(FID: SM.getFileID(SpellingLoc: L), Line: L.getExpansionLineNumber(), Col: col); |
146 | std::optional<llvm::MemoryBufferRef> Buffer = |
147 | SM.getBufferOrNone(FID: SM.getFileID(SpellingLoc: StartOfLine), Loc: StartOfLine); |
148 | if (!Buffer) |
149 | return {}; |
150 | |
151 | const char *BufferPos = SM.getCharacterData(SL: StartOfLine); |
152 | |
153 | Token Token; |
154 | Lexer Lexer(SM.getLocForStartOfFile(FID: SM.getFileID(SpellingLoc: StartOfLine)), LangOpts, |
155 | Buffer->getBufferStart(), BufferPos, Buffer->getBufferEnd()); |
156 | |
157 | size_t NextStart = 0; |
158 | std::ostringstream LineBuff; |
159 | while (!Lexer.LexFromRawLexer(Result&: Token) && NextStart < 2) { |
160 | if (Token.isAtStartOfLine() && NextStart++ > 0) |
161 | continue; |
162 | LineBuff << std::string(SM.getCharacterData(SL: Token.getLocation()), |
163 | Token.getLength()); |
164 | } |
165 | |
166 | return LineBuff.str(); |
167 | } |
168 | |
169 | static llvm::SmallString<32> GetMD5HashOfContent(StringRef Content) { |
170 | llvm::MD5 Hash; |
171 | llvm::MD5::MD5Result MD5Res; |
172 | SmallString<32> Res; |
173 | |
174 | Hash.update(Str: Content); |
175 | Hash.final(Result&: MD5Res); |
176 | llvm::MD5::stringifyResult(Result&: MD5Res, Str&: Res); |
177 | |
178 | return Res; |
179 | } |
180 | |
181 | std::string clang::getIssueString(const FullSourceLoc &IssueLoc, |
182 | StringRef CheckerName, |
183 | StringRef WarningMessage, |
184 | const Decl *IssueDecl, |
185 | const LangOptions &LangOpts) { |
186 | static StringRef Delimiter = "$" ; |
187 | |
188 | return (llvm::Twine(CheckerName) + Delimiter + |
189 | GetEnclosingDeclContextSignature(D: IssueDecl) + Delimiter + |
190 | Twine(IssueLoc.getExpansionColumnNumber()) + Delimiter + |
191 | NormalizeLine(SM: IssueLoc.getManager(), L: IssueLoc, LangOpts) + |
192 | Delimiter + WarningMessage) |
193 | .str(); |
194 | } |
195 | |
196 | SmallString<32> clang::getIssueHash(const FullSourceLoc &IssueLoc, |
197 | StringRef CheckerName, |
198 | StringRef WarningMessage, |
199 | const Decl *IssueDecl, |
200 | const LangOptions &LangOpts) { |
201 | |
202 | return GetMD5HashOfContent(Content: getIssueString( |
203 | IssueLoc, CheckerName, WarningMessage, IssueDecl, LangOpts)); |
204 | } |
205 | |