1 | //===--- Analysis.cpp -----------------------------------------------------===// |
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-include-cleaner/Analysis.h" |
10 | #include "AnalysisInternal.h" |
11 | #include "clang-include-cleaner/IncludeSpeller.h" |
12 | #include "clang-include-cleaner/Record.h" |
13 | #include "clang-include-cleaner/Types.h" |
14 | #include "clang/AST/Decl.h" |
15 | #include "clang/AST/DeclBase.h" |
16 | #include "clang/Basic/DirectoryEntry.h" |
17 | #include "clang/Basic/FileEntry.h" |
18 | #include "clang/Basic/SourceManager.h" |
19 | #include "clang/Format/Format.h" |
20 | #include "clang/Lex/HeaderSearch.h" |
21 | #include "clang/Lex/Preprocessor.h" |
22 | #include "clang/Tooling/Core/Replacement.h" |
23 | #include "clang/Tooling/Inclusions/StandardLibrary.h" |
24 | #include "llvm/ADT/ArrayRef.h" |
25 | #include "llvm/ADT/DenseSet.h" |
26 | #include "llvm/ADT/STLExtras.h" |
27 | #include "llvm/ADT/STLFunctionalExtras.h" |
28 | #include "llvm/ADT/SmallVector.h" |
29 | #include "llvm/ADT/StringMap.h" |
30 | #include "llvm/ADT/StringRef.h" |
31 | #include "llvm/Support/Error.h" |
32 | #include "llvm/Support/ErrorHandling.h" |
33 | #include <cassert> |
34 | #include <climits> |
35 | #include <string> |
36 | |
37 | namespace clang::include_cleaner { |
38 | |
39 | namespace { |
40 | bool shouldIgnoreMacroReference(const Preprocessor &PP, const Macro &M) { |
41 | auto *MI = PP.getMacroInfo(II: M.Name); |
42 | // Macros that expand to themselves are confusing from user's point of view. |
43 | // They usually aspect the usage to be attributed to the underlying decl and |
44 | // not the macro definition. So ignore such macros (e.g. std{in,out,err} are |
45 | // implementation defined macros, that just resolve to themselves in |
46 | // practice). |
47 | return MI && MI->getNumTokens() == 1 && MI->isObjectLike() && |
48 | MI->getReplacementToken(Tok: 0).getIdentifierInfo() == M.Name; |
49 | } |
50 | } // namespace |
51 | |
52 | void walkUsed(llvm::ArrayRef<Decl *> ASTRoots, |
53 | llvm::ArrayRef<SymbolReference> MacroRefs, |
54 | const PragmaIncludes *PI, const Preprocessor &PP, |
55 | UsedSymbolCB CB) { |
56 | const auto &SM = PP.getSourceManager(); |
57 | // This is duplicated in writeHTMLReport, changes should be mirrored there. |
58 | tooling::stdlib::Recognizer Recognizer; |
59 | for (auto *Root : ASTRoots) { |
60 | walkAST(*Root, [&](SourceLocation Loc, NamedDecl &ND, RefType RT) { |
61 | auto FID = SM.getFileID(SpellingLoc: SM.getSpellingLoc(Loc)); |
62 | if (FID != SM.getMainFileID() && FID != SM.getPreambleFileID()) |
63 | return; |
64 | // FIXME: Most of the work done here is repetitive. It might be useful to |
65 | // have a cache/batching. |
66 | SymbolReference SymRef{ND, .RefLocation: Loc, .RT: RT}; |
67 | return CB(SymRef, headersForSymbol(ND, PP, PI)); |
68 | }); |
69 | } |
70 | for (const SymbolReference &MacroRef : MacroRefs) { |
71 | assert(MacroRef.Target.kind() == Symbol::Macro); |
72 | if (!SM.isWrittenInMainFile(Loc: SM.getSpellingLoc(Loc: MacroRef.RefLocation)) || |
73 | shouldIgnoreMacroReference(PP, M: MacroRef.Target.macro())) |
74 | continue; |
75 | CB(MacroRef, headersForSymbol(S: MacroRef.Target, PP, PI)); |
76 | } |
77 | } |
78 | |
79 | AnalysisResults |
80 | analyze(llvm::ArrayRef<Decl *> ASTRoots, |
81 | llvm::ArrayRef<SymbolReference> MacroRefs, const Includes &Inc, |
82 | const PragmaIncludes *PI, const Preprocessor &PP, |
83 | llvm::function_ref<bool(llvm::StringRef)> ) { |
84 | auto &SM = PP.getSourceManager(); |
85 | const auto MainFile = *SM.getFileEntryRefForID(FID: SM.getMainFileID()); |
86 | llvm::DenseSet<const Include *> Used; |
87 | llvm::StringMap<Header> Missing; |
88 | constexpr auto = [](llvm::StringRef) { return false; }; |
89 | if (!HeaderFilter) |
90 | HeaderFilter = DefaultHeaderFilter; |
91 | OptionalDirectoryEntryRef ResourceDir = |
92 | PP.getHeaderSearchInfo().getModuleMap().getBuiltinDir(); |
93 | walkUsed(ASTRoots, MacroRefs, PI, PP, |
94 | CB: [&](const SymbolReference &Ref, llvm::ArrayRef<Header> Providers) { |
95 | bool Satisfied = false; |
96 | for (const Header &H : Providers) { |
97 | if (H.kind() == Header::Physical && |
98 | (H.physical() == MainFile || |
99 | H.physical().getDir() == ResourceDir)) { |
100 | Satisfied = true; |
101 | } |
102 | for (const Include *I : Inc.match(H)) { |
103 | Used.insert(V: I); |
104 | Satisfied = true; |
105 | } |
106 | } |
107 | // Bail out if we can't (or need not) insert an include. |
108 | if (Satisfied || Providers.empty() || Ref.RT != RefType::Explicit) |
109 | return; |
110 | if (HeaderFilter(Providers.front().resolvedPath())) |
111 | return; |
112 | // Check if we have any headers with the same spelling, in edge |
113 | // cases like `#include_next "foo.h"`, the user can't ever |
114 | // include the physical foo.h, but can have a spelling that |
115 | // refers to it. |
116 | auto Spelling = spellHeader( |
117 | Input: {.H: Providers.front(), .HS: PP.getHeaderSearchInfo(), .Main: MainFile}); |
118 | for (const Include *I : Inc.match(H: Header{Spelling})) { |
119 | Used.insert(V: I); |
120 | Satisfied = true; |
121 | } |
122 | if (!Satisfied) |
123 | Missing.try_emplace(Key: std::move(Spelling), Args: Providers.front()); |
124 | }); |
125 | |
126 | AnalysisResults Results; |
127 | for (const Include &I : Inc.all()) { |
128 | if (Used.contains(V: &I) || !I.Resolved || |
129 | HeaderFilter(I.Resolved->getName()) || |
130 | I.Resolved->getDir() == ResourceDir) |
131 | continue; |
132 | if (PI) { |
133 | if (PI->shouldKeep(FE: *I.Resolved)) |
134 | continue; |
135 | // Check if main file is the public interface for a private header. If so |
136 | // we shouldn't diagnose it as unused. |
137 | if (auto = PI->getPublic(File: *I.Resolved); !PHeader.empty()) { |
138 | PHeader = PHeader.trim(Chars: "<>\"" ); |
139 | // Since most private -> public mappings happen in a verbatim way, we |
140 | // check textually here. This might go wrong in presence of symlinks or |
141 | // header mappings. But that's not different than rest of the places. |
142 | if (MainFile.getName().ends_with(Suffix: PHeader)) |
143 | continue; |
144 | } |
145 | } |
146 | Results.Unused.push_back(x: &I); |
147 | } |
148 | for (auto &E : Missing) |
149 | Results.Missing.emplace_back(args: E.first().str(), args&: E.second); |
150 | llvm::sort(C&: Results.Missing); |
151 | return Results; |
152 | } |
153 | |
154 | std::string fixIncludes(const AnalysisResults &Results, |
155 | llvm::StringRef FileName, llvm::StringRef Code, |
156 | const format::FormatStyle &Style) { |
157 | assert(Style.isCpp() && "Only C++ style supports include insertions!" ); |
158 | tooling::Replacements R; |
159 | // Encode insertions/deletions in the magic way clang-format understands. |
160 | for (const Include *I : Results.Unused) |
161 | cantFail(Err: R.add(R: tooling::Replacement(FileName, UINT_MAX, 1, I->quote()))); |
162 | for (auto &[Spelled, _] : Results.Missing) |
163 | cantFail(Err: R.add( |
164 | R: tooling::Replacement(FileName, UINT_MAX, 0, "#include " + Spelled))); |
165 | // "cleanup" actually turns the UINT_MAX replacements into concrete edits. |
166 | auto Positioned = cantFail(ValOrErr: format::cleanupAroundReplacements(Code, Replaces: R, Style)); |
167 | return cantFail(ValOrErr: tooling::applyAllReplacements(Code, Replaces: Positioned)); |
168 | } |
169 | |
170 | } // namespace clang::include_cleaner |
171 | |