1 //===--- HeaderSourceSwitch.cpp - --------------------------------*- 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 "HeaderSourceSwitch.h"
10 #include "AST.h"
11 #include "SourceCode.h"
12 #include "index/SymbolCollector.h"
13 #include "support/Logger.h"
14 #include "clang/AST/Decl.h"
15 
16 namespace clang {
17 namespace clangd {
18 
19 llvm::Optional<Path> getCorrespondingHeaderOrSource(
20     PathRef OriginalFile, llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS) {
21   llvm::StringRef SourceExtensions[] = {".cpp", ".c", ".cc", ".cxx",
22                                         ".c++", ".m", ".mm"};
23   llvm::StringRef HeaderExtensions[] = {".h", ".hh", ".hpp", ".hxx", ".inc"};
24 
25   llvm::StringRef PathExt = llvm::sys::path::extension(OriginalFile);
26 
27   // Lookup in a list of known extensions.
28   auto SourceIter =
29       llvm::find_if(SourceExtensions, [&PathExt](PathRef SourceExt) {
30         return SourceExt.equals_insensitive(PathExt);
31       });
32   bool IsSource = SourceIter != std::end(SourceExtensions);
33 
34   auto HeaderIter =
35       llvm::find_if(HeaderExtensions, [&PathExt](PathRef HeaderExt) {
36         return HeaderExt.equals_insensitive(PathExt);
37       });
38   bool IsHeader = HeaderIter != std::end(HeaderExtensions);
39 
40   // We can only switch between the known extensions.
41   if (!IsSource && !IsHeader)
42     return None;
43 
44   // Array to lookup extensions for the switch. An opposite of where original
45   // extension was found.
46   llvm::ArrayRef<llvm::StringRef> NewExts;
47   if (IsSource)
48     NewExts = HeaderExtensions;
49   else
50     NewExts = SourceExtensions;
51 
52   // Storage for the new path.
53   llvm::SmallString<128> NewPath = OriginalFile;
54 
55   // Loop through switched extension candidates.
56   for (llvm::StringRef NewExt : NewExts) {
57     llvm::sys::path::replace_extension(NewPath, NewExt);
58     if (VFS->exists(NewPath))
59       return Path(NewPath);
60 
61     // Also check NewExt in upper-case, just in case.
62     llvm::sys::path::replace_extension(NewPath, NewExt.upper());
63     if (VFS->exists(NewPath))
64       return Path(NewPath);
65   }
66   return None;
67 }
68 
69 llvm::Optional<Path> getCorrespondingHeaderOrSource(PathRef OriginalFile,
70                                                     ParsedAST &AST,
71                                                     const SymbolIndex *Index) {
72   if (!Index) {
73     // FIXME: use the AST to do the inference.
74     return None;
75   }
76   LookupRequest Request;
77   // Find all symbols present in the original file.
78   for (const auto *D : getIndexableLocalDecls(AST)) {
79     if (auto ID = getSymbolID(D))
80       Request.IDs.insert(ID);
81   }
82   llvm::StringMap<int> Candidates; // Target path => score.
83   auto AwardTarget = [&](const char *TargetURI) {
84     if (auto TargetPath = URI::resolve(TargetURI, OriginalFile)) {
85       if (*TargetPath != OriginalFile) // exclude the original file.
86         ++Candidates[*TargetPath];
87     } else {
88       elog("Failed to resolve URI {0}: {1}", TargetURI, TargetPath.takeError());
89     }
90   };
91   // If we switch from a header, we are looking for the implementation
92   // file, so we use the definition loc; otherwise we look for the header file,
93   // we use the decl loc;
94   //
95   // For each symbol in the original file, we get its target location (decl or
96   // def) from the index, then award that target file.
97   bool IsHeader = isHeaderFile(OriginalFile, AST.getLangOpts());
98   Index->lookup(Request, [&](const Symbol &Sym) {
99     if (IsHeader)
100       AwardTarget(Sym.Definition.FileURI);
101     else
102       AwardTarget(Sym.CanonicalDeclaration.FileURI);
103   });
104   // FIXME: our index doesn't have any interesting information (this could be
105   // that the background-index is not finished), we should use the decl/def
106   // locations from the AST to do the inference (from .cc to .h).
107   if (Candidates.empty())
108     return None;
109 
110   // Pickup the winner, who contains most of symbols.
111   // FIXME: should we use other signals (file proximity) to help score?
112   auto Best = Candidates.begin();
113   for (auto It = Candidates.begin(); It != Candidates.end(); ++It) {
114     if (It->second > Best->second)
115       Best = It;
116     else if (It->second == Best->second && It->first() < Best->first())
117       // Select the first one in the lexical order if we have multiple
118       // candidates.
119       Best = It;
120   }
121   return Path(Best->first());
122 }
123 
124 std::vector<const Decl *> getIndexableLocalDecls(ParsedAST &AST) {
125   std::vector<const Decl *> Results;
126   std::function<void(Decl *)> TraverseDecl = [&](Decl *D) {
127     auto *ND = llvm::dyn_cast<NamedDecl>(D);
128     if (!ND || ND->isImplicit())
129       return;
130     if (!SymbolCollector::shouldCollectSymbol(*ND, D->getASTContext(), {},
131                                               /*IsMainFileSymbol=*/false))
132       return;
133     if (!llvm::isa<FunctionDecl>(ND)) {
134       // Visit the children, but we skip function decls as we are not interested
135       // in the function body.
136       if (auto *Scope = llvm::dyn_cast<DeclContext>(ND)) {
137         for (auto *D : Scope->decls())
138           TraverseDecl(D);
139       }
140     }
141     if (llvm::isa<NamespaceDecl>(D))
142       return; // namespace is indexable, but we're not interested.
143     Results.push_back(D);
144   };
145   // Traverses the ParsedAST directly to collect all decls present in the main
146   // file.
147   for (auto *TopLevel : AST.getLocalTopLevelDecls())
148     TraverseDecl(TopLevel);
149   return Results;
150 }
151 
152 } // namespace clangd
153 } // namespace clang
154