1 //===- JSONCompilationDatabase.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 //  This file contains the implementation of the JSONCompilationDatabase.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #include "clang/Tooling/JSONCompilationDatabase.h"
14 #include "clang/Basic/LLVM.h"
15 #include "clang/Tooling/CompilationDatabase.h"
16 #include "clang/Tooling/CompilationDatabasePluginRegistry.h"
17 #include "clang/Tooling/Tooling.h"
18 #include "llvm/ADT/Optional.h"
19 #include "llvm/ADT/STLExtras.h"
20 #include "llvm/ADT/SmallString.h"
21 #include "llvm/ADT/SmallVector.h"
22 #include "llvm/ADT/StringRef.h"
23 #include "llvm/ADT/Triple.h"
24 #include "llvm/Support/Allocator.h"
25 #include "llvm/Support/Casting.h"
26 #include "llvm/Support/CommandLine.h"
27 #include "llvm/Support/ErrorOr.h"
28 #include "llvm/Support/Host.h"
29 #include "llvm/Support/MemoryBuffer.h"
30 #include "llvm/Support/Path.h"
31 #include "llvm/Support/StringSaver.h"
32 #include "llvm/Support/VirtualFileSystem.h"
33 #include "llvm/Support/YAMLParser.h"
34 #include "llvm/Support/raw_ostream.h"
35 #include <cassert>
36 #include <memory>
37 #include <string>
38 #include <system_error>
39 #include <tuple>
40 #include <utility>
41 #include <vector>
42 
43 using namespace clang;
44 using namespace tooling;
45 
46 namespace {
47 
48 /// A parser for escaped strings of command line arguments.
49 ///
50 /// Assumes \-escaping for quoted arguments (see the documentation of
51 /// unescapeCommandLine(...)).
52 class CommandLineArgumentParser {
53  public:
54   CommandLineArgumentParser(StringRef CommandLine)
55       : Input(CommandLine), Position(Input.begin()-1) {}
56 
57   std::vector<std::string> parse() {
58     bool HasMoreInput = true;
59     while (HasMoreInput && nextNonWhitespace()) {
60       std::string Argument;
61       HasMoreInput = parseStringInto(Argument);
62       CommandLine.push_back(Argument);
63     }
64     return CommandLine;
65   }
66 
67  private:
68   // All private methods return true if there is more input available.
69 
70   bool parseStringInto(std::string &String) {
71     do {
72       if (*Position == '"') {
73         if (!parseDoubleQuotedStringInto(String)) return false;
74       } else if (*Position == '\'') {
75         if (!parseSingleQuotedStringInto(String)) return false;
76       } else {
77         if (!parseFreeStringInto(String)) return false;
78       }
79     } while (*Position != ' ');
80     return true;
81   }
82 
83   bool parseDoubleQuotedStringInto(std::string &String) {
84     if (!next()) return false;
85     while (*Position != '"') {
86       if (!skipEscapeCharacter()) return false;
87       String.push_back(*Position);
88       if (!next()) return false;
89     }
90     return next();
91   }
92 
93   bool parseSingleQuotedStringInto(std::string &String) {
94     if (!next()) return false;
95     while (*Position != '\'') {
96       String.push_back(*Position);
97       if (!next()) return false;
98     }
99     return next();
100   }
101 
102   bool parseFreeStringInto(std::string &String) {
103     do {
104       if (!skipEscapeCharacter()) return false;
105       String.push_back(*Position);
106       if (!next()) return false;
107     } while (*Position != ' ' && *Position != '"' && *Position != '\'');
108     return true;
109   }
110 
111   bool skipEscapeCharacter() {
112     if (*Position == '\\') {
113       return next();
114     }
115     return true;
116   }
117 
118   bool nextNonWhitespace() {
119     do {
120       if (!next()) return false;
121     } while (*Position == ' ');
122     return true;
123   }
124 
125   bool next() {
126     ++Position;
127     return Position != Input.end();
128   }
129 
130   const StringRef Input;
131   StringRef::iterator Position;
132   std::vector<std::string> CommandLine;
133 };
134 
135 std::vector<std::string> unescapeCommandLine(JSONCommandLineSyntax Syntax,
136                                              StringRef EscapedCommandLine) {
137   if (Syntax == JSONCommandLineSyntax::AutoDetect) {
138     Syntax = JSONCommandLineSyntax::Gnu;
139     llvm::Triple Triple(llvm::sys::getProcessTriple());
140     if (Triple.getOS() == llvm::Triple::OSType::Win32) {
141       // Assume Windows command line parsing on Win32 unless the triple
142       // explicitly tells us otherwise.
143       if (!Triple.hasEnvironment() ||
144           Triple.getEnvironment() == llvm::Triple::EnvironmentType::MSVC)
145         Syntax = JSONCommandLineSyntax::Windows;
146     }
147   }
148 
149   if (Syntax == JSONCommandLineSyntax::Windows) {
150     llvm::BumpPtrAllocator Alloc;
151     llvm::StringSaver Saver(Alloc);
152     llvm::SmallVector<const char *, 64> T;
153     llvm::cl::TokenizeWindowsCommandLine(EscapedCommandLine, Saver, T);
154     std::vector<std::string> Result(T.begin(), T.end());
155     return Result;
156   }
157   assert(Syntax == JSONCommandLineSyntax::Gnu);
158   CommandLineArgumentParser parser(EscapedCommandLine);
159   return parser.parse();
160 }
161 
162 // This plugin locates a nearby compile_command.json file, and also infers
163 // compile commands for files not present in the database.
164 class JSONCompilationDatabasePlugin : public CompilationDatabasePlugin {
165   std::unique_ptr<CompilationDatabase>
166   loadFromDirectory(StringRef Directory, std::string &ErrorMessage) override {
167     SmallString<1024> JSONDatabasePath(Directory);
168     llvm::sys::path::append(JSONDatabasePath, "compile_commands.json");
169     auto Base = JSONCompilationDatabase::loadFromFile(
170         JSONDatabasePath, ErrorMessage, JSONCommandLineSyntax::AutoDetect);
171     return Base ? inferTargetAndDriverMode(
172                       inferMissingCompileCommands(expandResponseFiles(
173                           std::move(Base), llvm::vfs::getRealFileSystem())))
174                 : nullptr;
175   }
176 };
177 
178 } // namespace
179 
180 // Register the JSONCompilationDatabasePlugin with the
181 // CompilationDatabasePluginRegistry using this statically initialized variable.
182 static CompilationDatabasePluginRegistry::Add<JSONCompilationDatabasePlugin>
183 X("json-compilation-database", "Reads JSON formatted compilation databases");
184 
185 namespace clang {
186 namespace tooling {
187 
188 // This anchor is used to force the linker to link in the generated object file
189 // and thus register the JSONCompilationDatabasePlugin.
190 volatile int JSONAnchorSource = 0;
191 
192 } // namespace tooling
193 } // namespace clang
194 
195 std::unique_ptr<JSONCompilationDatabase>
196 JSONCompilationDatabase::loadFromFile(StringRef FilePath,
197                                       std::string &ErrorMessage,
198                                       JSONCommandLineSyntax Syntax) {
199   // Don't mmap: if we're a long-lived process, the build system may overwrite.
200   llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> DatabaseBuffer =
201       llvm::MemoryBuffer::getFile(FilePath, /*FileSize=*/-1,
202                                   /*RequiresNullTerminator=*/true,
203                                   /*IsVolatile=*/true);
204   if (std::error_code Result = DatabaseBuffer.getError()) {
205     ErrorMessage = "Error while opening JSON database: " + Result.message();
206     return nullptr;
207   }
208   std::unique_ptr<JSONCompilationDatabase> Database(
209       new JSONCompilationDatabase(std::move(*DatabaseBuffer), Syntax));
210   if (!Database->parse(ErrorMessage))
211     return nullptr;
212   return Database;
213 }
214 
215 std::unique_ptr<JSONCompilationDatabase>
216 JSONCompilationDatabase::loadFromBuffer(StringRef DatabaseString,
217                                         std::string &ErrorMessage,
218                                         JSONCommandLineSyntax Syntax) {
219   std::unique_ptr<llvm::MemoryBuffer> DatabaseBuffer(
220       llvm::MemoryBuffer::getMemBuffer(DatabaseString));
221   std::unique_ptr<JSONCompilationDatabase> Database(
222       new JSONCompilationDatabase(std::move(DatabaseBuffer), Syntax));
223   if (!Database->parse(ErrorMessage))
224     return nullptr;
225   return Database;
226 }
227 
228 std::vector<CompileCommand>
229 JSONCompilationDatabase::getCompileCommands(StringRef FilePath) const {
230   SmallString<128> NativeFilePath;
231   llvm::sys::path::native(FilePath, NativeFilePath);
232 
233   std::string Error;
234   llvm::raw_string_ostream ES(Error);
235   StringRef Match = MatchTrie.findEquivalent(NativeFilePath, ES);
236   if (Match.empty())
237     return {};
238   const auto CommandsRefI = IndexByFile.find(Match);
239   if (CommandsRefI == IndexByFile.end())
240     return {};
241   std::vector<CompileCommand> Commands;
242   getCommands(CommandsRefI->getValue(), Commands);
243   return Commands;
244 }
245 
246 std::vector<std::string>
247 JSONCompilationDatabase::getAllFiles() const {
248   std::vector<std::string> Result;
249   for (const auto &CommandRef : IndexByFile)
250     Result.push_back(CommandRef.first().str());
251   return Result;
252 }
253 
254 std::vector<CompileCommand>
255 JSONCompilationDatabase::getAllCompileCommands() const {
256   std::vector<CompileCommand> Commands;
257   getCommands(AllCommands, Commands);
258   return Commands;
259 }
260 
261 static llvm::StringRef stripExecutableExtension(llvm::StringRef Name) {
262   Name.consume_back(".exe");
263   return Name;
264 }
265 
266 // There are compiler-wrappers (ccache, distcc, gomacc) that take the "real"
267 // compiler as an argument, e.g. distcc gcc -O3 foo.c.
268 // These end up in compile_commands.json when people set CC="distcc gcc".
269 // Clang's driver doesn't understand this, so we need to unwrap.
270 static bool unwrapCommand(std::vector<std::string> &Args) {
271   if (Args.size() < 2)
272     return false;
273   StringRef Wrapper =
274       stripExecutableExtension(llvm::sys::path::filename(Args.front()));
275   if (Wrapper == "distcc" || Wrapper == "gomacc" || Wrapper == "ccache") {
276     // Most of these wrappers support being invoked 3 ways:
277     // `distcc g++ file.c` This is the mode we're trying to match.
278     //                     We need to drop `distcc`.
279     // `distcc file.c`     This acts like compiler is cc or similar.
280     //                     Clang's driver can handle this, no change needed.
281     // `g++ file.c`        g++ is a symlink to distcc.
282     //                     We don't even notice this case, and all is well.
283     //
284     // We need to distinguish between the first and second case.
285     // The wrappers themselves don't take flags, so Args[1] is a compiler flag,
286     // an input file, or a compiler. Inputs have extensions, compilers don't.
287     bool HasCompiler =
288         (Args[1][0] != '-') &&
289         !llvm::sys::path::has_extension(stripExecutableExtension(Args[1]));
290     if (HasCompiler) {
291       Args.erase(Args.begin());
292       return true;
293     }
294     // If !HasCompiler, wrappers act like GCC. Fine: so do we.
295   }
296   return false;
297 }
298 
299 static std::vector<std::string>
300 nodeToCommandLine(JSONCommandLineSyntax Syntax,
301                   const std::vector<llvm::yaml::ScalarNode *> &Nodes) {
302   SmallString<1024> Storage;
303   std::vector<std::string> Arguments;
304   if (Nodes.size() == 1)
305     Arguments = unescapeCommandLine(Syntax, Nodes[0]->getValue(Storage));
306   else
307     for (const auto *Node : Nodes)
308       Arguments.push_back(std::string(Node->getValue(Storage)));
309   // There may be multiple wrappers: using distcc and ccache together is common.
310   while (unwrapCommand(Arguments))
311     ;
312   return Arguments;
313 }
314 
315 void JSONCompilationDatabase::getCommands(
316     ArrayRef<CompileCommandRef> CommandsRef,
317     std::vector<CompileCommand> &Commands) const {
318   for (const auto &CommandRef : CommandsRef) {
319     SmallString<8> DirectoryStorage;
320     SmallString<32> FilenameStorage;
321     SmallString<32> OutputStorage;
322     auto Output = std::get<3>(CommandRef);
323     Commands.emplace_back(
324         std::get<0>(CommandRef)->getValue(DirectoryStorage),
325         std::get<1>(CommandRef)->getValue(FilenameStorage),
326         nodeToCommandLine(Syntax, std::get<2>(CommandRef)),
327         Output ? Output->getValue(OutputStorage) : "");
328   }
329 }
330 
331 bool JSONCompilationDatabase::parse(std::string &ErrorMessage) {
332   llvm::yaml::document_iterator I = YAMLStream.begin();
333   if (I == YAMLStream.end()) {
334     ErrorMessage = "Error while parsing YAML.";
335     return false;
336   }
337   llvm::yaml::Node *Root = I->getRoot();
338   if (!Root) {
339     ErrorMessage = "Error while parsing YAML.";
340     return false;
341   }
342   auto *Array = dyn_cast<llvm::yaml::SequenceNode>(Root);
343   if (!Array) {
344     ErrorMessage = "Expected array.";
345     return false;
346   }
347   for (auto &NextObject : *Array) {
348     auto *Object = dyn_cast<llvm::yaml::MappingNode>(&NextObject);
349     if (!Object) {
350       ErrorMessage = "Expected object.";
351       return false;
352     }
353     llvm::yaml::ScalarNode *Directory = nullptr;
354     llvm::Optional<std::vector<llvm::yaml::ScalarNode *>> Command;
355     llvm::yaml::ScalarNode *File = nullptr;
356     llvm::yaml::ScalarNode *Output = nullptr;
357     for (auto& NextKeyValue : *Object) {
358       auto *KeyString = dyn_cast<llvm::yaml::ScalarNode>(NextKeyValue.getKey());
359       if (!KeyString) {
360         ErrorMessage = "Expected strings as key.";
361         return false;
362       }
363       SmallString<10> KeyStorage;
364       StringRef KeyValue = KeyString->getValue(KeyStorage);
365       llvm::yaml::Node *Value = NextKeyValue.getValue();
366       if (!Value) {
367         ErrorMessage = "Expected value.";
368         return false;
369       }
370       auto *ValueString = dyn_cast<llvm::yaml::ScalarNode>(Value);
371       auto *SequenceString = dyn_cast<llvm::yaml::SequenceNode>(Value);
372       if (KeyValue == "arguments" && !SequenceString) {
373         ErrorMessage = "Expected sequence as value.";
374         return false;
375       } else if (KeyValue != "arguments" && !ValueString) {
376         ErrorMessage = "Expected string as value.";
377         return false;
378       }
379       if (KeyValue == "directory") {
380         Directory = ValueString;
381       } else if (KeyValue == "arguments") {
382         Command = std::vector<llvm::yaml::ScalarNode *>();
383         for (auto &Argument : *SequenceString) {
384           auto *Scalar = dyn_cast<llvm::yaml::ScalarNode>(&Argument);
385           if (!Scalar) {
386             ErrorMessage = "Only strings are allowed in 'arguments'.";
387             return false;
388           }
389           Command->push_back(Scalar);
390         }
391       } else if (KeyValue == "command") {
392         if (!Command)
393           Command = std::vector<llvm::yaml::ScalarNode *>(1, ValueString);
394       } else if (KeyValue == "file") {
395         File = ValueString;
396       } else if (KeyValue == "output") {
397         Output = ValueString;
398       } else {
399         ErrorMessage = ("Unknown key: \"" +
400                         KeyString->getRawValue() + "\"").str();
401         return false;
402       }
403     }
404     if (!File) {
405       ErrorMessage = "Missing key: \"file\".";
406       return false;
407     }
408     if (!Command) {
409       ErrorMessage = "Missing key: \"command\" or \"arguments\".";
410       return false;
411     }
412     if (!Directory) {
413       ErrorMessage = "Missing key: \"directory\".";
414       return false;
415     }
416     SmallString<8> FileStorage;
417     StringRef FileName = File->getValue(FileStorage);
418     SmallString<128> NativeFilePath;
419     if (llvm::sys::path::is_relative(FileName)) {
420       SmallString<8> DirectoryStorage;
421       SmallString<128> AbsolutePath(
422           Directory->getValue(DirectoryStorage));
423       llvm::sys::path::append(AbsolutePath, FileName);
424       llvm::sys::path::remove_dots(AbsolutePath, /*remove_dot_dot=*/ true);
425       llvm::sys::path::native(AbsolutePath, NativeFilePath);
426     } else {
427       llvm::sys::path::native(FileName, NativeFilePath);
428     }
429     auto Cmd = CompileCommandRef(Directory, File, *Command, Output);
430     IndexByFile[NativeFilePath].push_back(Cmd);
431     AllCommands.push_back(Cmd);
432     MatchTrie.insert(NativeFilePath);
433   }
434   return true;
435 }
436