1 //===--- CompilationDatabase.cpp - ----------------------------------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 //  This file contains implementations of the CompilationDatabase base class
11 //  and the FixedCompilationDatabase.
12 //
13 //  FIXME: Various functions that take a string &ErrorMessage should be upgraded
14 //  to Expected.
15 //
16 //===----------------------------------------------------------------------===//
17 
18 #include "clang/Tooling/CompilationDatabase.h"
19 #include "clang/Basic/Diagnostic.h"
20 #include "clang/Basic/DiagnosticOptions.h"
21 #include "clang/Driver/Action.h"
22 #include "clang/Driver/Compilation.h"
23 #include "clang/Driver/Driver.h"
24 #include "clang/Driver/DriverDiagnostic.h"
25 #include "clang/Driver/Job.h"
26 #include "clang/Frontend/TextDiagnosticPrinter.h"
27 #include "clang/Tooling/CompilationDatabasePluginRegistry.h"
28 #include "clang/Tooling/Tooling.h"
29 #include "llvm/ADT/SmallString.h"
30 #include "llvm/Option/Arg.h"
31 #include "llvm/Support/Host.h"
32 #include "llvm/Support/LineIterator.h"
33 #include "llvm/Support/Path.h"
34 #include "llvm/Support/raw_ostream.h"
35 #include <sstream>
36 #include <system_error>
37 using namespace clang;
38 using namespace tooling;
39 
40 LLVM_INSTANTIATE_REGISTRY(CompilationDatabasePluginRegistry)
41 
42 CompilationDatabase::~CompilationDatabase() {}
43 
44 std::unique_ptr<CompilationDatabase>
45 CompilationDatabase::loadFromDirectory(StringRef BuildDirectory,
46                                        std::string &ErrorMessage) {
47   llvm::raw_string_ostream ErrorStream(ErrorMessage);
48   for (CompilationDatabasePluginRegistry::iterator
49        It = CompilationDatabasePluginRegistry::begin(),
50        Ie = CompilationDatabasePluginRegistry::end();
51        It != Ie; ++It) {
52     std::string DatabaseErrorMessage;
53     std::unique_ptr<CompilationDatabasePlugin> Plugin(It->instantiate());
54     if (std::unique_ptr<CompilationDatabase> DB =
55             Plugin->loadFromDirectory(BuildDirectory, DatabaseErrorMessage))
56       return DB;
57     ErrorStream << It->getName() << ": " << DatabaseErrorMessage << "\n";
58   }
59   return nullptr;
60 }
61 
62 static std::unique_ptr<CompilationDatabase>
63 findCompilationDatabaseFromDirectory(StringRef Directory,
64                                      std::string &ErrorMessage) {
65   std::stringstream ErrorStream;
66   bool HasErrorMessage = false;
67   while (!Directory.empty()) {
68     std::string LoadErrorMessage;
69 
70     if (std::unique_ptr<CompilationDatabase> DB =
71             CompilationDatabase::loadFromDirectory(Directory, LoadErrorMessage))
72       return DB;
73 
74     if (!HasErrorMessage) {
75       ErrorStream << "No compilation database found in " << Directory.str()
76                   << " or any parent directory\n" << LoadErrorMessage;
77       HasErrorMessage = true;
78     }
79 
80     Directory = llvm::sys::path::parent_path(Directory);
81   }
82   ErrorMessage = ErrorStream.str();
83   return nullptr;
84 }
85 
86 std::unique_ptr<CompilationDatabase>
87 CompilationDatabase::autoDetectFromSource(StringRef SourceFile,
88                                           std::string &ErrorMessage) {
89   SmallString<1024> AbsolutePath(getAbsolutePath(SourceFile));
90   StringRef Directory = llvm::sys::path::parent_path(AbsolutePath);
91 
92   std::unique_ptr<CompilationDatabase> DB =
93       findCompilationDatabaseFromDirectory(Directory, ErrorMessage);
94 
95   if (!DB)
96     ErrorMessage = ("Could not auto-detect compilation database for file \"" +
97                    SourceFile + "\"\n" + ErrorMessage).str();
98   return DB;
99 }
100 
101 std::unique_ptr<CompilationDatabase>
102 CompilationDatabase::autoDetectFromDirectory(StringRef SourceDir,
103                                              std::string &ErrorMessage) {
104   SmallString<1024> AbsolutePath(getAbsolutePath(SourceDir));
105 
106   std::unique_ptr<CompilationDatabase> DB =
107       findCompilationDatabaseFromDirectory(AbsolutePath, ErrorMessage);
108 
109   if (!DB)
110     ErrorMessage = ("Could not auto-detect compilation database from directory \"" +
111                    SourceDir + "\"\n" + ErrorMessage).str();
112   return DB;
113 }
114 
115 std::vector<CompileCommand> CompilationDatabase::getAllCompileCommands() const {
116   std::vector<CompileCommand> Result;
117   for (const auto &File : getAllFiles()) {
118     auto C = getCompileCommands(File);
119     std::move(C.begin(), C.end(), std::back_inserter(Result));
120   }
121   return Result;
122 }
123 
124 CompilationDatabasePlugin::~CompilationDatabasePlugin() {}
125 
126 namespace {
127 // Helper for recursively searching through a chain of actions and collecting
128 // all inputs, direct and indirect, of compile jobs.
129 struct CompileJobAnalyzer {
130   void run(const driver::Action *A) {
131     runImpl(A, false);
132   }
133 
134   SmallVector<std::string, 2> Inputs;
135 
136 private:
137 
138   void runImpl(const driver::Action *A, bool Collect) {
139     bool CollectChildren = Collect;
140     switch (A->getKind()) {
141     case driver::Action::CompileJobClass:
142       CollectChildren = true;
143       break;
144 
145     case driver::Action::InputClass: {
146       if (Collect) {
147         const driver::InputAction *IA = cast<driver::InputAction>(A);
148         Inputs.push_back(IA->getInputArg().getSpelling());
149       }
150     } break;
151 
152     default:
153       // Don't care about others
154       ;
155     }
156 
157     for (const driver::Action *AI : A->inputs())
158       runImpl(AI, CollectChildren);
159   }
160 };
161 
162 // Special DiagnosticConsumer that looks for warn_drv_input_file_unused
163 // diagnostics from the driver and collects the option strings for those unused
164 // options.
165 class UnusedInputDiagConsumer : public DiagnosticConsumer {
166 public:
167   UnusedInputDiagConsumer(DiagnosticConsumer &Other) : Other(Other) {}
168 
169   void HandleDiagnostic(DiagnosticsEngine::Level DiagLevel,
170                         const Diagnostic &Info) override {
171     if (Info.getID() == clang::diag::warn_drv_input_file_unused) {
172       // Arg 1 for this diagnostic is the option that didn't get used.
173       UnusedInputs.push_back(Info.getArgStdStr(0));
174     } else if (DiagLevel >= DiagnosticsEngine::Error) {
175       // If driver failed to create compilation object, show the diagnostics
176       // to user.
177       Other.HandleDiagnostic(DiagLevel, Info);
178     }
179   }
180 
181   DiagnosticConsumer &Other;
182   SmallVector<std::string, 2> UnusedInputs;
183 };
184 
185 // Unary functor for asking "Given a StringRef S1, does there exist a string
186 // S2 in Arr where S1 == S2?"
187 struct MatchesAny {
188   MatchesAny(ArrayRef<std::string> Arr) : Arr(Arr) {}
189   bool operator() (StringRef S) {
190     for (const std::string *I = Arr.begin(), *E = Arr.end(); I != E; ++I)
191       if (*I == S)
192         return true;
193     return false;
194   }
195 private:
196   ArrayRef<std::string> Arr;
197 };
198 } // namespace
199 
200 /// \brief Strips any positional args and possible argv[0] from a command-line
201 /// provided by the user to construct a FixedCompilationDatabase.
202 ///
203 /// FixedCompilationDatabase requires a command line to be in this format as it
204 /// constructs the command line for each file by appending the name of the file
205 /// to be compiled. FixedCompilationDatabase also adds its own argv[0] to the
206 /// start of the command line although its value is not important as it's just
207 /// ignored by the Driver invoked by the ClangTool using the
208 /// FixedCompilationDatabase.
209 ///
210 /// FIXME: This functionality should probably be made available by
211 /// clang::driver::Driver although what the interface should look like is not
212 /// clear.
213 ///
214 /// \param[in] Args Args as provided by the user.
215 /// \return Resulting stripped command line.
216 ///          \li true if successful.
217 ///          \li false if \c Args cannot be used for compilation jobs (e.g.
218 ///          contains an option like -E or -version).
219 static bool stripPositionalArgs(std::vector<const char *> Args,
220                                 std::vector<std::string> &Result,
221                                 std::string &ErrorMsg) {
222   IntrusiveRefCntPtr<DiagnosticOptions> DiagOpts = new DiagnosticOptions();
223   llvm::raw_string_ostream Output(ErrorMsg);
224   TextDiagnosticPrinter DiagnosticPrinter(Output, &*DiagOpts);
225   UnusedInputDiagConsumer DiagClient(DiagnosticPrinter);
226   DiagnosticsEngine Diagnostics(
227       IntrusiveRefCntPtr<clang::DiagnosticIDs>(new DiagnosticIDs()),
228       &*DiagOpts, &DiagClient, false);
229 
230   // The clang executable path isn't required since the jobs the driver builds
231   // will not be executed.
232   std::unique_ptr<driver::Driver> NewDriver(new driver::Driver(
233       /* ClangExecutable= */ "", llvm::sys::getDefaultTargetTriple(),
234       Diagnostics));
235   NewDriver->setCheckInputsExist(false);
236 
237   // This becomes the new argv[0]. The value is actually not important as it
238   // isn't used for invoking Tools.
239   Args.insert(Args.begin(), "clang-tool");
240 
241   // By adding -c, we force the driver to treat compilation as the last phase.
242   // It will then issue warnings via Diagnostics about un-used options that
243   // would have been used for linking. If the user provided a compiler name as
244   // the original argv[0], this will be treated as a linker input thanks to
245   // insertng a new argv[0] above. All un-used options get collected by
246   // UnusedInputdiagConsumer and get stripped out later.
247   Args.push_back("-c");
248 
249   // Put a dummy C++ file on to ensure there's at least one compile job for the
250   // driver to construct. If the user specified some other argument that
251   // prevents compilation, e.g. -E or something like -version, we may still end
252   // up with no jobs but then this is the user's fault.
253   Args.push_back("placeholder.cpp");
254 
255   // Remove -no-integrated-as; it's not used for syntax checking,
256   // and it confuses targets which don't support this option.
257   Args.erase(std::remove_if(Args.begin(), Args.end(),
258                             MatchesAny(std::string("-no-integrated-as"))),
259              Args.end());
260 
261   const std::unique_ptr<driver::Compilation> Compilation(
262       NewDriver->BuildCompilation(Args));
263   if (!Compilation)
264     return false;
265 
266   const driver::JobList &Jobs = Compilation->getJobs();
267 
268   CompileJobAnalyzer CompileAnalyzer;
269 
270   for (const auto &Cmd : Jobs) {
271     // Collect only for Assemble and Compile jobs. If we do all jobs we get
272     // duplicates since Link jobs point to Assemble jobs as inputs.
273     if (Cmd.getSource().getKind() == driver::Action::AssembleJobClass ||
274         Cmd.getSource().getKind() == driver::Action::CompileJobClass) {
275       CompileAnalyzer.run(&Cmd.getSource());
276     }
277   }
278 
279   if (CompileAnalyzer.Inputs.empty()) {
280     ErrorMsg = "warning: no compile jobs found\n";
281     return false;
282   }
283 
284   // Remove all compilation input files from the command line. This is
285   // necessary so that getCompileCommands() can construct a command line for
286   // each file.
287   std::vector<const char *>::iterator End = std::remove_if(
288       Args.begin(), Args.end(), MatchesAny(CompileAnalyzer.Inputs));
289 
290   // Remove all inputs deemed unused for compilation.
291   End = std::remove_if(Args.begin(), End, MatchesAny(DiagClient.UnusedInputs));
292 
293   // Remove the -c add above as well. It will be at the end right now.
294   assert(strcmp(*(End - 1), "-c") == 0);
295   --End;
296 
297   Result = std::vector<std::string>(Args.begin() + 1, End);
298   return true;
299 }
300 
301 std::unique_ptr<FixedCompilationDatabase>
302 FixedCompilationDatabase::loadFromCommandLine(int &Argc,
303                                               const char *const *Argv,
304                                               std::string &ErrorMsg,
305                                               Twine Directory) {
306   ErrorMsg.clear();
307   if (Argc == 0)
308     return nullptr;
309   const char *const *DoubleDash = std::find(Argv, Argv + Argc, StringRef("--"));
310   if (DoubleDash == Argv + Argc)
311     return nullptr;
312   std::vector<const char *> CommandLine(DoubleDash + 1, Argv + Argc);
313   Argc = DoubleDash - Argv;
314 
315   std::vector<std::string> StrippedArgs;
316   if (!stripPositionalArgs(CommandLine, StrippedArgs, ErrorMsg))
317     return nullptr;
318   return llvm::make_unique<FixedCompilationDatabase>(Directory, StrippedArgs);
319 }
320 
321 std::unique_ptr<FixedCompilationDatabase>
322 FixedCompilationDatabase::loadFromFile(StringRef Path, std::string &ErrorMsg) {
323   ErrorMsg.clear();
324   llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> File =
325       llvm::MemoryBuffer::getFile(Path);
326   if (std::error_code Result = File.getError()) {
327     ErrorMsg = "Error while opening fixed database: " + Result.message();
328     return nullptr;
329   }
330   std::vector<std::string> Args{llvm::line_iterator(**File),
331                                 llvm::line_iterator()};
332   return llvm::make_unique<FixedCompilationDatabase>(
333       llvm::sys::path::parent_path(Path), std::move(Args));
334 }
335 
336 FixedCompilationDatabase::
337 FixedCompilationDatabase(Twine Directory, ArrayRef<std::string> CommandLine) {
338   std::vector<std::string> ToolCommandLine(1, "clang-tool");
339   ToolCommandLine.insert(ToolCommandLine.end(),
340                          CommandLine.begin(), CommandLine.end());
341   CompileCommands.emplace_back(Directory, StringRef(),
342                                std::move(ToolCommandLine),
343                                StringRef());
344 }
345 
346 std::vector<CompileCommand>
347 FixedCompilationDatabase::getCompileCommands(StringRef FilePath) const {
348   std::vector<CompileCommand> Result(CompileCommands);
349   Result[0].CommandLine.push_back(FilePath);
350   Result[0].Filename = FilePath;
351   return Result;
352 }
353 
354 namespace {
355 
356 class FixedCompilationDatabasePlugin : public CompilationDatabasePlugin {
357   std::unique_ptr<CompilationDatabase>
358   loadFromDirectory(StringRef Directory, std::string &ErrorMessage) override {
359     SmallString<1024> DatabasePath(Directory);
360     llvm::sys::path::append(DatabasePath, "compile_flags.txt");
361     return FixedCompilationDatabase::loadFromFile(DatabasePath, ErrorMessage);
362   }
363 };
364 
365 static CompilationDatabasePluginRegistry::Add<FixedCompilationDatabasePlugin>
366 X("fixed-compilation-database", "Reads plain-text flags file");
367 
368 } // namespace
369 
370 namespace clang {
371 namespace tooling {
372 
373 // This anchor is used to force the linker to link in the generated object file
374 // and thus register the JSONCompilationDatabasePlugin.
375 extern volatile int JSONAnchorSource;
376 static int LLVM_ATTRIBUTE_UNUSED JSONAnchorDest = JSONAnchorSource;
377 
378 } // end namespace tooling
379 } // end namespace clang
380