1 //===--- Tooling.cpp - Running clang standalone tools ---------------------===//
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 implements functions to run clang tools standalone instead
11 //  of running them as a plugin.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "clang/Tooling/Tooling.h"
16 #include "clang/Driver/Compilation.h"
17 #include "clang/Driver/Driver.h"
18 #include "clang/Driver/Tool.h"
19 #include "clang/Frontend/CompilerInstance.h"
20 #include "clang/Frontend/FrontendDiagnostic.h"
21 #include "clang/Frontend/TextDiagnosticPrinter.h"
22 #include "clang/Tooling/ArgumentsAdjusters.h"
23 #include "clang/Tooling/CompilationDatabase.h"
24 #include "llvm/ADT/STLExtras.h"
25 #include "llvm/Option/Option.h"
26 #include "llvm/Support/Debug.h"
27 #include "llvm/Support/FileSystem.h"
28 #include "llvm/Support/Host.h"
29 #include "llvm/Support/raw_ostream.h"
30 
31 // For chdir, see the comment in ClangTool::run for more information.
32 #ifdef _WIN32
33 #  include <direct.h>
34 #else
35 #  include <unistd.h>
36 #endif
37 
38 namespace clang {
39 namespace tooling {
40 
41 FrontendActionFactory::~FrontendActionFactory() {}
42 
43 // FIXME: This file contains structural duplication with other parts of the
44 // code that sets up a compiler to run tools on it, and we should refactor
45 // it to be based on the same framework.
46 
47 /// \brief Builds a clang driver initialized for running clang tools.
48 static clang::driver::Driver *newDriver(clang::DiagnosticsEngine *Diagnostics,
49                                         const char *BinaryName) {
50   const std::string DefaultOutputName = "a.out";
51   clang::driver::Driver *CompilerDriver = new clang::driver::Driver(
52     BinaryName, llvm::sys::getDefaultTargetTriple(),
53     DefaultOutputName, *Diagnostics);
54   CompilerDriver->setTitle("clang_based_tool");
55   return CompilerDriver;
56 }
57 
58 /// \brief Retrieves the clang CC1 specific flags out of the compilation's jobs.
59 ///
60 /// Returns NULL on error.
61 static const llvm::opt::ArgStringList *getCC1Arguments(
62     clang::DiagnosticsEngine *Diagnostics,
63     clang::driver::Compilation *Compilation) {
64   // We expect to get back exactly one Command job, if we didn't something
65   // failed. Extract that job from the Compilation.
66   const clang::driver::JobList &Jobs = Compilation->getJobs();
67   if (Jobs.size() != 1 || !isa<clang::driver::Command>(*Jobs.begin())) {
68     SmallString<256> error_msg;
69     llvm::raw_svector_ostream error_stream(error_msg);
70     Jobs.Print(error_stream, "; ", true);
71     Diagnostics->Report(clang::diag::err_fe_expected_compiler_job)
72         << error_stream.str();
73     return NULL;
74   }
75 
76   // The one job we find should be to invoke clang again.
77   const clang::driver::Command *Cmd =
78       cast<clang::driver::Command>(*Jobs.begin());
79   if (StringRef(Cmd->getCreator().getName()) != "clang") {
80     Diagnostics->Report(clang::diag::err_fe_expected_clang_command);
81     return NULL;
82   }
83 
84   return &Cmd->getArguments();
85 }
86 
87 /// \brief Returns a clang build invocation initialized from the CC1 flags.
88 static clang::CompilerInvocation *newInvocation(
89     clang::DiagnosticsEngine *Diagnostics,
90     const llvm::opt::ArgStringList &CC1Args) {
91   assert(!CC1Args.empty() && "Must at least contain the program name!");
92   clang::CompilerInvocation *Invocation = new clang::CompilerInvocation;
93   clang::CompilerInvocation::CreateFromArgs(
94       *Invocation, CC1Args.data() + 1, CC1Args.data() + CC1Args.size(),
95       *Diagnostics);
96   Invocation->getFrontendOpts().DisableFree = false;
97   Invocation->getCodeGenOpts().DisableFree = false;
98   return Invocation;
99 }
100 
101 bool runToolOnCode(clang::FrontendAction *ToolAction, const Twine &Code,
102                    const Twine &FileName) {
103   return runToolOnCodeWithArgs(
104       ToolAction, Code, std::vector<std::string>(), FileName);
105 }
106 
107 bool runToolOnCodeWithArgs(clang::FrontendAction *ToolAction, const Twine &Code,
108                            const std::vector<std::string> &Args,
109                            const Twine &FileName) {
110   SmallString<16> FileNameStorage;
111   StringRef FileNameRef = FileName.toNullTerminatedStringRef(FileNameStorage);
112   std::vector<std::string> Commands;
113   Commands.push_back("clang-tool");
114   Commands.push_back("-fsyntax-only");
115   Commands.insert(Commands.end(), Args.begin(), Args.end());
116   Commands.push_back(FileNameRef.data());
117   FileManager Files((FileSystemOptions()));
118   ToolInvocation Invocation(Commands, ToolAction, &Files);
119 
120   SmallString<1024> CodeStorage;
121   Invocation.mapVirtualFile(FileNameRef,
122                             Code.toNullTerminatedStringRef(CodeStorage));
123   return Invocation.run();
124 }
125 
126 std::string getAbsolutePath(StringRef File) {
127   StringRef RelativePath(File);
128   // FIXME: Should '.\\' be accepted on Win32?
129   if (RelativePath.startswith("./")) {
130     RelativePath = RelativePath.substr(strlen("./"));
131   }
132 
133   SmallString<1024> AbsolutePath = RelativePath;
134   llvm::error_code EC = llvm::sys::fs::make_absolute(AbsolutePath);
135   assert(!EC);
136   (void)EC;
137   llvm::sys::path::native(AbsolutePath);
138   return AbsolutePath.str();
139 }
140 
141 ToolInvocation::ToolInvocation(
142     ArrayRef<std::string> CommandLine, FrontendAction *ToolAction,
143     FileManager *Files)
144     : CommandLine(CommandLine.vec()), ToolAction(ToolAction), Files(Files) {
145 }
146 
147 void ToolInvocation::mapVirtualFile(StringRef FilePath, StringRef Content) {
148   SmallString<1024> PathStorage;
149   llvm::sys::path::native(FilePath, PathStorage);
150   MappedFileContents[PathStorage] = Content;
151 }
152 
153 bool ToolInvocation::run() {
154   std::vector<const char*> Argv;
155   for (int I = 0, E = CommandLine.size(); I != E; ++I)
156     Argv.push_back(CommandLine[I].c_str());
157   const char *const BinaryName = Argv[0];
158   IntrusiveRefCntPtr<DiagnosticOptions> DiagOpts = new DiagnosticOptions();
159   TextDiagnosticPrinter DiagnosticPrinter(
160       llvm::errs(), &*DiagOpts);
161   DiagnosticsEngine Diagnostics(
162     IntrusiveRefCntPtr<clang::DiagnosticIDs>(new DiagnosticIDs()),
163     &*DiagOpts, &DiagnosticPrinter, false);
164 
165   const OwningPtr<clang::driver::Driver> Driver(
166       newDriver(&Diagnostics, BinaryName));
167   // Since the input might only be virtual, don't check whether it exists.
168   Driver->setCheckInputsExist(false);
169   const OwningPtr<clang::driver::Compilation> Compilation(
170       Driver->BuildCompilation(llvm::makeArrayRef(Argv)));
171   const llvm::opt::ArgStringList *const CC1Args = getCC1Arguments(
172       &Diagnostics, Compilation.get());
173   if (CC1Args == NULL) {
174     return false;
175   }
176   OwningPtr<clang::CompilerInvocation> Invocation(
177       newInvocation(&Diagnostics, *CC1Args));
178   return runInvocation(BinaryName, Compilation.get(), Invocation.take());
179 }
180 
181 bool ToolInvocation::runInvocation(
182     const char *BinaryName,
183     clang::driver::Compilation *Compilation,
184     clang::CompilerInvocation *Invocation) {
185   // Show the invocation, with -v.
186   if (Invocation->getHeaderSearchOpts().Verbose) {
187     llvm::errs() << "clang Invocation:\n";
188     Compilation->getJobs().Print(llvm::errs(), "\n", true);
189     llvm::errs() << "\n";
190   }
191 
192   // Create a compiler instance to handle the actual work.
193   clang::CompilerInstance Compiler;
194   Compiler.setInvocation(Invocation);
195   Compiler.setFileManager(Files);
196   // FIXME: What about LangOpts?
197 
198   // ToolAction can have lifetime requirements for Compiler or its members, and
199   // we need to ensure it's deleted earlier than Compiler. So we pass it to an
200   // OwningPtr declared after the Compiler variable.
201   OwningPtr<FrontendAction> ScopedToolAction(ToolAction.take());
202 
203   // Create the compilers actual diagnostics engine.
204   Compiler.createDiagnostics();
205   if (!Compiler.hasDiagnostics())
206     return false;
207 
208   Compiler.createSourceManager(*Files);
209   addFileMappingsTo(Compiler.getSourceManager());
210 
211   const bool Success = Compiler.ExecuteAction(*ScopedToolAction);
212 
213   Compiler.resetAndLeakFileManager();
214   Files->clearStatCaches();
215   return Success;
216 }
217 
218 void ToolInvocation::addFileMappingsTo(SourceManager &Sources) {
219   for (llvm::StringMap<StringRef>::const_iterator
220            It = MappedFileContents.begin(), End = MappedFileContents.end();
221        It != End; ++It) {
222     // Inject the code as the given file name into the preprocessor options.
223     const llvm::MemoryBuffer *Input =
224         llvm::MemoryBuffer::getMemBuffer(It->getValue());
225     // FIXME: figure out what '0' stands for.
226     const FileEntry *FromFile = Files->getVirtualFile(
227         It->getKey(), Input->getBufferSize(), 0);
228     Sources.overrideFileContents(FromFile, Input);
229   }
230 }
231 
232 ClangTool::ClangTool(const CompilationDatabase &Compilations,
233                      ArrayRef<std::string> SourcePaths)
234     : Files((FileSystemOptions())) {
235   ArgsAdjusters.push_back(new ClangStripOutputAdjuster());
236   ArgsAdjusters.push_back(new ClangSyntaxOnlyAdjuster());
237   for (unsigned I = 0, E = SourcePaths.size(); I != E; ++I) {
238     SmallString<1024> File(getAbsolutePath(SourcePaths[I]));
239 
240     std::vector<CompileCommand> CompileCommandsForFile =
241       Compilations.getCompileCommands(File.str());
242     if (!CompileCommandsForFile.empty()) {
243       for (int I = 0, E = CompileCommandsForFile.size(); I != E; ++I) {
244         CompileCommands.push_back(std::make_pair(File.str(),
245                                   CompileCommandsForFile[I]));
246       }
247     } else {
248       // FIXME: There are two use cases here: doing a fuzzy
249       // "find . -name '*.cc' |xargs tool" match, where as a user I don't care
250       // about the .cc files that were not found, and the use case where I
251       // specify all files I want to run over explicitly, where this should
252       // be an error. We'll want to add an option for this.
253       llvm::outs() << "Skipping " << File << ". Command line not found.\n";
254     }
255   }
256 }
257 
258 void ClangTool::mapVirtualFile(StringRef FilePath, StringRef Content) {
259   MappedFileContents.push_back(std::make_pair(FilePath, Content));
260 }
261 
262 void ClangTool::setArgumentsAdjuster(ArgumentsAdjuster *Adjuster) {
263   clearArgumentsAdjusters();
264   appendArgumentsAdjuster(Adjuster);
265 }
266 
267 void ClangTool::appendArgumentsAdjuster(ArgumentsAdjuster *Adjuster) {
268   ArgsAdjusters.push_back(Adjuster);
269 }
270 
271 void ClangTool::clearArgumentsAdjusters() {
272   for (unsigned I = 0, E = ArgsAdjusters.size(); I != E; ++I)
273     delete ArgsAdjusters[I];
274   ArgsAdjusters.clear();
275 }
276 
277 int ClangTool::run(FrontendActionFactory *ActionFactory) {
278   // Exists solely for the purpose of lookup of the resource path.
279   // This just needs to be some symbol in the binary.
280   static int StaticSymbol;
281   // The driver detects the builtin header path based on the path of the
282   // executable.
283   // FIXME: On linux, GetMainExecutable is independent of the value of the
284   // first argument, thus allowing ClangTool and runToolOnCode to just
285   // pass in made-up names here. Make sure this works on other platforms.
286   std::string MainExecutable =
287       llvm::sys::fs::getMainExecutable("clang_tool", &StaticSymbol);
288 
289   bool ProcessingFailed = false;
290   for (unsigned I = 0; I < CompileCommands.size(); ++I) {
291     std::string File = CompileCommands[I].first;
292     // FIXME: chdir is thread hostile; on the other hand, creating the same
293     // behavior as chdir is complex: chdir resolves the path once, thus
294     // guaranteeing that all subsequent relative path operations work
295     // on the same path the original chdir resulted in. This makes a difference
296     // for example on network filesystems, where symlinks might be switched
297     // during runtime of the tool. Fixing this depends on having a file system
298     // abstraction that allows openat() style interactions.
299     if (chdir(CompileCommands[I].second.Directory.c_str()))
300       llvm::report_fatal_error("Cannot chdir into \"" +
301                                CompileCommands[I].second.Directory + "\n!");
302     std::vector<std::string> CommandLine = CompileCommands[I].second.CommandLine;
303     for (unsigned I = 0, E = ArgsAdjusters.size(); I != E; ++I)
304       CommandLine = ArgsAdjusters[I]->Adjust(CommandLine);
305     assert(!CommandLine.empty());
306     CommandLine[0] = MainExecutable;
307     // FIXME: We need a callback mechanism for the tool writer to output a
308     // customized message for each file.
309     DEBUG({
310       llvm::dbgs() << "Processing: " << File << ".\n";
311     });
312     ToolInvocation Invocation(CommandLine, ActionFactory->create(), &Files);
313     for (int I = 0, E = MappedFileContents.size(); I != E; ++I) {
314       Invocation.mapVirtualFile(MappedFileContents[I].first,
315                                 MappedFileContents[I].second);
316     }
317     if (!Invocation.run()) {
318       // FIXME: Diagnostics should be used instead.
319       llvm::errs() << "Error while processing " << File << ".\n";
320       ProcessingFailed = true;
321     }
322   }
323   return ProcessingFailed ? 1 : 0;
324 }
325 
326 } // end namespace tooling
327 } // end namespace clang
328