1 //===--- QueryDriverDatabase.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 // Some compiler drivers have implicit search mechanism for system headers.
9 // This compilation database implementation tries to extract that information by
10 // executing the driver in verbose mode. gcc-compatible drivers print something
11 // like:
12 // ....
13 // ....
14 // #include <...> search starts here:
15 //  /usr/lib/gcc/x86_64-linux-gnu/7/include
16 //  /usr/local/include
17 //  /usr/lib/gcc/x86_64-linux-gnu/7/include-fixed
18 //  /usr/include/x86_64-linux-gnu
19 //  /usr/include
20 // End of search list.
21 // ....
22 // ....
23 // This component parses that output and adds each path to command line args
24 // provided by Base, after prepending them with -isystem. Therefore current
25 // implementation would not work with a driver that is not gcc-compatible.
26 //
27 // First argument of the command line received from underlying compilation
28 // database is used as compiler driver path. Due to this arbitrary binary
29 // execution, this mechanism is not used by default and only executes binaries
30 // in the paths that are explicitly included by the user.
31 
32 #include "GlobalCompilationDatabase.h"
33 #include "support/Logger.h"
34 #include "support/Path.h"
35 #include "support/Trace.h"
36 #include "clang/Driver/Types.h"
37 #include "clang/Tooling/CompilationDatabase.h"
38 #include "llvm/ADT/DenseMap.h"
39 #include "llvm/ADT/ScopeExit.h"
40 #include "llvm/ADT/SmallString.h"
41 #include "llvm/ADT/StringExtras.h"
42 #include "llvm/ADT/StringRef.h"
43 #include "llvm/ADT/iterator_range.h"
44 #include "llvm/Support/FileSystem.h"
45 #include "llvm/Support/MemoryBuffer.h"
46 #include "llvm/Support/Path.h"
47 #include "llvm/Support/Program.h"
48 #include "llvm/Support/Regex.h"
49 #include "llvm/Support/ScopedPrinter.h"
50 #include <algorithm>
51 #include <map>
52 #include <string>
53 #include <vector>
54 
55 namespace clang {
56 namespace clangd {
57 namespace {
58 
59 std::vector<std::string> parseDriverOutput(llvm::StringRef Output) {
60   std::vector<std::string> SystemIncludes;
61   const char SIS[] = "#include <...> search starts here:";
62   const char SIE[] = "End of search list.";
63   llvm::SmallVector<llvm::StringRef, 8> Lines;
64   Output.split(Lines, '\n', /*MaxSplit=*/-1, /*KeepEmpty=*/false);
65 
66   auto StartIt = llvm::find_if(
67       Lines, [SIS](llvm::StringRef Line) { return Line.trim() == SIS; });
68   if (StartIt == Lines.end()) {
69     elog("System include extraction: start marker not found: {0}", Output);
70     return {};
71   }
72   ++StartIt;
73   const auto EndIt =
74       llvm::find_if(llvm::make_range(StartIt, Lines.end()),
75                     [SIE](llvm::StringRef Line) { return Line.trim() == SIE; });
76   if (EndIt == Lines.end()) {
77     elog("System include extraction: end marker missing: {0}", Output);
78     return {};
79   }
80 
81   for (llvm::StringRef Line : llvm::make_range(StartIt, EndIt)) {
82     SystemIncludes.push_back(Line.trim().str());
83     vlog("System include extraction: adding {0}", Line);
84   }
85   return SystemIncludes;
86 }
87 
88 std::vector<std::string>
89 extractSystemIncludes(PathRef Driver, llvm::StringRef Lang,
90                       llvm::ArrayRef<std::string> CommandLine,
91                       const llvm::Regex &QueryDriverRegex) {
92   trace::Span Tracer("Extract system includes");
93   SPAN_ATTACH(Tracer, "driver", Driver);
94   SPAN_ATTACH(Tracer, "lang", Lang);
95 
96   if (!QueryDriverRegex.match(Driver)) {
97     vlog("System include extraction: not allowed driver {0}", Driver);
98     return {};
99   }
100 
101   if (!llvm::sys::fs::exists(Driver)) {
102     elog("System include extraction: {0} does not exist.", Driver);
103     return {};
104   }
105   if (!llvm::sys::fs::can_execute(Driver)) {
106     elog("System include extraction: {0} is not executable.", Driver);
107     return {};
108   }
109 
110   llvm::SmallString<128> StdErrPath;
111   if (auto EC = llvm::sys::fs::createTemporaryFile("system-includes", "clangd",
112                                                    StdErrPath)) {
113     elog("System include extraction: failed to create temporary file with "
114          "error {0}",
115          EC.message());
116     return {};
117   }
118   auto CleanUp = llvm::make_scope_exit(
119       [&StdErrPath]() { llvm::sys::fs::remove(StdErrPath); });
120 
121   llvm::Optional<llvm::StringRef> Redirects[] = {
122       {""}, {""}, llvm::StringRef(StdErrPath)};
123 
124   llvm::SmallVector<llvm::StringRef, 12> Args = {Driver, "-E", "-x",
125                                                  Lang,   "-",  "-v"};
126 
127   // These flags will be preserved
128   const llvm::StringRef FlagsToPreserve[] = {
129       "-nostdinc", "--no-standard-includes", "-nostdinc++", "-nobuiltininc"};
130   // Preserves these flags and their values, either as separate args or with an
131   // equalsbetween them
132   const llvm::StringRef ArgsToPreserve[] = {"--sysroot", "-isysroot"};
133 
134   for (size_t I = 0, E = CommandLine.size(); I < E; ++I) {
135     llvm::StringRef Arg = CommandLine[I];
136     if (llvm::any_of(FlagsToPreserve,
137                      [&Arg](llvm::StringRef S) { return S == Arg; })) {
138       Args.push_back(Arg);
139     } else {
140       const auto *Found =
141           llvm::find_if(ArgsToPreserve, [&Arg](llvm::StringRef S) {
142             return Arg.startswith(S);
143           });
144       if (Found == std::end(ArgsToPreserve))
145         continue;
146       Arg = Arg.drop_front(Found->size());
147       if (Arg.empty() && I + 1 < E) {
148         Args.push_back(CommandLine[I]);
149         Args.push_back(CommandLine[++I]);
150       } else if (Arg.startswith("=")) {
151         Args.push_back(CommandLine[I]);
152       }
153     }
154   }
155 
156   if (int RC = llvm::sys::ExecuteAndWait(Driver, Args, /*Env=*/llvm::None,
157                                          Redirects)) {
158     elog("System include extraction: driver execution failed with return code: "
159          "{0}. Args: ['{1}']",
160          llvm::to_string(RC), llvm::join(Args, "', '"));
161     return {};
162   }
163 
164   auto BufOrError = llvm::MemoryBuffer::getFile(StdErrPath);
165   if (!BufOrError) {
166     elog("System include extraction: failed to read {0} with error {1}",
167          StdErrPath, BufOrError.getError().message());
168     return {};
169   }
170 
171   auto Includes = parseDriverOutput(BufOrError->get()->getBuffer());
172   log("System include extractor: successfully executed {0}, got includes: "
173       "\"{1}\"",
174       Driver, llvm::join(Includes, ", "));
175   return Includes;
176 }
177 
178 tooling::CompileCommand &
179 addSystemIncludes(tooling::CompileCommand &Cmd,
180                   llvm::ArrayRef<std::string> SystemIncludes) {
181   for (llvm::StringRef Include : SystemIncludes) {
182     // FIXME(kadircet): This doesn't work when we have "--driver-mode=cl"
183     Cmd.CommandLine.push_back("-isystem");
184     Cmd.CommandLine.push_back(Include.str());
185   }
186   return Cmd;
187 }
188 
189 /// Converts a glob containing only ** or * into a regex.
190 std::string convertGlobToRegex(llvm::StringRef Glob) {
191   std::string RegText;
192   llvm::raw_string_ostream RegStream(RegText);
193   RegStream << '^';
194   for (size_t I = 0, E = Glob.size(); I < E; ++I) {
195     if (Glob[I] == '*') {
196       if (I + 1 < E && Glob[I + 1] == '*') {
197         // Double star, accept any sequence.
198         RegStream << ".*";
199         // Also skip the second star.
200         ++I;
201       } else {
202         // Single star, accept any sequence without a slash.
203         RegStream << "[^/]*";
204       }
205     } else {
206       RegStream << llvm::Regex::escape(Glob.substr(I, 1));
207     }
208   }
209   RegStream << '$';
210   RegStream.flush();
211   return RegText;
212 }
213 
214 /// Converts a glob containing only ** or * into a regex.
215 llvm::Regex convertGlobsToRegex(llvm::ArrayRef<std::string> Globs) {
216   assert(!Globs.empty() && "Globs cannot be empty!");
217   std::vector<std::string> RegTexts;
218   RegTexts.reserve(Globs.size());
219   for (llvm::StringRef Glob : Globs)
220     RegTexts.push_back(convertGlobToRegex(Glob));
221 
222   llvm::Regex Reg(llvm::join(RegTexts, "|"));
223   assert(Reg.isValid(RegTexts.front()) &&
224          "Created an invalid regex from globs");
225   return Reg;
226 }
227 
228 /// Extracts system includes from a trusted driver by parsing the output of
229 /// include search path and appends them to the commands coming from underlying
230 /// compilation database.
231 class QueryDriverDatabase : public GlobalCompilationDatabase {
232 public:
233   QueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
234                       std::unique_ptr<GlobalCompilationDatabase> Base)
235       : QueryDriverRegex(convertGlobsToRegex(QueryDriverGlobs)),
236         Base(std::move(Base)) {
237     assert(this->Base);
238     BaseChanged =
239         this->Base->watch([this](const std::vector<std::string> &Changes) {
240           OnCommandChanged.broadcast(Changes);
241         });
242   }
243 
244   llvm::Optional<tooling::CompileCommand>
245   getCompileCommand(PathRef File) const override {
246     auto Cmd = Base->getCompileCommand(File);
247     if (!Cmd || Cmd->CommandLine.empty())
248       return Cmd;
249 
250     llvm::StringRef Lang;
251     for (size_t I = 0, E = Cmd->CommandLine.size(); I < E; ++I) {
252       llvm::StringRef Arg = Cmd->CommandLine[I];
253       if (Arg == "-x" && I + 1 < E)
254         Lang = Cmd->CommandLine[I + 1];
255       else if (Arg.startswith("-x"))
256         Lang = Arg.drop_front(2).trim();
257     }
258     if (Lang.empty()) {
259       llvm::StringRef Ext = llvm::sys::path::extension(File).trim('.');
260       auto Type = driver::types::lookupTypeForExtension(Ext);
261       if (Type == driver::types::TY_INVALID) {
262         elog("System include extraction: invalid file type for {0}", Ext);
263         return {};
264       }
265       Lang = driver::types::getTypeName(Type);
266     }
267 
268     llvm::SmallString<128> Driver(Cmd->CommandLine.front());
269     llvm::sys::fs::make_absolute(Cmd->Directory, Driver);
270 
271     std::vector<std::string> SystemIncludes =
272         QueriedDrivers.get(/*Key=*/(Driver + ":" + Lang).str(), [&] {
273           return extractSystemIncludes(Driver, Lang, Cmd->CommandLine,
274                                        QueryDriverRegex);
275         });
276 
277     return addSystemIncludes(*Cmd, SystemIncludes);
278   }
279 
280   llvm::Optional<ProjectInfo> getProjectInfo(PathRef File) const override {
281     return Base->getProjectInfo(File);
282   }
283 
284 private:
285   // Caches includes extracted from a driver. Key is driver:lang.
286   Memoize<llvm::StringMap<std::vector<std::string>>> QueriedDrivers;
287   llvm::Regex QueryDriverRegex;
288 
289   std::unique_ptr<GlobalCompilationDatabase> Base;
290   CommandChanged::Subscription BaseChanged;
291 };
292 } // namespace
293 
294 std::unique_ptr<GlobalCompilationDatabase>
295 getQueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
296                        std::unique_ptr<GlobalCompilationDatabase> Base) {
297   assert(Base && "Null base to SystemIncludeExtractor");
298   if (QueryDriverGlobs.empty())
299     return Base;
300   return std::make_unique<QueryDriverDatabase>(QueryDriverGlobs,
301                                                std::move(Base));
302 }
303 
304 } // namespace clangd
305 } // namespace clang
306