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 whitelisted by the user.
31 
32 #include "GlobalCompilationDatabase.h"
33 #include "Logger.h"
34 #include "Path.h"
35 #include "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> extractSystemIncludes(PathRef Driver,
89                                                llvm::StringRef Lang,
90                                                llvm::Regex &QueryDriverRegex) {
91   trace::Span Tracer("Extract system includes");
92   SPAN_ATTACH(Tracer, "driver", Driver);
93   SPAN_ATTACH(Tracer, "lang", Lang);
94 
95   if (!QueryDriverRegex.match(Driver)) {
96     vlog("System include extraction: not whitelisted driver {0}", Driver);
97     return {};
98   }
99 
100   if (!llvm::sys::fs::exists(Driver)) {
101     elog("System include extraction: {0} does not exist.", Driver);
102     return {};
103   }
104   if (!llvm::sys::fs::can_execute(Driver)) {
105     elog("System include extraction: {0} is not executable.", Driver);
106     return {};
107   }
108 
109   llvm::SmallString<128> StdErrPath;
110   if (auto EC = llvm::sys::fs::createTemporaryFile("system-includes", "clangd",
111                                                    StdErrPath)) {
112     elog("System include extraction: failed to create temporary file with "
113          "error {0}",
114          EC.message());
115     return {};
116   }
117   auto CleanUp = llvm::make_scope_exit(
118       [&StdErrPath]() { llvm::sys::fs::remove(StdErrPath); });
119 
120   llvm::Optional<llvm::StringRef> Redirects[] = {
121       {""}, {""}, llvm::StringRef(StdErrPath)};
122 
123   // Should we also preserve flags like "-sysroot", "-nostdinc" ?
124   const llvm::StringRef Args[] = {Driver, "-E", "-x", Lang, "-", "-v"};
125 
126   if (int RC = llvm::sys::ExecuteAndWait(Driver, Args, /*Env=*/llvm::None,
127                                          Redirects)) {
128     elog("System include extraction: driver execution failed with return code: "
129          "{0}",
130          llvm::to_string(RC));
131     return {};
132   }
133 
134   auto BufOrError = llvm::MemoryBuffer::getFile(StdErrPath);
135   if (!BufOrError) {
136     elog("System include extraction: failed to read {0} with error {1}",
137          StdErrPath, BufOrError.getError().message());
138     return {};
139   }
140 
141   auto Includes = parseDriverOutput(BufOrError->get()->getBuffer());
142   log("System include extractor: succesfully executed {0}, got includes: "
143       "\"{1}\"",
144       Driver, llvm::join(Includes, ", "));
145   return Includes;
146 }
147 
148 tooling::CompileCommand &
149 addSystemIncludes(tooling::CompileCommand &Cmd,
150                   llvm::ArrayRef<std::string> SystemIncludes) {
151   for (llvm::StringRef Include : SystemIncludes) {
152     // FIXME(kadircet): This doesn't work when we have "--driver-mode=cl"
153     Cmd.CommandLine.push_back("-isystem");
154     Cmd.CommandLine.push_back(Include.str());
155   }
156   return Cmd;
157 }
158 
159 /// Converts a glob containing only ** or * into a regex.
160 std::string convertGlobToRegex(llvm::StringRef Glob) {
161   std::string RegText;
162   llvm::raw_string_ostream RegStream(RegText);
163   RegStream << '^';
164   for (size_t I = 0, E = Glob.size(); I < E; ++I) {
165     if (Glob[I] == '*') {
166       if (I + 1 < E && Glob[I + 1] == '*') {
167         // Double star, accept any sequence.
168         RegStream << ".*";
169         // Also skip the second star.
170         ++I;
171       } else {
172         // Single star, accept any sequence without a slash.
173         RegStream << "[^/]*";
174       }
175     } else {
176       RegStream << llvm::Regex::escape(Glob.substr(I, 1));
177     }
178   }
179   RegStream << '$';
180   RegStream.flush();
181   return RegText;
182 }
183 
184 /// Converts a glob containing only ** or * into a regex.
185 llvm::Regex convertGlobsToRegex(llvm::ArrayRef<std::string> Globs) {
186   assert(!Globs.empty() && "Globs cannot be empty!");
187   std::vector<std::string> RegTexts;
188   RegTexts.reserve(Globs.size());
189   for (llvm::StringRef Glob : Globs)
190     RegTexts.push_back(convertGlobToRegex(Glob));
191 
192   llvm::Regex Reg(llvm::join(RegTexts, "|"));
193   assert(Reg.isValid(RegTexts.front()) &&
194          "Created an invalid regex from globs");
195   return Reg;
196 }
197 
198 /// Extracts system includes from a trusted driver by parsing the output of
199 /// include search path and appends them to the commands coming from underlying
200 /// compilation database.
201 class QueryDriverDatabase : public GlobalCompilationDatabase {
202 public:
203   QueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
204                       std::unique_ptr<GlobalCompilationDatabase> Base)
205       : QueryDriverRegex(convertGlobsToRegex(QueryDriverGlobs)),
206         Base(std::move(Base)) {
207     assert(this->Base);
208     BaseChanged =
209         this->Base->watch([this](const std::vector<std::string> &Changes) {
210           OnCommandChanged.broadcast(Changes);
211         });
212   }
213 
214   llvm::Optional<tooling::CompileCommand>
215   getCompileCommand(PathRef File) const override {
216     auto Cmd = Base->getCompileCommand(File);
217     if (!Cmd || Cmd->CommandLine.empty())
218       return Cmd;
219 
220     llvm::StringRef Lang;
221     for (size_t I = 0, E = Cmd->CommandLine.size(); I < E; ++I) {
222       llvm::StringRef Arg = Cmd->CommandLine[I];
223       if (Arg == "-x" && I + 1 < E)
224         Lang = Cmd->CommandLine[I + 1];
225       else if (Arg.startswith("-x"))
226         Lang = Arg.drop_front(2).trim();
227     }
228     if (Lang.empty()) {
229       llvm::StringRef Ext = llvm::sys::path::extension(File).trim('.');
230       auto Type = driver::types::lookupTypeForExtension(Ext);
231       if (Type == driver::types::TY_INVALID) {
232         elog("System include extraction: invalid file type for {0}", Ext);
233         return {};
234       }
235       Lang = driver::types::getTypeName(Type);
236     }
237 
238     llvm::SmallString<128> Driver(Cmd->CommandLine.front());
239     llvm::sys::fs::make_absolute(Cmd->Directory, Driver);
240     auto Key = std::make_pair(Driver.str(), Lang);
241 
242     std::vector<std::string> SystemIncludes;
243     {
244       std::lock_guard<std::mutex> Lock(Mu);
245 
246       auto It = DriverToIncludesCache.find(Key);
247       if (It != DriverToIncludesCache.end())
248         SystemIncludes = It->second;
249       else
250         DriverToIncludesCache[Key] = SystemIncludes =
251             extractSystemIncludes(Key.first, Key.second, QueryDriverRegex);
252     }
253 
254     return addSystemIncludes(*Cmd, SystemIncludes);
255   }
256 
257   llvm::Optional<ProjectInfo> getProjectInfo(PathRef File) const override {
258     return Base->getProjectInfo(File);
259   }
260 
261 private:
262   mutable std::mutex Mu;
263   // Caches includes extracted from a driver.
264   mutable std::map<std::pair<std::string, std::string>,
265                    std::vector<std::string>>
266       DriverToIncludesCache;
267   mutable llvm::Regex QueryDriverRegex;
268 
269   std::unique_ptr<GlobalCompilationDatabase> Base;
270   CommandChanged::Subscription BaseChanged;
271 };
272 } // namespace
273 
274 std::unique_ptr<GlobalCompilationDatabase>
275 getQueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
276                        std::unique_ptr<GlobalCompilationDatabase> Base) {
277   assert(Base && "Null base to SystemIncludeExtractor");
278   if (QueryDriverGlobs.empty())
279     return Base;
280   return llvm::make_unique<QueryDriverDatabase>(QueryDriverGlobs,
281                                                 std::move(Base));
282 }
283 
284 } // namespace clangd
285 } // namespace clang
286