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   constexpr char const *SIS = "#include <...> search starts here:";
62   constexpr char const *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 = std::find(Lines.begin(), Lines.end(), SIS);
67   if (StartIt == Lines.end()) {
68     elog("System include extraction: start marker not found: {0}", Output);
69     return {};
70   }
71   ++StartIt;
72   const auto EndIt = std::find(StartIt, Lines.end(), SIE);
73   if (EndIt == Lines.end()) {
74     elog("System include extraction: end marker missing: {0}", Output);
75     return {};
76   }
77 
78   for (llvm::StringRef Line : llvm::make_range(StartIt, EndIt)) {
79     SystemIncludes.push_back(Line.str());
80     vlog("System include extraction: adding {0}", Line);
81   }
82   return SystemIncludes;
83 }
84 
85 std::vector<std::string> extractSystemIncludes(PathRef Driver,
86                                                llvm::StringRef Ext,
87                                                llvm::Regex &QueryDriverRegex) {
88   trace::Span Tracer("Extract system includes");
89   SPAN_ATTACH(Tracer, "driver", Driver);
90   SPAN_ATTACH(Tracer, "ext", Ext);
91 
92   if (!QueryDriverRegex.match(Driver)) {
93     vlog("System include extraction: not whitelisted driver {0}", Driver);
94     return {};
95   }
96 
97   if (!llvm::sys::fs::exists(Driver)) {
98     elog("System include extraction: {0} does not exist.", Driver);
99     return {};
100   }
101   if (!llvm::sys::fs::can_execute(Driver)) {
102     elog("System include extraction: {0} is not executable.", Driver);
103     return {};
104   }
105 
106   llvm::SmallString<128> StdErrPath;
107   if (auto EC = llvm::sys::fs::createTemporaryFile("system-includes", "clangd",
108                                                    StdErrPath)) {
109     elog("System include extraction: failed to create temporary file with "
110          "error {0}",
111          EC.message());
112     return {};
113   }
114   auto CleanUp = llvm::make_scope_exit(
115       [&StdErrPath]() { llvm::sys::fs::remove(StdErrPath); });
116 
117   llvm::Optional<llvm::StringRef> Redirects[] = {
118       {""}, {""}, llvm::StringRef(StdErrPath)};
119 
120   auto Type = driver::types::lookupTypeForExtension(Ext);
121   if (Type == driver::types::TY_INVALID) {
122     elog("System include extraction: invalid file type for {0}", Ext);
123     return {};
124   }
125   // Should we also preserve flags like "-sysroot", "-nostdinc" ?
126   const llvm::StringRef Args[] = {
127       Driver, "-E", "-x", driver::types::getTypeName(Type), "-", "-v"};
128 
129   if (int RC = llvm::sys::ExecuteAndWait(Driver, Args, /*Env=*/llvm::None,
130                                          Redirects)) {
131     elog("System include extraction: driver execution failed with return code: "
132          "{0}",
133          llvm::to_string(RC));
134     return {};
135   }
136 
137   auto BufOrError = llvm::MemoryBuffer::getFile(StdErrPath);
138   if (!BufOrError) {
139     elog("System include extraction: failed to read {0} with error {1}",
140          StdErrPath, BufOrError.getError().message());
141     return {};
142   }
143 
144   auto Includes = parseDriverOutput(BufOrError->get()->getBuffer());
145   log("System include extractor: succesfully executed {0}, got includes: "
146       "\"{1}\"",
147       Driver, llvm::join(Includes, ", "));
148   return Includes;
149 }
150 
151 tooling::CompileCommand &
152 addSystemIncludes(tooling::CompileCommand &Cmd,
153                   llvm::ArrayRef<std::string> SystemIncludes) {
154   for (llvm::StringRef Include : SystemIncludes) {
155     // FIXME(kadircet): This doesn't work when we have "--driver-mode=cl"
156     Cmd.CommandLine.push_back("-isystem");
157     Cmd.CommandLine.push_back(Include.str());
158   }
159   return Cmd;
160 }
161 
162 /// Converts a glob containing only ** or * into a regex.
163 std::string convertGlobToRegex(llvm::StringRef Glob) {
164   std::string RegText;
165   llvm::raw_string_ostream RegStream(RegText);
166   RegStream << '^';
167   for (size_t I = 0, E = Glob.size(); I < E; ++I) {
168     if (Glob[I] == '*') {
169       if (I + 1 < E && Glob[I + 1] == '*') {
170         // Double star, accept any sequence.
171         RegStream << ".*";
172         // Also skip the second star.
173         ++I;
174       } else {
175         // Single star, accept any sequence without a slash.
176         RegStream << "[^/]*";
177       }
178     } else {
179       RegStream << llvm::Regex::escape(Glob.substr(I, 1));
180     }
181   }
182   RegStream << '$';
183   RegStream.flush();
184   return RegText;
185 }
186 
187 /// Converts a glob containing only ** or * into a regex.
188 llvm::Regex convertGlobsToRegex(llvm::ArrayRef<std::string> Globs) {
189   assert(!Globs.empty() && "Globs cannot be empty!");
190   std::vector<std::string> RegTexts;
191   RegTexts.reserve(Globs.size());
192   for (llvm::StringRef Glob : Globs)
193     RegTexts.push_back(convertGlobToRegex(Glob));
194 
195   llvm::Regex Reg(llvm::join(RegTexts, "|"));
196   assert(Reg.isValid(RegTexts.front()) &&
197          "Created an invalid regex from globs");
198   return Reg;
199 }
200 
201 /// Extracts system includes from a trusted driver by parsing the output of
202 /// include search path and appends them to the commands coming from underlying
203 /// compilation database.
204 class QueryDriverDatabase : public GlobalCompilationDatabase {
205 public:
206   QueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
207                       std::unique_ptr<GlobalCompilationDatabase> Base)
208       : QueryDriverRegex(convertGlobsToRegex(QueryDriverGlobs)),
209         Base(std::move(Base)) {
210     assert(this->Base);
211     BaseChanged =
212         this->Base->watch([this](const std::vector<std::string> &Changes) {
213           OnCommandChanged.broadcast(Changes);
214         });
215   }
216 
217   llvm::Optional<tooling::CompileCommand>
218   getCompileCommand(PathRef File, ProjectInfo *PI = nullptr) const override {
219     auto Cmd = Base->getCompileCommand(File, PI);
220     if (!Cmd || Cmd->CommandLine.empty())
221       return Cmd;
222 
223     llvm::SmallString<128> Driver(Cmd->CommandLine.front());
224     llvm::sys::fs::make_absolute(Cmd->Directory, Driver);
225     llvm::StringRef Ext = llvm::sys::path::extension(File).trim('.');
226     auto Key = std::make_pair(Driver.str(), Ext);
227 
228     std::vector<std::string> SystemIncludes;
229     {
230       std::lock_guard<std::mutex> Lock(Mu);
231 
232       auto It = DriverToIncludesCache.find(Key);
233       if (It != DriverToIncludesCache.end())
234         SystemIncludes = It->second;
235       else
236         DriverToIncludesCache[Key] = SystemIncludes =
237             extractSystemIncludes(Key.first, Key.second, QueryDriverRegex);
238     }
239 
240     return addSystemIncludes(*Cmd, SystemIncludes);
241   }
242 
243 private:
244   mutable std::mutex Mu;
245   // Caches includes extracted from a driver.
246   mutable std::map<std::pair<std::string, std::string>,
247                    std::vector<std::string>>
248       DriverToIncludesCache;
249   mutable llvm::Regex QueryDriverRegex;
250 
251   std::unique_ptr<GlobalCompilationDatabase> Base;
252   CommandChanged::Subscription BaseChanged;
253 };
254 } // namespace
255 
256 std::unique_ptr<GlobalCompilationDatabase>
257 getQueryDriverDatabase(llvm::ArrayRef<std::string> QueryDriverGlobs,
258                        std::unique_ptr<GlobalCompilationDatabase> Base) {
259   assert(Base && "Null base to SystemIncludeExtractor");
260   if (QueryDriverGlobs.empty())
261     return Base;
262   return llvm::make_unique<QueryDriverDatabase>(QueryDriverGlobs,
263                                                 std::move(Base));
264 }
265 
266 } // namespace clangd
267 } // namespace clang
268