clang-tools-extra/clangd/CompileCommands.cpp

   1 //===--- CompileCommands.cpp ----------------------------------------------===//
   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
   9 #include "CompileCommands.h"
  10 #include "Config.h"
  11 #include "support/Logger.h"
  12 #include "support/Trace.h"
  13 #include "clang/Driver/Driver.h"
  14 #include "clang/Driver/Options.h"
  15 #include "clang/Frontend/CompilerInvocation.h"
  16 #include "clang/Tooling/CompilationDatabase.h"
  17 #include "clang/Tooling/Tooling.h"
  18 #include "llvm/ADT/ArrayRef.h"
  19 #include "llvm/ADT/STLExtras.h"
  20 #include "llvm/ADT/SmallVector.h"
  21 #include "llvm/ADT/StringRef.h"
  22 #include "llvm/Option/ArgList.h"
  23 #include "llvm/Option/Option.h"
  24 #include "llvm/Support/Allocator.h"
  25 #include "llvm/Support/Debug.h"
  26 #include "llvm/Support/FileSystem.h"
  27 #include "llvm/Support/FileUtilities.h"
  28 #include "llvm/Support/MemoryBuffer.h"
  29 #include "llvm/Support/Path.h"
  30 #include "llvm/Support/Program.h"
  31 #include "llvm/TargetParser/Host.h"
  32 #include <iterator>
  33 #include <optional>
  34 #include <string>
  35 #include <vector>
  36
  37 namespace clang {
  38 namespace clangd {
  39 namespace {
  40
  41 // Query apple's `xcrun` launcher, which is the source of truth for "how should"
  42 // clang be invoked on this system.
  43 std::optional<std::string> queryXcrun(llvm::ArrayRef<llvm::StringRef> Argv) {
  44   auto Xcrun = llvm::sys::findProgramByName("xcrun");
  45   if (!Xcrun) {
  46     log("Couldn't find xcrun. Hopefully you have a non-apple toolchain...");
  47     return std::nullopt;
  48   }
  49   llvm::SmallString<64> OutFile;
  50   llvm::sys::fs::createTemporaryFile("clangd-xcrun", "", OutFile);
  51   llvm::FileRemover OutRemover(OutFile);
  52   std::optional<llvm::StringRef> Redirects[3] = {
  53       /*stdin=*/{""}, /*stdout=*/{OutFile.str()}, /*stderr=*/{""}};
  54   vlog("Invoking {0} to find clang installation", *Xcrun);
  55   int Ret = llvm::sys::ExecuteAndWait(*Xcrun, Argv,
  56                                       /*Env=*/std::nullopt, Redirects,
  57                                       /*SecondsToWait=*/10);
  58   if (Ret != 0) {
  59     log("xcrun exists but failed with code {0}. "
  60         "If you have a non-apple toolchain, this is OK. "
  61         "Otherwise, try xcode-select --install.",
  62         Ret);
  63     return std::nullopt;
  64   }
  65
  66   auto Buf = llvm::MemoryBuffer::getFile(OutFile);
  67   if (!Buf) {
  68     log("Can't read xcrun output: {0}", Buf.getError().message());
  69     return std::nullopt;
  70   }
  71   StringRef Path = Buf->get()->getBuffer().trim();
  72   if (Path.empty()) {
  73     log("xcrun produced no output");
  74     return std::nullopt;
  75   }
  76   return Path.str();
  77 }
  78
  79 // Resolve symlinks if possible.
  80 std::string resolve(std::string Path) {
  81   llvm::SmallString<128> Resolved;
  82   if (llvm::sys::fs::real_path(Path, Resolved)) {
  83     log("Failed to resolve possible symlink {0}", Path);
  84     return Path;
  85   }
  86   return std::string(Resolved.str());
  87 }
  88
  89 // Get a plausible full `clang` path.
  90 // This is used in the fallback compile command, or when the CDB returns a
  91 // generic driver with no path.
  92 std::string detectClangPath() {
  93   // The driver and/or cc1 sometimes depend on the binary name to compute
  94   // useful things like the standard library location.
  95   // We need to emulate what clang on this system is likely to see.
  96   // cc1 in particular looks at the "real path" of the running process, and
  97   // so if /usr/bin/clang is a symlink, it sees the resolved path.
  98   // clangd doesn't have that luxury, so we resolve symlinks ourselves.
  99
 100   // On Mac, `which clang` is /usr/bin/clang. It runs `xcrun clang`, which knows
 101   // where the real clang is kept. We need to do the same thing,
 102   // because cc1 (not the driver!) will find libc++ relative to argv[0].
 103 #ifdef __APPLE__
 104   if (auto MacClang = queryXcrun({"xcrun", "--find", "clang"}))
 105     return resolve(std::move(*MacClang));
 106 #endif
 107   // On other platforms, just look for compilers on the PATH.
 108   for (const char *Name : {"clang", "gcc", "cc"})
 109     if (auto PathCC = llvm::sys::findProgramByName(Name))
 110       return resolve(std::move(*PathCC));
 111   // Fallback: a nonexistent 'clang' binary next to clangd.
 112   static int StaticForMainAddr;
 113   std::string ClangdExecutable =
 114       llvm::sys::fs::getMainExecutable("clangd", (void *)&StaticForMainAddr);
 115   SmallString<128> ClangPath;
 116   ClangPath = llvm::sys::path::parent_path(ClangdExecutable);
 117   llvm::sys::path::append(ClangPath, "clang");
 118   return std::string(ClangPath.str());
 119 }
 120
 121 // On mac, /usr/bin/clang sets SDKROOT and then invokes the real clang.
 122 // The effect of this is to set -isysroot correctly. We do the same.
 123 std::optional<std::string> detectSysroot() {
 124 #ifndef __APPLE__
 125   return std::nullopt;
 126 #endif
 127
 128   // SDKROOT overridden in environment, respect it. Driver will set isysroot.
 129   if (::getenv("SDKROOT"))
 130     return std::nullopt;
 131   return queryXcrun({"xcrun", "--show-sdk-path"});
 132 }
 133
 134 std::string detectStandardResourceDir() {
 135   static int StaticForMainAddr; // Just an address in this process.
 136   return CompilerInvocation::GetResourcesPath("clangd",
 137                                               (void *)&StaticForMainAddr);
 138 }
 139
 140 // The path passed to argv[0] is important:
 141 //  - its parent directory is Driver::Dir, used for library discovery
 142 //  - its basename affects CLI parsing (clang-cl) and other settings
 143 // Where possible it should be an absolute path with sensible directory, but
 144 // with the original basename.
 145 static std::string resolveDriver(llvm::StringRef Driver, bool FollowSymlink,
 146                                  std::optional<std::string> ClangPath) {
 147   auto SiblingOf = [&](llvm::StringRef AbsPath) {
 148     llvm::SmallString<128> Result = llvm::sys::path::parent_path(AbsPath);
 149     llvm::sys::path::append(Result, llvm::sys::path::filename(Driver));
 150     return Result.str().str();
 151   };
 152
 153   // First, eliminate relative paths.
 154   std::string Storage;
 155   if (!llvm::sys::path::is_absolute(Driver)) {
 156     // If it's working-dir relative like bin/clang, we can't resolve it.
 157     // FIXME: we could if we had the working directory here.
 158     // Let's hope it's not a symlink.
 159     if (llvm::any_of(Driver,
 160                      [](char C) { return llvm::sys::path::is_separator(C); }))
 161       return Driver.str();
 162     // If the driver is a generic like "g++" with no path, add clang dir.
 163     if (ClangPath &&
 164         (Driver == "clang" || Driver == "clang++" || Driver == "gcc" ||
 165          Driver == "g++" || Driver == "cc" || Driver == "c++")) {
 166       return SiblingOf(*ClangPath);
 167     }
 168     // Otherwise try to look it up on PATH. This won't change basename.
 169     auto Absolute = llvm::sys::findProgramByName(Driver);
 170     if (Absolute && llvm::sys::path::is_absolute(*Absolute))
 171       Driver = Storage = std::move(*Absolute);
 172     else if (ClangPath) // If we don't find it, use clang dir again.
 173       return SiblingOf(*ClangPath);
 174     else // Nothing to do: can't find the command and no detected dir.
 175       return Driver.str();
 176   }
 177
 178   // Now we have an absolute path, but it may be a symlink.
 179   assert(llvm::sys::path::is_absolute(Driver));
 180   if (FollowSymlink) {
 181     llvm::SmallString<256> Resolved;
 182     if (!llvm::sys::fs::real_path(Driver, Resolved))
 183       return SiblingOf(Resolved);
 184   }
 185   return Driver.str();
 186 }
 187
 188 } // namespace
 189
 190 CommandMangler::CommandMangler() {
 191   Tokenizer = llvm::Triple(llvm::sys::getProcessTriple()).isOSWindows()
 192                   ? llvm::cl::TokenizeWindowsCommandLine
 193                   : llvm::cl::TokenizeGNUCommandLine;
 194 }
 195
 196 CommandMangler CommandMangler::detect() {
 197   CommandMangler Result;
 198   Result.ClangPath = detectClangPath();
 199   Result.ResourceDir = detectStandardResourceDir();
 200   Result.Sysroot = detectSysroot();
 201   return Result;
 202 }
 203
 204 CommandMangler CommandMangler::forTests() { return CommandMangler(); }
 205
 206 void CommandMangler::operator()(tooling::CompileCommand &Command,
 207                                 llvm::StringRef File) const {
 208   std::vector<std::string> &Cmd = Command.CommandLine;
 209   trace::Span S("AdjustCompileFlags");
 210   // Most of the modifications below assumes the Cmd starts with a driver name.
 211   // We might consider injecting a generic driver name like "cc" or "c++", but
 212   // a Cmd missing the driver is probably rare enough in practice and erroneous.
 213   if (Cmd.empty())
 214     return;
 215
 216   // FS used for expanding response files.
 217   // FIXME: ExpandResponseFiles appears not to provide the usual
 218   // thread-safety guarantees, as the access to FS is not locked!
 219   // For now, use the real FS, which is known to be threadsafe (if we don't
 220   // use/change working directory, which ExpandResponseFiles doesn't).
 221   auto FS = llvm::vfs::getRealFileSystem();
 222   tooling::addExpandedResponseFiles(Cmd, Command.Directory, Tokenizer, *FS);
 223
 224   auto &OptTable = clang::driver::getDriverOptTable();
 225   // OriginalArgs needs to outlive ArgList.
 226   llvm::SmallVector<const char *, 16> OriginalArgs;
 227   OriginalArgs.reserve(Cmd.size());
 228   for (const auto &S : Cmd)
 229     OriginalArgs.push_back(S.c_str());
 230   bool IsCLMode = driver::IsClangCL(driver::getDriverMode(
 231       OriginalArgs[0], llvm::ArrayRef(OriginalArgs).slice(1)));
 232   // ParseArgs propagates missing arg/opt counts on error, but preserves
 233   // everything it could parse in ArgList. So we just ignore those counts.
 234   unsigned IgnoredCount;
 235   // Drop the executable name, as ParseArgs doesn't expect it. This means
 236   // indices are actually of by one between ArgList and OriginalArgs.
 237   llvm::opt::InputArgList ArgList;
 238   ArgList = OptTable.ParseArgs(
 239       llvm::ArrayRef(OriginalArgs).drop_front(), IgnoredCount, IgnoredCount,
 240       llvm::opt::Visibility(IsCLMode ? driver::options::CLOption
 241                                      : driver::options::ClangOption));
 242
 243   llvm::SmallVector<unsigned, 1> IndicesToDrop;
 244   // Having multiple architecture options (e.g. when building fat binaries)
 245   // results in multiple compiler jobs, which clangd cannot handle. In such
 246   // cases strip all the `-arch` options and fallback to default architecture.
 247   // As there are no signals to figure out which one user actually wants. They
 248   // can explicitly specify one through `CompileFlags.Add` if need be.
 249   unsigned ArchOptCount = 0;
 250   for (auto *Input : ArgList.filtered(driver::options::OPT_arch)) {
 251     ++ArchOptCount;
 252     for (auto I = 0U; I <= Input->getNumValues(); ++I)
 253       IndicesToDrop.push_back(Input->getIndex() + I);
 254   }
 255   // If there is a single `-arch` option, keep it.
 256   if (ArchOptCount < 2)
 257     IndicesToDrop.clear();
 258
 259   // In some cases people may try to reuse the command from another file, e.g.
 260   //   { File: "foo.h", CommandLine: "clang foo.cpp" }.
 261   // We assume the intent is to parse foo.h the same way as foo.cpp, or as if
 262   // it were being included from foo.cpp.
 263   //
 264   // We're going to rewrite the command to refer to foo.h, and this may change
 265   // its semantics (e.g. by parsing the file as C). If we do this, we should
 266   // use transferCompileCommand to adjust the argv.
 267   // In practice only the extension of the file matters, so do this only when
 268   // it differs.
 269   llvm::StringRef FileExtension = llvm::sys::path::extension(File);
 270   std::optional<std::string> TransferFrom;
 271   auto SawInput = [&](llvm::StringRef Input) {
 272     if (llvm::sys::path::extension(Input) != FileExtension)
 273       TransferFrom.emplace(Input);
 274   };
 275
 276   // Strip all the inputs and `--`. We'll put the input for the requested file
 277   // explicitly at the end of the flags. This ensures modifications done in the
 278   // following steps apply in more cases (like setting -x, which only affects
 279   // inputs that come after it).
 280   for (auto *Input : ArgList.filtered(driver::options::OPT_INPUT)) {
 281     SawInput(Input->getValue(0));
 282     IndicesToDrop.push_back(Input->getIndex());
 283   }
 284   // Anything after `--` is also treated as input, drop them as well.
 285   if (auto *DashDash =
 286           ArgList.getLastArgNoClaim(driver::options::OPT__DASH_DASH)) {
 287     auto DashDashIndex = DashDash->getIndex() + 1; // +1 accounts for Cmd[0]
 288     for (unsigned I = DashDashIndex; I < Cmd.size(); ++I)
 289       SawInput(Cmd[I]);
 290     Cmd.resize(DashDashIndex);
 291   }
 292   llvm::sort(IndicesToDrop);
 293   for (unsigned Idx : llvm::reverse(IndicesToDrop))
 294     // +1 to account for the executable name in Cmd[0] that
 295     // doesn't exist in ArgList.
 296     Cmd.erase(Cmd.begin() + Idx + 1);
 297   // All the inputs are stripped, append the name for the requested file. Rest
 298   // of the modifications should respect `--`.
 299   Cmd.push_back("--");
 300   Cmd.push_back(File.str());
 301
 302   if (TransferFrom) {
 303     tooling::CompileCommand TransferCmd;
 304     TransferCmd.Filename = std::move(*TransferFrom);
 305     TransferCmd.CommandLine = std::move(Cmd);
 306     TransferCmd = transferCompileCommand(std::move(TransferCmd), File);
 307     Cmd = std::move(TransferCmd.CommandLine);
 308     assert(Cmd.size() >= 2 && Cmd.back() == File &&
 309            Cmd[Cmd.size() - 2] == "--" &&
 310            "TransferCommand should produce a command ending in -- filename");
 311   }
 312
 313   for (auto &Edit : Config::current().CompileFlags.Edits)
 314     Edit(Cmd);
 315
 316   // The system include extractor needs to run:
 317   //  - AFTER transferCompileCommand(), because the -x flag it adds may be
 318   //    necessary for the system include extractor to identify the file type
 319   //  - AFTER applying CompileFlags.Edits, because the name of the compiler
 320   //    that needs to be invoked may come from the CompileFlags->Compiler key
 321   //  - BEFORE addTargetAndModeForProgramName(), because gcc doesn't support
 322   //    the target flag that might be added.
 323   //  - BEFORE resolveDriver() because that can mess up the driver path,
 324   //    e.g. changing gcc to /path/to/clang/bin/gcc
 325   if (SystemIncludeExtractor) {
 326     SystemIncludeExtractor(Command, File);
 327   }
 328
 329   tooling::addTargetAndModeForProgramName(Cmd, Cmd.front());
 330
 331   // Check whether the flag exists, either as -flag or -flag=*
 332   auto Has = [&](llvm::StringRef Flag) {
 333     for (llvm::StringRef Arg : Cmd) {
 334       if (Arg.consume_front(Flag) && (Arg.empty() || Arg[0] == '='))
 335         return true;
 336     }
 337     return false;
 338   };
 339
 340   llvm::erase_if(Cmd, [](llvm::StringRef Elem) {
 341     return Elem.startswith("--save-temps") || Elem.startswith("-save-temps");
 342   });
 343
 344   std::vector<std::string> ToAppend;
 345   if (ResourceDir && !Has("-resource-dir"))
 346     ToAppend.push_back(("-resource-dir=" + *ResourceDir));
 347
 348   // Don't set `-isysroot` if it is already set or if `--sysroot` is set.
 349   // `--sysroot` is a superset of the `-isysroot` argument.
 350   if (Sysroot && !Has("-isysroot") && !Has("--sysroot")) {
 351     ToAppend.push_back("-isysroot");
 352     ToAppend.push_back(*Sysroot);
 353   }
 354
 355   if (!ToAppend.empty()) {
 356     Cmd.insert(llvm::find(Cmd, "--"), std::make_move_iterator(ToAppend.begin()),
 357                std::make_move_iterator(ToAppend.end()));
 358   }
 359
 360   if (!Cmd.empty()) {
 361     bool FollowSymlink = !Has("-no-canonical-prefixes");
 362     Cmd.front() =
 363         (FollowSymlink ? ResolvedDrivers : ResolvedDriversNoFollow)
 364             .get(Cmd.front(), [&, this] {
 365               return resolveDriver(Cmd.front(), FollowSymlink, ClangPath);
 366             });
 367   }
 368 }
 369
 370 // ArgStripper implementation
 371 namespace {
 372
 373 // Determine total number of args consumed by this option.
 374 // Return answers for {Exact, Prefix} match. 0 means not allowed.
 375 std::pair<unsigned, unsigned> getArgCount(const llvm::opt::Option &Opt) {
 376   constexpr static unsigned Rest = 10000; // Should be all the rest!
 377   // Reference is llvm::opt::Option::acceptInternal()
 378   using llvm::opt::Option;
 379   switch (Opt.getKind()) {
 380   case Option::FlagClass:
 381     return {1, 0};
 382   case Option::JoinedClass:
 383   case Option::CommaJoinedClass:
 384     return {1, 1};
 385   case Option::GroupClass:
 386   case Option::InputClass:
 387   case Option::UnknownClass:
 388   case Option::ValuesClass:
 389     return {1, 0};
 390   case Option::JoinedAndSeparateClass:
 391     return {2, 2};
 392   case Option::SeparateClass:
 393     return {2, 0};
 394   case Option::MultiArgClass:
 395     return {1 + Opt.getNumArgs(), 0};
 396   case Option::JoinedOrSeparateClass:
 397     return {2, 1};
 398   case Option::RemainingArgsClass:
 399     return {Rest, 0};
 400   case Option::RemainingArgsJoinedClass:
 401     return {Rest, Rest};
 402   }
 403   llvm_unreachable("Unhandled option kind");
 404 }
 405
 406 // Flag-parsing mode, which affects which flags are available.
 407 enum DriverMode : unsigned char {
 408   DM_None = 0,
 409   DM_GCC = 1, // Default mode e.g. when invoked as 'clang'
 410   DM_CL = 2,  // MS CL.exe compatible mode e.g. when invoked as 'clang-cl'
 411   DM_CC1 = 4, // When invoked as 'clang -cc1' or after '-Xclang'
 412   DM_All = 7
 413 };
 414
 415 // Examine args list to determine if we're in GCC, CL-compatible, or cc1 mode.
 416 DriverMode getDriverMode(const std::vector<std::string> &Args) {
 417   DriverMode Mode = DM_GCC;
 418   llvm::StringRef Argv0 = Args.front();
 419   if (Argv0.ends_with_insensitive(".exe"))
 420     Argv0 = Argv0.drop_back(strlen(".exe"));
 421   if (Argv0.ends_with_insensitive("cl"))
 422     Mode = DM_CL;
 423   for (const llvm::StringRef Arg : Args) {
 424     if (Arg == "--driver-mode=cl") {
 425       Mode = DM_CL;
 426       break;
 427     }
 428     if (Arg == "-cc1") {
 429       Mode = DM_CC1;
 430       break;
 431     }
 432   }
 433   return Mode;
 434 }
 435
 436 // Returns the set of DriverModes where an option may be used.
 437 unsigned char getModes(const llvm::opt::Option &Opt) {
 438   unsigned char Result = DM_None;
 439   if (Opt.hasVisibilityFlag(driver::options::ClangOption))
 440     Result |= DM_GCC;
 441   if (Opt.hasVisibilityFlag(driver::options::CC1Option))
 442     Result |= DM_CC1;
 443   if (Opt.hasVisibilityFlag(driver::options::CLOption))
 444     Result |= DM_CL;
 445   return Result;
 446 }
 447
 448 } // namespace
 449
 450 llvm::ArrayRef<ArgStripper::Rule> ArgStripper::rulesFor(llvm::StringRef Arg) {
 451   // All the hard work is done once in a static initializer.
 452   // We compute a table containing strings to look for and #args to skip.
 453   // e.g. "-x" => {-x 2 args, -x* 1 arg, --language 2 args, --language=* 1 arg}
 454   using TableTy =
 455       llvm::StringMap<llvm::SmallVector<Rule, 4>, llvm::BumpPtrAllocator>;
 456   static TableTy *Table = [] {
 457     auto &DriverTable = driver::getDriverOptTable();
 458     using DriverID = clang::driver::options::ID;
 459
 460     // Collect sets of aliases, so we can treat -foo and -foo= as synonyms.
 461     // Conceptually a double-linked list: PrevAlias[I] -> I -> NextAlias[I].
 462     // If PrevAlias[I] is INVALID, then I is canonical.
 463     DriverID PrevAlias[DriverID::LastOption] = {DriverID::OPT_INVALID};
 464     DriverID NextAlias[DriverID::LastOption] = {DriverID::OPT_INVALID};
 465     auto AddAlias = [&](DriverID Self, DriverID T) {
 466       if (NextAlias[T]) {
 467         PrevAlias[NextAlias[T]] = Self;
 468         NextAlias[Self] = NextAlias[T];
 469       }
 470       PrevAlias[Self] = T;
 471       NextAlias[T] = Self;
 472     };
 473     // Also grab prefixes for each option, these are not fully exposed.
 474     llvm::ArrayRef<llvm::StringLiteral> Prefixes[DriverID::LastOption];
 475
 476 #define PREFIX(NAME, VALUE)                                                    \
 477   static constexpr llvm::StringLiteral NAME##_init[] = VALUE;                  \
 478   static constexpr llvm::ArrayRef<llvm::StringLiteral> NAME(                   \
 479       NAME##_init, std::size(NAME##_init) - 1);
 480 #define OPTION(PREFIX, PREFIXED_NAME, ID, KIND, GROUP, ALIAS, ALIASARGS,       \
 481                FLAGS, VISIBILITY, PARAM, HELP, METAVAR, VALUES)                \
 482   Prefixes[DriverID::OPT_##ID] = PREFIX;
 483 #include "clang/Driver/Options.inc"
 484 #undef OPTION
 485 #undef PREFIX
 486
 487     struct {
 488       DriverID ID;
 489       DriverID AliasID;
 490       const void *AliasArgs;
 491     } AliasTable[] = {
 492 #define OPTION(PREFIX, PREFIXED_NAME, ID, KIND, GROUP, ALIAS, ALIASARGS,       \
 493                FLAGS, VISIBILITY, PARAM, HELP, METAVAR, VALUES)                \
 494   {DriverID::OPT_##ID, DriverID::OPT_##ALIAS, ALIASARGS},
 495 #include "clang/Driver/Options.inc"
 496 #undef OPTION
 497     };
 498     for (auto &E : AliasTable)
 499       if (E.AliasID != DriverID::OPT_INVALID && E.AliasArgs == nullptr)
 500         AddAlias(E.ID, E.AliasID);
 501
 502     auto Result = std::make_unique<TableTy>();
 503     // Iterate over distinct options (represented by the canonical alias).
 504     // Every spelling of this option will get the same set of rules.
 505     for (unsigned ID = 1 /*Skip INVALID */; ID < DriverID::LastOption; ++ID) {
 506       if (PrevAlias[ID] || ID == DriverID::OPT_Xclang)
 507         continue; // Not canonical, or specially handled.
 508       llvm::SmallVector<Rule> Rules;
 509       // Iterate over each alias, to add rules for parsing it.
 510       for (unsigned A = ID; A != DriverID::OPT_INVALID; A = NextAlias[A]) {
 511         if (!Prefixes[A].size()) // option groups.
 512           continue;
 513         auto Opt = DriverTable.getOption(A);
 514         // Exclude - and -foo pseudo-options.
 515         if (Opt.getName().empty())
 516           continue;
 517         auto Modes = getModes(Opt);
 518         std::pair<unsigned, unsigned> ArgCount = getArgCount(Opt);
 519         // Iterate over each spelling of the alias, e.g. -foo vs --foo.
 520         for (StringRef Prefix : Prefixes[A]) {
 521           llvm::SmallString<64> Buf(Prefix);
 522           Buf.append(Opt.getName());
 523           llvm::StringRef Spelling = Result->try_emplace(Buf).first->getKey();
 524           Rules.emplace_back();
 525           Rule &R = Rules.back();
 526           R.Text = Spelling;
 527           R.Modes = Modes;
 528           R.ExactArgs = ArgCount.first;
 529           R.PrefixArgs = ArgCount.second;
 530           // Concrete priority is the index into the option table.
 531           // Effectively, earlier entries take priority over later ones.
 532           assert(ID < std::numeric_limits<decltype(R.Priority)>::max() &&
 533                  "Rules::Priority overflowed by options table");
 534           R.Priority = ID;
 535         }
 536       }
 537       // Register the set of rules under each possible name.
 538       for (const auto &R : Rules)
 539         Result->find(R.Text)->second.append(Rules.begin(), Rules.end());
 540     }
 541 #ifndef NDEBUG
 542     // Dump the table and various measures of its size.
 543     unsigned RuleCount = 0;
 544     dlog("ArgStripper Option spelling table");
 545     for (const auto &Entry : *Result) {
 546       dlog("{0}", Entry.first());
 547       RuleCount += Entry.second.size();
 548       for (const auto &R : Entry.second)
 549         dlog("  {0} #={1} *={2} Mode={3}", R.Text, R.ExactArgs, R.PrefixArgs,
 550              int(R.Modes));
 551     }
 552     dlog("Table spellings={0} rules={1} string-bytes={2}", Result->size(),
 553          RuleCount, Result->getAllocator().getBytesAllocated());
 554 #endif
 555     // The static table will never be destroyed.
 556     return Result.release();
 557   }();
 558
 559   auto It = Table->find(Arg);
 560   return (It == Table->end()) ? llvm::ArrayRef<Rule>() : It->second;
 561 }
 562
 563 void ArgStripper::strip(llvm::StringRef Arg) {
 564   auto OptionRules = rulesFor(Arg);
 565   if (OptionRules.empty()) {
 566     // Not a recognized flag. Strip it literally.
 567     Storage.emplace_back(Arg);
 568     Rules.emplace_back();
 569     Rules.back().Text = Storage.back();
 570     Rules.back().ExactArgs = 1;
 571     if (Rules.back().Text.consume_back("*"))
 572       Rules.back().PrefixArgs = 1;
 573     Rules.back().Modes = DM_All;
 574     Rules.back().Priority = -1; // Max unsigned = lowest priority.
 575   } else {
 576     Rules.append(OptionRules.begin(), OptionRules.end());
 577   }
 578 }
 579
 580 const ArgStripper::Rule *ArgStripper::matchingRule(llvm::StringRef Arg,
 581                                                    unsigned Mode,
 582                                                    unsigned &ArgCount) const {
 583   const ArgStripper::Rule *BestRule = nullptr;
 584   for (const Rule &R : Rules) {
 585     // Rule can fail to match if...
 586     if (!(R.Modes & Mode))
 587       continue; // not applicable to current driver mode
 588     if (BestRule && BestRule->Priority < R.Priority)
 589       continue; // lower-priority than best candidate.
 590     if (!Arg.startswith(R.Text))
 591       continue; // current arg doesn't match the prefix string
 592     bool PrefixMatch = Arg.size() > R.Text.size();
 593     // Can rule apply as an exact/prefix match?
 594     if (unsigned Count = PrefixMatch ? R.PrefixArgs : R.ExactArgs) {
 595       BestRule = &R;
 596       ArgCount = Count;
 597     }
 598     // Continue in case we find a higher-priority rule.
 599   }
 600   return BestRule;
 601 }
 602
 603 void ArgStripper::process(std::vector<std::string> &Args) const {
 604   if (Args.empty())
 605     return;
 606
 607   // We're parsing the args list in some mode (e.g. gcc-compatible) but may
 608   // temporarily switch to another mode with the -Xclang flag.
 609   DriverMode MainMode = getDriverMode(Args);
 610   DriverMode CurrentMode = MainMode;
 611
 612   // Read and write heads for in-place deletion.
 613   unsigned Read = 0, Write = 0;
 614   bool WasXclang = false;
 615   while (Read < Args.size()) {
 616     unsigned ArgCount = 0;
 617     if (matchingRule(Args[Read], CurrentMode, ArgCount)) {
 618       // Delete it and its args.
 619       if (WasXclang) {
 620         assert(Write > 0);
 621         --Write; // Drop previous -Xclang arg
 622         CurrentMode = MainMode;
 623         WasXclang = false;
 624       }
 625       // Advance to last arg. An arg may be foo or -Xclang foo.
 626       for (unsigned I = 1; Read < Args.size() && I < ArgCount; ++I) {
 627         ++Read;
 628         if (Read < Args.size() && Args[Read] == "-Xclang")
 629           ++Read;
 630       }
 631     } else {
 632       // No match, just copy the arg through.
 633       WasXclang = Args[Read] == "-Xclang";
 634       CurrentMode = WasXclang ? DM_CC1 : MainMode;
 635       if (Write != Read)
 636         Args[Write] = std::move(Args[Read]);
 637       ++Write;
 638     }
 639     ++Read;
 640   }
 641   Args.resize(Write);
 642 }
 643
 644 std::string printArgv(llvm::ArrayRef<llvm::StringRef> Args) {
 645   std::string Buf;
 646   llvm::raw_string_ostream OS(Buf);
 647   bool Sep = false;
 648   for (llvm::StringRef Arg : Args) {
 649     if (Sep)
 650       OS << ' ';
 651     Sep = true;
 652     if (llvm::all_of(Arg, llvm::isPrint) &&
 653         Arg.find_first_of(" \t\n\"\\") == llvm::StringRef::npos) {
 654       OS << Arg;
 655       continue;
 656     }
 657     OS << '"';
 658     OS.write_escaped(Arg, /*UseHexEscapes=*/true);
 659     OS << '"';
 660   }
 661   return std::move(OS.str());
 662 }
 663
 664 std::string printArgv(llvm::ArrayRef<std::string> Args) {
 665   std::vector<llvm::StringRef> Refs(Args.size());
 666   llvm::copy(Args, Refs.begin());
 667   return printArgv(Refs);
 668 }
 669
 670 } // namespace clangd
 671 } // namespace clang