Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 1 | //===--- ConfigYAML.cpp - Loading configuration fragments from YAML files -===// |
| 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 | //===----------------------------------------------------------------------===// |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 8 | #include "ConfigFragment.h" |
Sam McCall | dbf486c | 2020-07-14 15:17:16 +0200 | [diff] [blame] | 9 | #include "llvm/ADT/Optional.h" |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 10 | #include "llvm/ADT/SmallSet.h" |
| 11 | #include "llvm/ADT/StringRef.h" |
| 12 | #include "llvm/Support/MemoryBuffer.h" |
| 13 | #include "llvm/Support/SourceMgr.h" |
| 14 | #include "llvm/Support/YAMLParser.h" |
| 15 | #include <system_error> |
| 16 | |
| 17 | namespace clang { |
| 18 | namespace clangd { |
| 19 | namespace config { |
| 20 | namespace { |
| 21 | using llvm::yaml::BlockScalarNode; |
| 22 | using llvm::yaml::MappingNode; |
| 23 | using llvm::yaml::Node; |
| 24 | using llvm::yaml::ScalarNode; |
| 25 | using llvm::yaml::SequenceNode; |
| 26 | |
| 27 | class Parser { |
| 28 | llvm::SourceMgr &SM; |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 29 | bool HadError = false; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 30 | |
| 31 | public: |
| 32 | Parser(llvm::SourceMgr &SM) : SM(SM) {} |
| 33 | |
| 34 | // Tries to parse N into F, returning false if it failed and we couldn't |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 35 | // meaningfully recover (YAML syntax error, or hard semantic error). |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 36 | bool parse(Fragment &F, Node &N) { |
| 37 | DictParser Dict("Config", this); |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 38 | Dict.handle("If", [&](Node &N) { parse(F.If, N); }); |
| 39 | Dict.handle("CompileFlags", [&](Node &N) { parse(F.CompileFlags, N); }); |
Adam Czachorowski | 7029e5d | 2020-09-15 20:13:00 +0200 | [diff] [blame] | 40 | Dict.handle("Index", [&](Node &N) { parse(F.Index, N); }); |
Adam Czachorowski | c894bfd | 2020-09-15 19:47:50 +0200 | [diff] [blame] | 41 | Dict.handle("Style", [&](Node &N) { parse(F.Style, N); }); |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 42 | Dict.handle("ClangTidy", [&](Node &N) { parse(F.ClangTidy, N); }); |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 43 | Dict.parse(N); |
| 44 | return !(N.failed() || HadError); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 45 | } |
| 46 | |
| 47 | private: |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 48 | void parse(Fragment::IfBlock &F, Node &N) { |
Sam McCall | f12cd99 | 2020-06-26 01:49:53 +0200 | [diff] [blame] | 49 | DictParser Dict("If", this); |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 50 | Dict.unrecognized([&](Located<std::string>, Node &) { |
| 51 | F.HasUnrecognizedCondition = true; |
| 52 | return true; // Emit a warning for the unrecognized key. |
| 53 | }); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 54 | Dict.handle("PathMatch", [&](Node &N) { |
| 55 | if (auto Values = scalarValues(N)) |
| 56 | F.PathMatch = std::move(*Values); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 57 | }); |
Sam McCall | 86f1313 | 2020-07-09 23:33:46 +0200 | [diff] [blame] | 58 | Dict.handle("PathExclude", [&](Node &N) { |
| 59 | if (auto Values = scalarValues(N)) |
| 60 | F.PathExclude = std::move(*Values); |
| 61 | }); |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 62 | Dict.parse(N); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 63 | } |
| 64 | |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 65 | void parse(Fragment::CompileFlagsBlock &F, Node &N) { |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 66 | DictParser Dict("CompileFlags", this); |
| 67 | Dict.handle("Add", [&](Node &N) { |
| 68 | if (auto Values = scalarValues(N)) |
| 69 | F.Add = std::move(*Values); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 70 | }); |
Sam McCall | 6c16fbd | 2020-07-13 20:37:54 +0200 | [diff] [blame] | 71 | Dict.handle("Remove", [&](Node &N) { |
| 72 | if (auto Values = scalarValues(N)) |
| 73 | F.Remove = std::move(*Values); |
| 74 | }); |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 75 | Dict.parse(N); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 76 | } |
| 77 | |
Adam Czachorowski | c894bfd | 2020-09-15 19:47:50 +0200 | [diff] [blame] | 78 | void parse(Fragment::StyleBlock &F, Node &N) { |
| 79 | DictParser Dict("Style", this); |
| 80 | Dict.handle("FullyQualifiedNamespaces", [&](Node &N) { |
| 81 | if (auto Values = scalarValues(N)) |
| 82 | F.FullyQualifiedNamespaces = std::move(*Values); |
| 83 | }); |
| 84 | Dict.parse(N); |
| 85 | } |
| 86 | |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 87 | void parse(Fragment::ClangTidyBlock &F, Node &N) { |
| 88 | DictParser Dict("ClangTidy", this); |
| 89 | Dict.handle("Add", [&](Node &N) { |
| 90 | if (auto Values = scalarValues(N)) |
| 91 | F.Add = std::move(*Values); |
| 92 | }); |
| 93 | Dict.handle("Remove", [&](Node &N) { |
| 94 | if (auto Values = scalarValues(N)) |
| 95 | F.Remove = std::move(*Values); |
| 96 | }); |
| 97 | Dict.handle("CheckOptions", [&](Node &N) { |
| 98 | DictParser CheckOptDict("CheckOptions", this); |
| 99 | CheckOptDict.unrecognized([&](Located<std::string> &&Key, Node &Val) { |
| 100 | if (auto Value = scalarValue(Val, *Key)) |
| 101 | F.CheckOptions.emplace_back(std::move(Key), std::move(*Value)); |
| 102 | return false; // Don't emit a warning |
| 103 | }); |
| 104 | CheckOptDict.parse(N); |
| 105 | }); |
| 106 | Dict.parse(N); |
| 107 | } |
| 108 | |
Sam McCall | dbf486c | 2020-07-14 15:17:16 +0200 | [diff] [blame] | 109 | void parse(Fragment::IndexBlock &F, Node &N) { |
| 110 | DictParser Dict("Index", this); |
| 111 | Dict.handle("Background", |
| 112 | [&](Node &N) { F.Background = scalarValue(N, "Background"); }); |
Kadir Cetinkaya | 359e2f9 | 2020-10-01 12:29:55 +0100 | [diff] [blame] | 113 | Dict.handle("External", [&](Node &N) { |
| 114 | Fragment::IndexBlock::ExternalBlock External; |
| 115 | parse(External, N); |
| 116 | F.External.emplace(std::move(External)); |
| 117 | F.External->Range = N.getSourceRange(); |
| 118 | }); |
| 119 | Dict.parse(N); |
| 120 | } |
| 121 | |
| 122 | void parse(Fragment::IndexBlock::ExternalBlock &F, Node &N) { |
| 123 | DictParser Dict("External", this); |
| 124 | Dict.handle("File", [&](Node &N) { F.File = scalarValue(N, "File"); }); |
| 125 | Dict.handle("Server", |
| 126 | [&](Node &N) { F.Server = scalarValue(N, "Server"); }); |
| 127 | Dict.handle("MountPoint", |
| 128 | [&](Node &N) { F.MountPoint = scalarValue(N, "MountPoint"); }); |
Sam McCall | dbf486c | 2020-07-14 15:17:16 +0200 | [diff] [blame] | 129 | Dict.parse(N); |
| 130 | } |
| 131 | |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 132 | // Helper for parsing mapping nodes (dictionaries). |
| 133 | // We don't use YamlIO as we want to control over unknown keys. |
| 134 | class DictParser { |
| 135 | llvm::StringRef Description; |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 136 | std::vector<std::pair<llvm::StringRef, std::function<void(Node &)>>> Keys; |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 137 | std::function<bool(Located<std::string>, Node &)> UnknownHandler; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 138 | Parser *Outer; |
| 139 | |
| 140 | public: |
| 141 | DictParser(llvm::StringRef Description, Parser *Outer) |
| 142 | : Description(Description), Outer(Outer) {} |
| 143 | |
| 144 | // Parse is called when Key is encountered, and passed the associated value. |
| 145 | // It should emit diagnostics if the value is invalid (e.g. wrong type). |
| 146 | // If Key is seen twice, Parse runs only once and an error is reported. |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 147 | void handle(llvm::StringLiteral Key, std::function<void(Node &)> Parse) { |
Tres Popp | 1a30eab | 2020-06-26 10:12:04 +0200 | [diff] [blame] | 148 | for (const auto &Entry : Keys) { |
| 149 | (void) Entry; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 150 | assert(Entry.first != Key && "duplicate key handler"); |
Tres Popp | 1a30eab | 2020-06-26 10:12:04 +0200 | [diff] [blame] | 151 | } |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 152 | Keys.emplace_back(Key, std::move(Parse)); |
| 153 | } |
| 154 | |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 155 | // Handler is called when a Key is not matched by any handle(). |
| 156 | // If this is unset or the Handler returns true, a warning is emitted for |
| 157 | // the unknown key. |
| 158 | void |
| 159 | unrecognized(std::function<bool(Located<std::string>, Node &)> Handler) { |
| 160 | UnknownHandler = std::move(Handler); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 161 | } |
| 162 | |
| 163 | // Process a mapping node and call handlers for each key/value pair. |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 164 | void parse(Node &N) const { |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 165 | if (N.getType() != Node::NK_Mapping) { |
| 166 | Outer->error(Description + " should be a dictionary", N); |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 167 | return; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 168 | } |
| 169 | llvm::SmallSet<std::string, 8> Seen; |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 170 | // We *must* consume all items, even on error, or the parser will assert. |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 171 | for (auto &KV : llvm::cast<MappingNode>(N)) { |
| 172 | auto *K = KV.getKey(); |
| 173 | if (!K) // YAMLParser emitted an error. |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 174 | continue; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 175 | auto Key = Outer->scalarValue(*K, "Dictionary key"); |
| 176 | if (!Key) |
| 177 | continue; |
| 178 | if (!Seen.insert(**Key).second) { |
| 179 | Outer->warning("Duplicate key " + **Key + " is ignored", *K); |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 180 | if (auto *Value = KV.getValue()) |
| 181 | Value->skip(); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 182 | continue; |
| 183 | } |
| 184 | auto *Value = KV.getValue(); |
| 185 | if (!Value) // YAMLParser emitted an error. |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 186 | continue; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 187 | bool Matched = false; |
| 188 | for (const auto &Handler : Keys) { |
| 189 | if (Handler.first == **Key) { |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 190 | Matched = true; |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 191 | Handler.second(*Value); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 192 | break; |
| 193 | } |
| 194 | } |
| 195 | if (!Matched) { |
Nathan James | 20b69af | 2020-11-22 10:04:00 +0000 | [diff] [blame] | 196 | bool Warn = !UnknownHandler; |
| 197 | if (UnknownHandler) |
| 198 | Warn = UnknownHandler( |
| 199 | Located<std::string>(**Key, K->getSourceRange()), *Value); |
| 200 | if (Warn) |
| 201 | Outer->warning("Unknown " + Description + " key " + **Key, *K); |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 202 | } |
| 203 | } |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 204 | } |
| 205 | }; |
| 206 | |
| 207 | // Try to parse a single scalar value from the node, warn on failure. |
| 208 | llvm::Optional<Located<std::string>> scalarValue(Node &N, |
| 209 | llvm::StringRef Desc) { |
| 210 | llvm::SmallString<256> Buf; |
| 211 | if (auto *S = llvm::dyn_cast<ScalarNode>(&N)) |
| 212 | return Located<std::string>(S->getValue(Buf).str(), N.getSourceRange()); |
| 213 | if (auto *BS = llvm::dyn_cast<BlockScalarNode>(&N)) |
| 214 | return Located<std::string>(BS->getValue().str(), N.getSourceRange()); |
| 215 | warning(Desc + " should be scalar", N); |
| 216 | return llvm::None; |
| 217 | } |
| 218 | |
| 219 | // Try to parse a list of single scalar values, or just a single value. |
| 220 | llvm::Optional<std::vector<Located<std::string>>> scalarValues(Node &N) { |
| 221 | std::vector<Located<std::string>> Result; |
| 222 | if (auto *S = llvm::dyn_cast<ScalarNode>(&N)) { |
| 223 | llvm::SmallString<256> Buf; |
| 224 | Result.emplace_back(S->getValue(Buf).str(), N.getSourceRange()); |
| 225 | } else if (auto *S = llvm::dyn_cast<BlockScalarNode>(&N)) { |
| 226 | Result.emplace_back(S->getValue().str(), N.getSourceRange()); |
| 227 | } else if (auto *S = llvm::dyn_cast<SequenceNode>(&N)) { |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 228 | // We *must* consume all items, even on error, or the parser will assert. |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 229 | for (auto &Child : *S) { |
| 230 | if (auto Value = scalarValue(Child, "List item")) |
| 231 | Result.push_back(std::move(*Value)); |
| 232 | } |
| 233 | } else { |
| 234 | warning("Expected scalar or list of scalars", N); |
| 235 | return llvm::None; |
| 236 | } |
| 237 | return Result; |
| 238 | } |
| 239 | |
| 240 | // Report a "hard" error, reflecting a config file that can never be valid. |
| 241 | void error(const llvm::Twine &Msg, const Node &N) { |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 242 | HadError = true; |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 243 | SM.PrintMessage(N.getSourceRange().Start, llvm::SourceMgr::DK_Error, Msg, |
| 244 | N.getSourceRange()); |
| 245 | } |
| 246 | |
| 247 | // Report a "soft" error that could be caused by e.g. version skew. |
| 248 | void warning(const llvm::Twine &Msg, const Node &N) { |
| 249 | SM.PrintMessage(N.getSourceRange().Start, llvm::SourceMgr::DK_Warning, Msg, |
| 250 | N.getSourceRange()); |
| 251 | } |
| 252 | }; |
| 253 | |
| 254 | } // namespace |
| 255 | |
| 256 | std::vector<Fragment> Fragment::parseYAML(llvm::StringRef YAML, |
| 257 | llvm::StringRef BufferName, |
| 258 | DiagnosticCallback Diags) { |
| 259 | // The YAML document may contain multiple conditional fragments. |
| 260 | // The SourceManager is shared for all of them. |
| 261 | auto SM = std::make_shared<llvm::SourceMgr>(); |
| 262 | auto Buf = llvm::MemoryBuffer::getMemBufferCopy(YAML, BufferName); |
| 263 | // Adapt DiagnosticCallback to function-pointer interface. |
| 264 | // Callback receives both errors we emit and those from the YAML parser. |
| 265 | SM->setDiagHandler( |
| 266 | [](const llvm::SMDiagnostic &Diag, void *Ctx) { |
| 267 | (*reinterpret_cast<DiagnosticCallback *>(Ctx))(Diag); |
| 268 | }, |
| 269 | &Diags); |
| 270 | std::vector<Fragment> Result; |
| 271 | for (auto &Doc : llvm::yaml::Stream(*Buf, *SM)) { |
Sam McCall | f365186 | 2020-07-09 00:13:54 +0200 | [diff] [blame] | 272 | if (Node *N = Doc.getRoot()) { |
Sam McCall | e9fb150 | 2020-06-23 17:21:56 +0200 | [diff] [blame] | 273 | Fragment Fragment; |
| 274 | Fragment.Source.Manager = SM; |
| 275 | Fragment.Source.Location = N->getSourceRange().Start; |
| 276 | if (Parser(*SM).parse(Fragment, *N)) |
| 277 | Result.push_back(std::move(Fragment)); |
| 278 | } |
| 279 | } |
| 280 | // Hack: stash the buffer in the SourceMgr to keep it alive. |
| 281 | // SM has two entries: "main" non-owning buffer, and ignored owning buffer. |
| 282 | SM->AddNewSourceBuffer(std::move(Buf), llvm::SMLoc()); |
| 283 | return Result; |
| 284 | } |
| 285 | |
| 286 | } // namespace config |
| 287 | } // namespace clangd |
| 288 | } // namespace clang |