src/llvm/include/llvm/Support/YAMLParser.h

   1 //===--- YAMLParser.h - Simple YAML parser --------------------------------===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file is distributed under the University of Illinois Open Source
   6 // License. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9 //
  10 //  This is a YAML 1.2 parser.
  11 //
  12 //  See http://www.yaml.org/spec/1.2/spec.html for the full standard.
  13 //
  14 //  This currently does not implement the following:
  15 //    * Multi-line literal folding.
  16 //    * Tag resolution.
  17 //    * UTF-16.
  18 //    * BOMs anywhere other than the first Unicode scalar value in the file.
  19 //
  20 //  The most important class here is Stream. This represents a YAML stream with
  21 //  0, 1, or many documents.
  22 //
  23 //  SourceMgr sm;
  24 //  StringRef input = getInput();
  25 //  yaml::Stream stream(input, sm);
  26 //
  27 //  for (yaml::document_iterator di = stream.begin(), de = stream.end();
  28 //       di != de; ++di) {
  29 //    yaml::Node *n = di->getRoot();
  30 //    if (n) {
  31 //      // Do something with n...
  32 //    } else
  33 //      break;
  34 //  }
  35 //
  36 //===----------------------------------------------------------------------===//
  37
  38 #ifndef LLVM_SUPPORT_YAMLPARSER_H
  39 #define LLVM_SUPPORT_YAMLPARSER_H
  40
  41 #include "llvm/ADT/SmallString.h"
  42 #include "llvm/ADT/StringRef.h"
  43 #include "llvm/Support/Allocator.h"
  44 #include "llvm/Support/MemoryBuffer.h"
  45 #include "llvm/Support/SMLoc.h"
  46 #include <limits>
  47 #include <map>
  48 #include <utility>
  49
  50 namespace llvm {
  51 class SourceMgr;
  52 class raw_ostream;
  53 class Twine;
  54
  55 namespace yaml {
  56
  57 class document_iterator;
  58 class Document;
  59 class Node;
  60 class Scanner;
  61 struct Token;
  62
  63 /// \brief Dump all the tokens in this stream to OS.
  64 /// \returns true if there was an error, false otherwise.
  65 bool dumpTokens(StringRef Input, raw_ostream &);
  66
  67 /// \brief Scans all tokens in input without outputting anything. This is used
  68 ///        for benchmarking the tokenizer.
  69 /// \returns true if there was an error, false otherwise.
  70 bool scanTokens(StringRef Input);
  71
  72 /// \brief Escape \a Input for a double quoted scalar.
  73 std::string escape(StringRef Input);
  74
  75 /// \brief This class represents a YAML stream potentially containing multiple
  76 ///        documents.
  77 class Stream {
  78 public:
  79   /// \brief This keeps a reference to the string referenced by \p Input.
  80   Stream(StringRef Input, SourceMgr &);
  81
  82   Stream(MemoryBufferRef InputBuffer, SourceMgr &);
  83   ~Stream();
  84
  85   document_iterator begin();
  86   document_iterator end();
  87   void skip();
  88   bool failed();
  89   bool validate() {
  90     skip();
  91     return !failed();
  92   }
  93
  94   void printError(Node *N, const Twine &Msg);
  95
  96 private:
  97   std::unique_ptr<Scanner> scanner;
  98   std::unique_ptr<Document> CurrentDoc;
  99
 100   friend class Document;
 101 };
 102
 103 /// \brief Abstract base class for all Nodes.
 104 class Node {
 105   virtual void anchor();
 106
 107 public:
 108   enum NodeKind {
 109     NK_Null,
 110     NK_Scalar,
 111     NK_KeyValue,
 112     NK_Mapping,
 113     NK_Sequence,
 114     NK_Alias
 115   };
 116
 117   Node(unsigned int Type, std::unique_ptr<Document> &, StringRef Anchor,
 118        StringRef Tag);
 119
 120   /// \brief Get the value of the anchor attached to this node. If it does not
 121   ///        have one, getAnchor().size() will be 0.
 122   StringRef getAnchor() const { return Anchor; }
 123
 124   /// \brief Get the tag as it was written in the document. This does not
 125   ///   perform tag resolution.
 126   StringRef getRawTag() const { return Tag; }
 127
 128   /// \brief Get the verbatium tag for a given Node. This performs tag resoluton
 129   ///   and substitution.
 130   std::string getVerbatimTag() const;
 131
 132   SMRange getSourceRange() const { return SourceRange; }
 133   void setSourceRange(SMRange SR) { SourceRange = SR; }
 134
 135   // These functions forward to Document and Scanner.
 136   Token &peekNext();
 137   Token getNext();
 138   Node *parseBlockNode();
 139   BumpPtrAllocator &getAllocator();
 140   void setError(const Twine &Message, Token &Location) const;
 141   bool failed() const;
 142
 143   virtual void skip() {}
 144
 145   unsigned int getType() const { return TypeID; }
 146
 147   void *operator new(size_t Size, BumpPtrAllocator &Alloc,
 148                      size_t Alignment = 16) throw() {
 149     return Alloc.Allocate(Size, Alignment);
 150   }
 151
 152   void operator delete(void *Ptr, BumpPtrAllocator &Alloc, size_t Size) throw() {
 153     Alloc.Deallocate(Ptr, Size);
 154   }
 155
 156 protected:
 157   std::unique_ptr<Document> &Doc;
 158   SMRange SourceRange;
 159
 160   void operator delete(void *) throw() {}
 161
 162   virtual ~Node() {}
 163
 164 private:
 165   unsigned int TypeID;
 166   StringRef Anchor;
 167   /// \brief The tag as typed in the document.
 168   StringRef Tag;
 169 };
 170
 171 /// \brief A null value.
 172 ///
 173 /// Example:
 174 ///   !!null null
 175 class NullNode : public Node {
 176   void anchor() override;
 177
 178 public:
 179   NullNode(std::unique_ptr<Document> &D)
 180       : Node(NK_Null, D, StringRef(), StringRef()) {}
 181
 182   static inline bool classof(const Node *N) { return N->getType() == NK_Null; }
 183 };
 184
 185 /// \brief A scalar node is an opaque datum that can be presented as a
 186 ///        series of zero or more Unicode scalar values.
 187 ///
 188 /// Example:
 189 ///   Adena
 190 class ScalarNode : public Node {
 191   void anchor() override;
 192
 193 public:
 194   ScalarNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
 195              StringRef Val)
 196       : Node(NK_Scalar, D, Anchor, Tag), Value(Val) {
 197     SMLoc Start = SMLoc::getFromPointer(Val.begin());
 198     SMLoc End = SMLoc::getFromPointer(Val.end());
 199     SourceRange = SMRange(Start, End);
 200   }
 201
 202   // Return Value without any escaping or folding or other fun YAML stuff. This
 203   // is the exact bytes that are contained in the file (after conversion to
 204   // utf8).
 205   StringRef getRawValue() const { return Value; }
 206
 207   /// \brief Gets the value of this node as a StringRef.
 208   ///
 209   /// \param Storage is used to store the content of the returned StringRef iff
 210   ///        it requires any modification from how it appeared in the source.
 211   ///        This happens with escaped characters and multi-line literals.
 212   StringRef getValue(SmallVectorImpl<char> &Storage) const;
 213
 214   static inline bool classof(const Node *N) {
 215     return N->getType() == NK_Scalar;
 216   }
 217
 218 private:
 219   StringRef Value;
 220
 221   StringRef unescapeDoubleQuoted(StringRef UnquotedValue,
 222                                  StringRef::size_type Start,
 223                                  SmallVectorImpl<char> &Storage) const;
 224 };
 225
 226 /// \brief A key and value pair. While not technically a Node under the YAML
 227 ///        representation graph, it is easier to treat them this way.
 228 ///
 229 /// TODO: Consider making this not a child of Node.
 230 ///
 231 /// Example:
 232 ///   Section: .text
 233 class KeyValueNode : public Node {
 234   void anchor() override;
 235
 236 public:
 237   KeyValueNode(std::unique_ptr<Document> &D)
 238       : Node(NK_KeyValue, D, StringRef(), StringRef()), Key(nullptr),
 239         Value(nullptr) {}
 240
 241   /// \brief Parse and return the key.
 242   ///
 243   /// This may be called multiple times.
 244   ///
 245   /// \returns The key, or nullptr if failed() == true.
 246   Node *getKey();
 247
 248   /// \brief Parse and return the value.
 249   ///
 250   /// This may be called multiple times.
 251   ///
 252   /// \returns The value, or nullptr if failed() == true.
 253   Node *getValue();
 254
 255   void skip() override {
 256     getKey()->skip();
 257     getValue()->skip();
 258   }
 259
 260   static inline bool classof(const Node *N) {
 261     return N->getType() == NK_KeyValue;
 262   }
 263
 264 private:
 265   Node *Key;
 266   Node *Value;
 267 };
 268
 269 /// \brief This is an iterator abstraction over YAML collections shared by both
 270 ///        sequences and maps.
 271 ///
 272 /// BaseT must have a ValueT* member named CurrentEntry and a member function
 273 /// increment() which must set CurrentEntry to 0 to create an end iterator.
 274 template <class BaseT, class ValueT>
 275 class basic_collection_iterator
 276     : public std::iterator<std::forward_iterator_tag, ValueT> {
 277 public:
 278   basic_collection_iterator() : Base(nullptr) {}
 279   basic_collection_iterator(BaseT *B) : Base(B) {}
 280
 281   ValueT *operator->() const {
 282     assert(Base && Base->CurrentEntry && "Attempted to access end iterator!");
 283     return Base->CurrentEntry;
 284   }
 285
 286   ValueT &operator*() const {
 287     assert(Base && Base->CurrentEntry &&
 288            "Attempted to dereference end iterator!");
 289     return *Base->CurrentEntry;
 290   }
 291
 292   operator ValueT *() const {
 293     assert(Base && Base->CurrentEntry && "Attempted to access end iterator!");
 294     return Base->CurrentEntry;
 295   }
 296
 297   bool operator!=(const basic_collection_iterator &Other) const {
 298     if (Base != Other.Base)
 299       return true;
 300     return (Base && Other.Base) &&
 301            Base->CurrentEntry != Other.Base->CurrentEntry;
 302   }
 303
 304   basic_collection_iterator &operator++() {
 305     assert(Base && "Attempted to advance iterator past end!");
 306     Base->increment();
 307     // Create an end iterator.
 308     if (!Base->CurrentEntry)
 309       Base = nullptr;
 310     return *this;
 311   }
 312
 313 private:
 314   BaseT *Base;
 315 };
 316
 317 // The following two templates are used for both MappingNode and Sequence Node.
 318 template <class CollectionType>
 319 typename CollectionType::iterator begin(CollectionType &C) {
 320   assert(C.IsAtBeginning && "You may only iterate over a collection once!");
 321   C.IsAtBeginning = false;
 322   typename CollectionType::iterator ret(&C);
 323   ++ret;
 324   return ret;
 325 }
 326
 327 template <class CollectionType> void skip(CollectionType &C) {
 328   // TODO: support skipping from the middle of a parsed collection ;/
 329   assert((C.IsAtBeginning || C.IsAtEnd) && "Cannot skip mid parse!");
 330   if (C.IsAtBeginning)
 331     for (typename CollectionType::iterator i = begin(C), e = C.end(); i != e;
 332          ++i)
 333       i->skip();
 334 }
 335
 336 /// \brief Represents a YAML map created from either a block map for a flow map.
 337 ///
 338 /// This parses the YAML stream as increment() is called.
 339 ///
 340 /// Example:
 341 ///   Name: _main
 342 ///   Scope: Global
 343 class MappingNode : public Node {
 344   void anchor() override;
 345
 346 public:
 347   enum MappingType {
 348     MT_Block,
 349     MT_Flow,
 350     MT_Inline ///< An inline mapping node is used for "[key: value]".
 351   };
 352
 353   MappingNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
 354               MappingType MT)
 355       : Node(NK_Mapping, D, Anchor, Tag), Type(MT), IsAtBeginning(true),
 356         IsAtEnd(false), CurrentEntry(nullptr) {}
 357
 358   friend class basic_collection_iterator<MappingNode, KeyValueNode>;
 359   typedef basic_collection_iterator<MappingNode, KeyValueNode> iterator;
 360   template <class T> friend typename T::iterator yaml::begin(T &);
 361   template <class T> friend void yaml::skip(T &);
 362
 363   iterator begin() { return yaml::begin(*this); }
 364
 365   iterator end() { return iterator(); }
 366
 367   void skip() override { yaml::skip(*this); }
 368
 369   static inline bool classof(const Node *N) {
 370     return N->getType() == NK_Mapping;
 371   }
 372
 373 private:
 374   MappingType Type;
 375   bool IsAtBeginning;
 376   bool IsAtEnd;
 377   KeyValueNode *CurrentEntry;
 378
 379   void increment();
 380 };
 381
 382 /// \brief Represents a YAML sequence created from either a block sequence for a
 383 ///        flow sequence.
 384 ///
 385 /// This parses the YAML stream as increment() is called.
 386 ///
 387 /// Example:
 388 ///   - Hello
 389 ///   - World
 390 class SequenceNode : public Node {
 391   void anchor() override;
 392
 393 public:
 394   enum SequenceType {
 395     ST_Block,
 396     ST_Flow,
 397     // Use for:
 398     //
 399     // key:
 400     // - val1
 401     // - val2
 402     //
 403     // As a BlockMappingEntry and BlockEnd are not created in this case.
 404     ST_Indentless
 405   };
 406
 407   SequenceNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
 408                SequenceType ST)
 409       : Node(NK_Sequence, D, Anchor, Tag), SeqType(ST), IsAtBeginning(true),
 410         IsAtEnd(false),
 411         WasPreviousTokenFlowEntry(true), // Start with an imaginary ','.
 412         CurrentEntry(nullptr) {}
 413
 414   friend class basic_collection_iterator<SequenceNode, Node>;
 415   typedef basic_collection_iterator<SequenceNode, Node> iterator;
 416   template <class T> friend typename T::iterator yaml::begin(T &);
 417   template <class T> friend void yaml::skip(T &);
 418
 419   void increment();
 420
 421   iterator begin() { return yaml::begin(*this); }
 422
 423   iterator end() { return iterator(); }
 424
 425   void skip() override { yaml::skip(*this); }
 426
 427   static inline bool classof(const Node *N) {
 428     return N->getType() == NK_Sequence;
 429   }
 430
 431 private:
 432   SequenceType SeqType;
 433   bool IsAtBeginning;
 434   bool IsAtEnd;
 435   bool WasPreviousTokenFlowEntry;
 436   Node *CurrentEntry;
 437 };
 438
 439 /// \brief Represents an alias to a Node with an anchor.
 440 ///
 441 /// Example:
 442 ///   *AnchorName
 443 class AliasNode : public Node {
 444   void anchor() override;
 445
 446 public:
 447   AliasNode(std::unique_ptr<Document> &D, StringRef Val)
 448       : Node(NK_Alias, D, StringRef(), StringRef()), Name(Val) {}
 449
 450   StringRef getName() const { return Name; }
 451   Node *getTarget();
 452
 453   static inline bool classof(const Node *N) { return N->getType() == NK_Alias; }
 454
 455 private:
 456   StringRef Name;
 457 };
 458
 459 /// \brief A YAML Stream is a sequence of Documents. A document contains a root
 460 ///        node.
 461 class Document {
 462 public:
 463   /// \brief Root for parsing a node. Returns a single node.
 464   Node *parseBlockNode();
 465
 466   Document(Stream &ParentStream);
 467
 468   /// \brief Finish parsing the current document and return true if there are
 469   ///        more. Return false otherwise.
 470   bool skip();
 471
 472   /// \brief Parse and return the root level node.
 473   Node *getRoot() {
 474     if (Root)
 475       return Root;
 476     return Root = parseBlockNode();
 477   }
 478
 479   const std::map<StringRef, StringRef> &getTagMap() const { return TagMap; }
 480
 481 private:
 482   friend class Node;
 483   friend class document_iterator;
 484
 485   /// \brief Stream to read tokens from.
 486   Stream &stream;
 487
 488   /// \brief Used to allocate nodes to. All are destroyed without calling their
 489   ///        destructor when the document is destroyed.
 490   BumpPtrAllocator NodeAllocator;
 491
 492   /// \brief The root node. Used to support skipping a partially parsed
 493   ///        document.
 494   Node *Root;
 495
 496   /// \brief Maps tag prefixes to their expansion.
 497   std::map<StringRef, StringRef> TagMap;
 498
 499   Token &peekNext();
 500   Token getNext();
 501   void setError(const Twine &Message, Token &Location) const;
 502   bool failed() const;
 503
 504   /// \brief Parse %BLAH directives and return true if any were encountered.
 505   bool parseDirectives();
 506
 507   /// \brief Parse %YAML
 508   void parseYAMLDirective();
 509
 510   /// \brief Parse %TAG
 511   void parseTAGDirective();
 512
 513   /// \brief Consume the next token and error if it is not \a TK.
 514   bool expectToken(int TK);
 515 };
 516
 517 /// \brief Iterator abstraction for Documents over a Stream.
 518 class document_iterator {
 519 public:
 520   document_iterator() : Doc(nullptr) {}
 521   document_iterator(std::unique_ptr<Document> &D) : Doc(&D) {}
 522
 523   bool operator==(const document_iterator &Other) {
 524     if (isAtEnd() || Other.isAtEnd())
 525       return isAtEnd() && Other.isAtEnd();
 526
 527     return Doc == Other.Doc;
 528   }
 529   bool operator!=(const document_iterator &Other) { return !(*this == Other); }
 530
 531   document_iterator operator++() {
 532     assert(Doc && "incrementing iterator past the end.");
 533     if (!(*Doc)->skip()) {
 534       Doc->reset(nullptr);
 535     } else {
 536       Stream &S = (*Doc)->stream;
 537       Doc->reset(new Document(S));
 538     }
 539     return *this;
 540   }
 541
 542   Document &operator*() { return *Doc->get(); }
 543
 544   std::unique_ptr<Document> &operator->() { return *Doc; }
 545
 546 private:
 547   bool isAtEnd() const { return !Doc || !*Doc; }
 548
 549   std::unique_ptr<Document> *Doc;
 550 };
 551
 552 } // End namespace yaml.
 553
 554 } // End namespace llvm.
 555
 556 #endif