1 //===--- YAMLParser.h - Simple YAML parser --------------------------------===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This is a YAML 1.2 parser.
12 // See http://www.yaml.org/spec/1.2/spec.html for the full standard.
14 // This currently does not implement the following:
15 // * Multi-line literal folding.
18 // * BOMs anywhere other than the first Unicode scalar value in the file.
20 // The most important class here is Stream. This represents a YAML stream with
21 // 0, 1, or many documents.
24 // StringRef input = getInput();
25 // yaml::Stream stream(input, sm);
27 // for (yaml::document_iterator di = stream.begin(), de = stream.end();
29 // yaml::Node *n = di->getRoot();
31 // // Do something with n...
36 //===----------------------------------------------------------------------===//
38 #ifndef LLVM_SUPPORT_YAMLPARSER_H
39 #define LLVM_SUPPORT_YAMLPARSER_H
41 #include "llvm/ADT/SmallString.h"
42 #include "llvm/ADT/StringRef.h"
43 #include "llvm/Support/Allocator.h"
44 #include "llvm/Support/SMLoc.h"
57 class document_iterator;
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 &);
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);
72 /// \brief Escape \a Input for a double quoted scalar.
73 std::string escape(StringRef Input);
75 /// \brief This class represents a YAML stream potentially containing multiple
79 /// \brief This keeps a reference to the string referenced by \p Input.
80 Stream(StringRef Input, SourceMgr &);
82 /// \brief This takes ownership of \p InputBuffer.
83 Stream(MemoryBuffer *InputBuffer, SourceMgr &);
86 document_iterator begin();
87 document_iterator end();
95 void printError(Node *N, const Twine &Msg);
98 std::unique_ptr<Scanner> scanner;
99 std::unique_ptr<Document> CurrentDoc;
101 friend class Document;
104 /// \brief Abstract base class for all Nodes.
106 virtual void anchor();
118 Node(unsigned int Type, std::unique_ptr<Document> &, StringRef Anchor,
121 /// \brief Get the value of the anchor attached to this node. If it does not
122 /// have one, getAnchor().size() will be 0.
123 StringRef getAnchor() const { return Anchor; }
125 /// \brief Get the tag as it was written in the document. This does not
126 /// perform tag resolution.
127 StringRef getRawTag() const { return Tag; }
129 /// \brief Get the verbatium tag for a given Node. This performs tag resoluton
130 /// and substitution.
131 std::string getVerbatimTag() const;
133 SMRange getSourceRange() const { return SourceRange; }
134 void setSourceRange(SMRange SR) { SourceRange = SR; }
136 // These functions forward to Document and Scanner.
139 Node *parseBlockNode();
140 BumpPtrAllocator &getAllocator();
141 void setError(const Twine &Message, Token &Location) const;
144 virtual void skip() {}
146 unsigned int getType() const { return TypeID; }
148 void *operator new(size_t Size, BumpPtrAllocator &Alloc,
149 size_t Alignment = 16) throw() {
150 return Alloc.Allocate(Size, Alignment);
153 void operator delete(void *Ptr, BumpPtrAllocator &Alloc, size_t Size) throw() {
154 Alloc.Deallocate(Ptr, Size);
158 std::unique_ptr<Document> &Doc;
161 void operator delete(void *) throw() {}
168 /// \brief The tag as typed in the document.
172 /// \brief A null value.
176 class NullNode : public Node {
177 void anchor() override;
180 NullNode(std::unique_ptr<Document> &D)
181 : Node(NK_Null, D, StringRef(), StringRef()) {}
183 static inline bool classof(const Node *N) { return N->getType() == NK_Null; }
186 /// \brief A scalar node is an opaque datum that can be presented as a
187 /// series of zero or more Unicode scalar values.
191 class ScalarNode : public Node {
192 void anchor() override;
195 ScalarNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
197 : Node(NK_Scalar, D, Anchor, Tag), Value(Val) {
198 SMLoc Start = SMLoc::getFromPointer(Val.begin());
199 SMLoc End = SMLoc::getFromPointer(Val.end());
200 SourceRange = SMRange(Start, End);
203 // Return Value without any escaping or folding or other fun YAML stuff. This
204 // is the exact bytes that are contained in the file (after conversion to
206 StringRef getRawValue() const { return Value; }
208 /// \brief Gets the value of this node as a StringRef.
210 /// \param Storage is used to store the content of the returned StringRef iff
211 /// it requires any modification from how it appeared in the source.
212 /// This happens with escaped characters and multi-line literals.
213 StringRef getValue(SmallVectorImpl<char> &Storage) const;
215 static inline bool classof(const Node *N) {
216 return N->getType() == NK_Scalar;
222 StringRef unescapeDoubleQuoted(StringRef UnquotedValue,
223 StringRef::size_type Start,
224 SmallVectorImpl<char> &Storage) const;
227 /// \brief A key and value pair. While not technically a Node under the YAML
228 /// representation graph, it is easier to treat them this way.
230 /// TODO: Consider making this not a child of Node.
234 class KeyValueNode : public Node {
235 void anchor() override;
238 KeyValueNode(std::unique_ptr<Document> &D)
239 : Node(NK_KeyValue, D, StringRef(), StringRef()), Key(nullptr),
242 /// \brief Parse and return the key.
244 /// This may be called multiple times.
246 /// \returns The key, or nullptr if failed() == true.
249 /// \brief Parse and return the value.
251 /// This may be called multiple times.
253 /// \returns The value, or nullptr if failed() == true.
256 void skip() override {
261 static inline bool classof(const Node *N) {
262 return N->getType() == NK_KeyValue;
270 /// \brief This is an iterator abstraction over YAML collections shared by both
271 /// sequences and maps.
273 /// BaseT must have a ValueT* member named CurrentEntry and a member function
274 /// increment() which must set CurrentEntry to 0 to create an end iterator.
275 template <class BaseT, class ValueT>
276 class basic_collection_iterator
277 : public std::iterator<std::forward_iterator_tag, ValueT> {
279 basic_collection_iterator() : Base(nullptr) {}
280 basic_collection_iterator(BaseT *B) : Base(B) {}
282 ValueT *operator->() const {
283 assert(Base && Base->CurrentEntry && "Attempted to access end iterator!");
284 return Base->CurrentEntry;
287 ValueT &operator*() const {
288 assert(Base && Base->CurrentEntry &&
289 "Attempted to dereference end iterator!");
290 return *Base->CurrentEntry;
293 operator ValueT *() const {
294 assert(Base && Base->CurrentEntry && "Attempted to access end iterator!");
295 return Base->CurrentEntry;
298 bool operator!=(const basic_collection_iterator &Other) const {
299 if (Base != Other.Base)
301 return (Base && Other.Base) &&
302 Base->CurrentEntry != Other.Base->CurrentEntry;
305 basic_collection_iterator &operator++() {
306 assert(Base && "Attempted to advance iterator past end!");
308 // Create an end iterator.
309 if (!Base->CurrentEntry)
318 // The following two templates are used for both MappingNode and Sequence Node.
319 template <class CollectionType>
320 typename CollectionType::iterator begin(CollectionType &C) {
321 assert(C.IsAtBeginning && "You may only iterate over a collection once!");
322 C.IsAtBeginning = false;
323 typename CollectionType::iterator ret(&C);
328 template <class CollectionType> void skip(CollectionType &C) {
329 // TODO: support skipping from the middle of a parsed collection ;/
330 assert((C.IsAtBeginning || C.IsAtEnd) && "Cannot skip mid parse!");
332 for (typename CollectionType::iterator i = begin(C), e = C.end(); i != e;
337 /// \brief Represents a YAML map created from either a block map for a flow map.
339 /// This parses the YAML stream as increment() is called.
344 class MappingNode : public Node {
345 void anchor() override;
351 MT_Inline ///< An inline mapping node is used for "[key: value]".
354 MappingNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
356 : Node(NK_Mapping, D, Anchor, Tag), Type(MT), IsAtBeginning(true),
357 IsAtEnd(false), CurrentEntry(nullptr) {}
359 friend class basic_collection_iterator<MappingNode, KeyValueNode>;
360 typedef basic_collection_iterator<MappingNode, KeyValueNode> iterator;
361 template <class T> friend typename T::iterator yaml::begin(T &);
362 template <class T> friend void yaml::skip(T &);
364 iterator begin() { return yaml::begin(*this); }
366 iterator end() { return iterator(); }
368 void skip() override { yaml::skip(*this); }
370 static inline bool classof(const Node *N) {
371 return N->getType() == NK_Mapping;
378 KeyValueNode *CurrentEntry;
383 /// \brief Represents a YAML sequence created from either a block sequence for a
386 /// This parses the YAML stream as increment() is called.
391 class SequenceNode : public Node {
392 void anchor() override;
404 // As a BlockMappingEntry and BlockEnd are not created in this case.
408 SequenceNode(std::unique_ptr<Document> &D, StringRef Anchor, StringRef Tag,
410 : Node(NK_Sequence, D, Anchor, Tag), SeqType(ST), IsAtBeginning(true),
412 WasPreviousTokenFlowEntry(true), // Start with an imaginary ','.
413 CurrentEntry(nullptr) {}
415 friend class basic_collection_iterator<SequenceNode, Node>;
416 typedef basic_collection_iterator<SequenceNode, Node> iterator;
417 template <class T> friend typename T::iterator yaml::begin(T &);
418 template <class T> friend void yaml::skip(T &);
422 iterator begin() { return yaml::begin(*this); }
424 iterator end() { return iterator(); }
426 void skip() override { yaml::skip(*this); }
428 static inline bool classof(const Node *N) {
429 return N->getType() == NK_Sequence;
433 SequenceType SeqType;
436 bool WasPreviousTokenFlowEntry;
440 /// \brief Represents an alias to a Node with an anchor.
444 class AliasNode : public Node {
445 void anchor() override;
448 AliasNode(std::unique_ptr<Document> &D, StringRef Val)
449 : Node(NK_Alias, D, StringRef(), StringRef()), Name(Val) {}
451 StringRef getName() const { return Name; }
454 static inline bool classof(const Node *N) { return N->getType() == NK_Alias; }
460 /// \brief A YAML Stream is a sequence of Documents. A document contains a root
464 /// \brief Root for parsing a node. Returns a single node.
465 Node *parseBlockNode();
467 Document(Stream &ParentStream);
469 /// \brief Finish parsing the current document and return true if there are
470 /// more. Return false otherwise.
473 /// \brief Parse and return the root level node.
477 return Root = parseBlockNode();
480 const std::map<StringRef, StringRef> &getTagMap() const { return TagMap; }
484 friend class document_iterator;
486 /// \brief Stream to read tokens from.
489 /// \brief Used to allocate nodes to. All are destroyed without calling their
490 /// destructor when the document is destroyed.
491 BumpPtrAllocator NodeAllocator;
493 /// \brief The root node. Used to support skipping a partially parsed
497 /// \brief Maps tag prefixes to their expansion.
498 std::map<StringRef, StringRef> TagMap;
502 void setError(const Twine &Message, Token &Location) const;
505 /// \brief Parse %BLAH directives and return true if any were encountered.
506 bool parseDirectives();
508 /// \brief Parse %YAML
509 void parseYAMLDirective();
511 /// \brief Parse %TAG
512 void parseTAGDirective();
514 /// \brief Consume the next token and error if it is not \a TK.
515 bool expectToken(int TK);
518 /// \brief Iterator abstraction for Documents over a Stream.
519 class document_iterator {
521 document_iterator() : Doc(nullptr) {}
522 document_iterator(std::unique_ptr<Document> &D) : Doc(&D) {}
524 bool operator==(const document_iterator &Other) {
525 if (isAtEnd() || Other.isAtEnd())
526 return isAtEnd() && Other.isAtEnd();
528 return Doc == Other.Doc;
530 bool operator!=(const document_iterator &Other) { return !(*this == Other); }
532 document_iterator operator++() {
533 assert(Doc && "incrementing iterator past the end.");
534 if (!(*Doc)->skip()) {
537 Stream &S = (*Doc)->stream;
538 Doc->reset(new Document(S));
543 Document &operator*() { return *Doc->get(); }
545 std::unique_ptr<Document> &operator->() { return *Doc; }
548 bool isAtEnd() const { return !Doc || !*Doc; }
550 std::unique_ptr<Document> *Doc;
553 } // End namespace yaml.
555 } // End namespace llvm.