X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FSupport%2FYAMLParser.cpp;h=73ce5e009be49db530fb296e0ae98953a0b8975d;hb=2ca626570fa1f2837bd0cea06303ed46890c6480;hp=2ae33f5b99634f6bb181e9f260aa805c4d19f0fc;hpb=525398e1376b0d61fe80a83ed2261e2e6e6ab4a3;p=oota-llvm.git diff --git a/lib/Support/YAMLParser.cpp b/lib/Support/YAMLParser.cpp index 2ae33f5b996..73ce5e009be 100644 --- a/lib/Support/YAMLParser.cpp +++ b/lib/Support/YAMLParser.cpp @@ -12,16 +12,15 @@ //===----------------------------------------------------------------------===// #include "llvm/Support/YAMLParser.h" - -#include "llvm/ADT/ilist.h" -#include "llvm/ADT/ilist_node.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringExtras.h" #include "llvm/ADT/Twine.h" +#include "llvm/ADT/ilist.h" +#include "llvm/ADT/ilist_node.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/MemoryBuffer.h" -#include "llvm/Support/raw_ostream.h" #include "llvm/Support/SourceMgr.h" +#include "llvm/Support/raw_ostream.h" using namespace llvm; using namespace yaml; @@ -97,6 +96,15 @@ static EncodingInfo getUnicodeEncoding(StringRef Input) { namespace llvm { namespace yaml { +/// Pin the vtables to this file. +void Node::anchor() {} +void NullNode::anchor() {} +void ScalarNode::anchor() {} +void KeyValueNode::anchor() {} +void MappingNode::anchor() {} +void SequenceNode::anchor() {} +void AliasNode::anchor() {} + /// Token - A single YAML token. struct Token : ilist_node { enum TokenKind { @@ -261,7 +269,7 @@ public: Token getNext(); void printError(SMLoc Loc, SourceMgr::DiagKind Kind, const Twine &Message, - ArrayRef Ranges = ArrayRef()) { + ArrayRef Ranges = None) { SM.PrintMessage(Loc, Kind, Message, Ranges); } @@ -370,9 +378,6 @@ private: /// sequence of ns-uri-char. StringRef scan_ns_uri_char(); - /// @brief Scan ns-plain-one-line[133] starting at \a Cur. - StringRef scan_ns_plain_one_line(); - /// @brief Consume a minimal well-formed code unit subsequence starting at /// \a Cur. Return false if it is not the same Unicode scalar value as /// \a Expected. This updates \a Column. @@ -865,42 +870,6 @@ StringRef Scanner::scan_ns_uri_char() { return StringRef(Start, Current - Start); } -StringRef Scanner::scan_ns_plain_one_line() { - StringRef::iterator start = Current; - // The first character must already be verified. - ++Current; - while (true) { - if (Current == End) { - break; - } else if (*Current == ':') { - // Check if the next character is a ns-char. - if (Current + 1 == End) - break; - StringRef::iterator i = skip_ns_char(Current + 1); - if (Current + 1 != i) { - Current = i; - Column += 2; // Consume both the ':' and ns-char. - } else - break; - } else if (*Current == '#') { - // Check if the previous character was a ns-char. - // The & 0x80 check is to check for the trailing byte of a utf-8 - if (*(Current - 1) & 0x80 || skip_ns_char(Current - 1) == Current) { - ++Current; - ++Column; - } else - break; - } else { - StringRef::iterator i = skip_nb_char(Current); - if (i == Current) - break; - Current = i; - ++Column; - } - } - return StringRef(start, Current - start); -} - bool Scanner::consume(uint32_t Expected) { if (Expected >= 0x80) report_fatal_error("Not dealing with this yet"); @@ -1071,14 +1040,22 @@ bool Scanner::scanDirective() { Current = skip_while(&Scanner::skip_ns_char, Current); StringRef Name(NameStart, Current - NameStart); Current = skip_while(&Scanner::skip_s_white, Current); - + + Token T; if (Name == "YAML") { Current = skip_while(&Scanner::skip_ns_char, Current); - Token T; T.Kind = Token::TK_VersionDirective; T.Range = StringRef(Start, Current - Start); TokenQueue.push_back(T); return true; + } else if(Name == "TAG") { + Current = skip_while(&Scanner::skip_ns_char, Current); + Current = skip_while(&Scanner::skip_s_white, Current); + Current = skip_while(&Scanner::skip_ns_char, Current); + T.Kind = Token::TK_TagDirective; + T.Range = StringRef(Start, Current - Start); + TokenQueue.push_back(T); + return true; } return false; } @@ -1545,12 +1522,10 @@ bool Scanner::fetchMoreTokens() { } Stream::Stream(StringRef Input, SourceMgr &SM) - : scanner(new Scanner(Input, SM)) - , CurrentDoc(0) {} + : scanner(new Scanner(Input, SM)), CurrentDoc() {} Stream::Stream(MemoryBuffer *InputBuffer, SourceMgr &SM) - : scanner(new Scanner(InputBuffer, SM)) - , CurrentDoc(0) {} + : scanner(new Scanner(InputBuffer, SM)), CurrentDoc() {} Stream::~Stream() {} @@ -1565,10 +1540,6 @@ void Stream::printError(Node *N, const Twine &Msg) { , Ranges); } -void Stream::handleYAMLDirective(const Token &t) { - // TODO: Ensure version is 1.x. -} - document_iterator Stream::begin() { if (CurrentDoc) report_fatal_error("Can only iterate over the stream once"); @@ -1589,14 +1560,57 @@ void Stream::skip() { i->skip(); } -Node::Node(unsigned int Type, OwningPtr &D, StringRef A) - : Doc(D) - , TypeID(Type) - , Anchor(A) { +Node::Node(unsigned int Type, std::unique_ptr &D, StringRef A, + StringRef T) + : Doc(D), TypeID(Type), Anchor(A), Tag(T) { SMLoc Start = SMLoc::getFromPointer(peekNext().Range.begin()); SourceRange = SMRange(Start, Start); } +std::string Node::getVerbatimTag() const { + StringRef Raw = getRawTag(); + if (!Raw.empty() && Raw != "!") { + std::string Ret; + if (Raw.find_last_of('!') == 0) { + Ret = Doc->getTagMap().find("!")->second; + Ret += Raw.substr(1); + return std::move(Ret); + } else if (Raw.startswith("!!")) { + Ret = Doc->getTagMap().find("!!")->second; + Ret += Raw.substr(2); + return std::move(Ret); + } else { + StringRef TagHandle = Raw.substr(0, Raw.find_last_of('!') + 1); + std::map::const_iterator It = + Doc->getTagMap().find(TagHandle); + if (It != Doc->getTagMap().end()) + Ret = It->second; + else { + Token T; + T.Kind = Token::TK_Tag; + T.Range = TagHandle; + setError(Twine("Unknown tag handle ") + TagHandle, T); + } + Ret += Raw.substr(Raw.find_last_of('!') + 1); + return std::move(Ret); + } + } + + switch (getType()) { + case NK_Null: + return "tag:yaml.org,2002:null"; + case NK_Scalar: + // TODO: Tag resolution. + return "tag:yaml.org,2002:str"; + case NK_Mapping: + return "tag:yaml.org,2002:map"; + case NK_Sequence: + return "tag:yaml.org,2002:seq"; + } + + return ""; +} + Token &Node::peekNext() { return Doc->peekNext(); } @@ -2000,6 +2014,10 @@ void SequenceNode::increment() { } Document::Document(Stream &S) : stream(S), Root(0) { + // Tag maps starts with two default mappings. + TagMap["!"] = "!"; + TagMap["!!"] = "tag:yaml.org,2002:"; + if (parseDirectives()) expectToken(Token::TK_DocumentStart); Token &T = peekNext(); @@ -2043,6 +2061,7 @@ Node *Document::parseBlockNode() { Token T = peekNext(); // Handle properties. Token AnchorInfo; + Token TagInfo; parse_property: switch (T.Kind) { case Token::TK_Alias: @@ -2057,7 +2076,11 @@ parse_property: T = peekNext(); goto parse_property; case Token::TK_Tag: - getNext(); // Skip TK_Tag. + if (TagInfo.Kind == Token::TK_Tag) { + setError("Already encountered a tag for this node!", T); + return 0; + } + TagInfo = getNext(); // Consume TK_Tag. T = peekNext(); goto parse_property; default: @@ -2071,42 +2094,49 @@ parse_property: // Don't eat the TK_BlockEntry, SequenceNode needs it. return new (NodeAllocator) SequenceNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , SequenceNode::ST_Indentless); case Token::TK_BlockSequenceStart: getNext(); return new (NodeAllocator) SequenceNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , SequenceNode::ST_Block); case Token::TK_BlockMappingStart: getNext(); return new (NodeAllocator) MappingNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , MappingNode::MT_Block); case Token::TK_FlowSequenceStart: getNext(); return new (NodeAllocator) SequenceNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , SequenceNode::ST_Flow); case Token::TK_FlowMappingStart: getNext(); return new (NodeAllocator) MappingNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , MappingNode::MT_Flow); case Token::TK_Scalar: getNext(); return new (NodeAllocator) ScalarNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , T.Range); case Token::TK_Key: // Don't eat the TK_Key, KeyValueNode expects it. return new (NodeAllocator) MappingNode( stream.CurrentDoc , AnchorInfo.Range.substr(1) + , TagInfo.Range , MappingNode::MT_Inline); case Token::TK_DocumentStart: case Token::TK_DocumentEnd: @@ -2127,10 +2157,10 @@ bool Document::parseDirectives() { while (true) { Token T = peekNext(); if (T.Kind == Token::TK_TagDirective) { - handleTagDirective(getNext()); + parseTAGDirective(); isDirective = true; } else if (T.Kind == Token::TK_VersionDirective) { - stream.handleYAMLDirective(getNext()); + parseYAMLDirective(); isDirective = true; } else break; @@ -2138,6 +2168,21 @@ bool Document::parseDirectives() { return isDirective; } +void Document::parseYAMLDirective() { + getNext(); // Eat %YAML +} + +void Document::parseTAGDirective() { + Token Tag = getNext(); // %TAG + StringRef T = Tag.Range; + // Strip %TAG + T = T.substr(T.find_first_of(" \t")).ltrim(" \t"); + std::size_t HandleEnd = T.find_first_of(" \t"); + StringRef TagHandle = T.substr(0, HandleEnd); + StringRef TagPrefix = T.substr(HandleEnd).ltrim(" \t"); + TagMap[TagHandle] = TagPrefix; +} + bool Document::expectToken(int TK) { Token T = getNext(); if (T.Kind != TK) {