mirror of
https://github.com/jbeder/yaml-cpp.git
synced 2025-09-09 12:41:17 +00:00
Update documentation for Scanner and AnchorDict, and fix formatting.
This commit is contained in:
@@ -12,11 +12,13 @@
|
||||
#include "../anchor.h"
|
||||
|
||||
namespace YAML {
|
||||
/// AnchorDict
|
||||
/// . An object that stores and retrieves values correlating to anchor_t
|
||||
/// values.
|
||||
/// . Efficient implementation that can make assumptions about how anchor_t
|
||||
/// values are assigned by the Parser class.
|
||||
/**
|
||||
* An object that stores and retrieves values correlating to {@link anchor_t}
|
||||
* values.
|
||||
*
|
||||
* <p>Efficient implementation that can make assumptions about how
|
||||
* {@code anchor_t} values are assigned by the {@link Parser} class.
|
||||
*/
|
||||
template <class T>
|
||||
class AnchorDict {
|
||||
public:
|
||||
|
164
src/scanner.cpp
164
src/scanner.cpp
@@ -16,23 +16,17 @@ Scanner::Scanner(std::istream& in)
|
||||
|
||||
Scanner::~Scanner() {}
|
||||
|
||||
// empty
|
||||
// . Returns true if there are no more tokens to be read
|
||||
bool Scanner::empty() {
|
||||
EnsureTokensInQueue();
|
||||
return m_tokens.empty();
|
||||
}
|
||||
|
||||
// pop
|
||||
// . Simply removes the next token on the queue.
|
||||
void Scanner::pop() {
|
||||
EnsureTokensInQueue();
|
||||
if (!m_tokens.empty())
|
||||
m_tokens.pop();
|
||||
}
|
||||
|
||||
// peek
|
||||
// . Returns (but does not remove) the next token on the queue.
|
||||
Token& Scanner::peek() {
|
||||
EnsureTokensInQueue();
|
||||
assert(!m_tokens.empty()); // should we be asserting here? I mean, we really
|
||||
@@ -49,21 +43,17 @@ Token& Scanner::peek() {
|
||||
return m_tokens.front();
|
||||
}
|
||||
|
||||
// mark
|
||||
// . Returns the current mark in the stream
|
||||
Mark Scanner::mark() const { return INPUT.mark(); }
|
||||
|
||||
// EnsureTokensInQueue
|
||||
// . Scan until there's a valid token at the front of the queue,
|
||||
// or we're sure the queue is empty.
|
||||
void Scanner::EnsureTokensInQueue() {
|
||||
while (1) {
|
||||
if (!m_tokens.empty()) {
|
||||
Token& token = m_tokens.front();
|
||||
|
||||
// if this guy's valid, then we're done
|
||||
if (token.status == Token::VALID)
|
||||
if (token.status == Token::VALID) {
|
||||
return;
|
||||
}
|
||||
|
||||
// here's where we clean up the impossible tokens
|
||||
if (token.status == Token::INVALID) {
|
||||
@@ -75,23 +65,23 @@ void Scanner::EnsureTokensInQueue() {
|
||||
}
|
||||
|
||||
// no token? maybe we've actually finished
|
||||
if (m_endedStream)
|
||||
if (m_endedStream) {
|
||||
return;
|
||||
}
|
||||
|
||||
// no? then scan...
|
||||
ScanNextToken();
|
||||
}
|
||||
}
|
||||
|
||||
// ScanNextToken
|
||||
// . The main scanning function; here we branch out and
|
||||
// scan whatever the next token should be.
|
||||
void Scanner::ScanNextToken() {
|
||||
if (m_endedStream)
|
||||
if (m_endedStream) {
|
||||
return;
|
||||
}
|
||||
|
||||
if (!m_startedStream)
|
||||
if (!m_startedStream) {
|
||||
return StartStream();
|
||||
}
|
||||
|
||||
// get rid of whitespace, etc. (in between tokens it should be irrelevent)
|
||||
ScanToNextToken();
|
||||
@@ -104,85 +94,102 @@ void Scanner::ScanNextToken() {
|
||||
// *****
|
||||
|
||||
// end of stream
|
||||
if (!INPUT)
|
||||
if (!INPUT) {
|
||||
return EndStream();
|
||||
}
|
||||
|
||||
if (INPUT.column() == 0 && INPUT.peek() == Keys::Directive)
|
||||
if (INPUT.column() == 0 && INPUT.peek() == Keys::Directive) {
|
||||
return ScanDirective();
|
||||
}
|
||||
|
||||
// document token
|
||||
if (INPUT.column() == 0 && Exp::DocStart().Matches(INPUT))
|
||||
if (INPUT.column() == 0 && Exp::DocStart().Matches(INPUT)) {
|
||||
return ScanDocStart();
|
||||
}
|
||||
|
||||
if (INPUT.column() == 0 && Exp::DocEnd().Matches(INPUT))
|
||||
if (INPUT.column() == 0 && Exp::DocEnd().Matches(INPUT)) {
|
||||
return ScanDocEnd();
|
||||
}
|
||||
|
||||
// flow start/end/entry
|
||||
if (INPUT.peek() == Keys::FlowSeqStart || INPUT.peek() == Keys::FlowMapStart)
|
||||
if (INPUT.peek() == Keys::FlowSeqStart ||
|
||||
INPUT.peek() == Keys::FlowMapStart) {
|
||||
return ScanFlowStart();
|
||||
}
|
||||
|
||||
if (INPUT.peek() == Keys::FlowSeqEnd || INPUT.peek() == Keys::FlowMapEnd)
|
||||
if (INPUT.peek() == Keys::FlowSeqEnd || INPUT.peek() == Keys::FlowMapEnd) {
|
||||
return ScanFlowEnd();
|
||||
}
|
||||
|
||||
if (INPUT.peek() == Keys::FlowEntry)
|
||||
if (INPUT.peek() == Keys::FlowEntry) {
|
||||
return ScanFlowEntry();
|
||||
}
|
||||
|
||||
// block/map stuff
|
||||
if (Exp::BlockEntry().Matches(INPUT))
|
||||
if (Exp::BlockEntry().Matches(INPUT)) {
|
||||
return ScanBlockEntry();
|
||||
}
|
||||
|
||||
if ((InBlockContext() ? Exp::Key() : Exp::KeyInFlow()).Matches(INPUT))
|
||||
if ((InBlockContext() ? Exp::Key() : Exp::KeyInFlow()).Matches(INPUT)) {
|
||||
return ScanKey();
|
||||
}
|
||||
|
||||
if (GetValueRegex().Matches(INPUT))
|
||||
if (GetValueRegex().Matches(INPUT)) {
|
||||
return ScanValue();
|
||||
}
|
||||
|
||||
// alias/anchor
|
||||
if (INPUT.peek() == Keys::Alias || INPUT.peek() == Keys::Anchor)
|
||||
if (INPUT.peek() == Keys::Alias || INPUT.peek() == Keys::Anchor) {
|
||||
return ScanAnchorOrAlias();
|
||||
}
|
||||
|
||||
// tag
|
||||
if (INPUT.peek() == Keys::Tag)
|
||||
if (INPUT.peek() == Keys::Tag) {
|
||||
return ScanTag();
|
||||
}
|
||||
|
||||
// special scalars
|
||||
if (InBlockContext() && (INPUT.peek() == Keys::LiteralScalar ||
|
||||
INPUT.peek() == Keys::FoldedScalar))
|
||||
INPUT.peek() == Keys::FoldedScalar)) {
|
||||
return ScanBlockScalar();
|
||||
}
|
||||
|
||||
if (INPUT.peek() == '\'' || INPUT.peek() == '\"')
|
||||
if (INPUT.peek() == '\'' || INPUT.peek() == '\"') {
|
||||
return ScanQuotedScalar();
|
||||
}
|
||||
|
||||
// plain scalars
|
||||
if ((InBlockContext() ? Exp::PlainScalar() : Exp::PlainScalarInFlow())
|
||||
.Matches(INPUT))
|
||||
.Matches(INPUT)) {
|
||||
return ScanPlainScalar();
|
||||
}
|
||||
|
||||
// don't know what it is!
|
||||
throw ParserException(INPUT.mark(), ErrorMsg::UNKNOWN_TOKEN);
|
||||
}
|
||||
|
||||
// ScanToNextToken
|
||||
// . Eats input until we reach the next token-like thing.
|
||||
void Scanner::ScanToNextToken() {
|
||||
while (1) {
|
||||
// first eat whitespace
|
||||
while (INPUT && IsWhitespaceToBeEaten(INPUT.peek())) {
|
||||
if (InBlockContext() && Exp::Tab().Matches(INPUT))
|
||||
if (InBlockContext() && Exp::Tab().Matches(INPUT)) {
|
||||
m_simpleKeyAllowed = false;
|
||||
}
|
||||
INPUT.eat(1);
|
||||
}
|
||||
|
||||
// then eat a comment
|
||||
if (Exp::Comment().Matches(INPUT)) {
|
||||
// eat until line break
|
||||
while (INPUT && !Exp::Break().Matches(INPUT))
|
||||
while (INPUT && !Exp::Break().Matches(INPUT)) {
|
||||
INPUT.eat(1);
|
||||
}
|
||||
}
|
||||
|
||||
// if it's NOT a line break, then we're done!
|
||||
if (!Exp::Break().Matches(INPUT))
|
||||
if (!Exp::Break().Matches(INPUT)) {
|
||||
break;
|
||||
}
|
||||
|
||||
// otherwise, let's eat the line break and keep going
|
||||
int n = Exp::Break().Match(INPUT);
|
||||
@@ -192,10 +199,11 @@ void Scanner::ScanToNextToken() {
|
||||
InvalidateSimpleKey();
|
||||
|
||||
// new line - we may be able to accept a simple key now
|
||||
if (InBlockContext())
|
||||
if (InBlockContext()) {
|
||||
m_simpleKeyAllowed = true;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
///////////////////////////////////////////////////////////////////////
|
||||
// Misc. helpers
|
||||
@@ -210,40 +218,39 @@ void Scanner::ScanToNextToken() {
|
||||
// that they can't contribute to indentation, so once you've seen a tab in a
|
||||
// line, you can't start a simple key
|
||||
bool Scanner::IsWhitespaceToBeEaten(char ch) {
|
||||
if (ch == ' ')
|
||||
if (ch == ' ') {
|
||||
return true;
|
||||
}
|
||||
|
||||
if (ch == '\t')
|
||||
if (ch == '\t') {
|
||||
return true;
|
||||
}
|
||||
|
||||
return false;
|
||||
}
|
||||
|
||||
// GetValueRegex
|
||||
// . Get the appropriate regex to check if it's a value token
|
||||
const RegEx& Scanner::GetValueRegex() const {
|
||||
if (InBlockContext())
|
||||
if (InBlockContext()) {
|
||||
return Exp::Value();
|
||||
}
|
||||
|
||||
return m_canBeJSONFlow ? Exp::ValueInJSONFlow() : Exp::ValueInFlow();
|
||||
}
|
||||
|
||||
// StartStream
|
||||
// . Set the initial conditions for starting a stream.
|
||||
void Scanner::StartStream() {
|
||||
m_startedStream = true;
|
||||
m_simpleKeyAllowed = true;
|
||||
std::unique_ptr<IndentMarker> pIndent(new IndentMarker(-1, IndentMarker::NONE));
|
||||
std::unique_ptr<IndentMarker> pIndent(
|
||||
new IndentMarker(-1, IndentMarker::NONE));
|
||||
m_indentRefs.push_back(std::move(pIndent));
|
||||
m_indents.push(&m_indentRefs.back());
|
||||
}
|
||||
|
||||
// EndStream
|
||||
// . Close out the stream, finish up, etc.
|
||||
void Scanner::EndStream() {
|
||||
// force newline
|
||||
if (INPUT.column() > 0)
|
||||
if (INPUT.column() > 0) {
|
||||
INPUT.ResetColumn();
|
||||
}
|
||||
|
||||
PopAllIndents();
|
||||
PopAllSimpleKeys();
|
||||
@@ -271,27 +278,26 @@ Token::TYPE Scanner::GetStartTokenFor(IndentMarker::INDENT_TYPE type) const {
|
||||
throw std::runtime_error("yaml-cpp: internal error, invalid indent type");
|
||||
}
|
||||
|
||||
// PushIndentTo
|
||||
// . Pushes an indentation onto the stack, and enqueues the
|
||||
// proper token (sequence start or mapping start).
|
||||
// . Returns the indent marker it generates (if any).
|
||||
Scanner::IndentMarker* Scanner::PushIndentTo(int column,
|
||||
IndentMarker::INDENT_TYPE type) {
|
||||
// are we in flow?
|
||||
if (InFlowContext())
|
||||
if (InFlowContext()) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
std::unique_ptr<IndentMarker> pIndent(new IndentMarker(column, type));
|
||||
IndentMarker& indent = *pIndent;
|
||||
const IndentMarker& lastIndent = *m_indents.top();
|
||||
|
||||
// is this actually an indentation?
|
||||
if (indent.column < lastIndent.column)
|
||||
if (indent.column < lastIndent.column) {
|
||||
return 0;
|
||||
}
|
||||
if (indent.column == lastIndent.column &&
|
||||
!(indent.type == IndentMarker::SEQ &&
|
||||
lastIndent.type == IndentMarker::MAP))
|
||||
lastIndent.type == IndentMarker::MAP)) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
// push a start token
|
||||
indent.pStartToken = PushToken(GetStartTokenFor(type));
|
||||
@@ -302,53 +308,50 @@ Scanner::IndentMarker* Scanner::PushIndentTo(int column,
|
||||
return &m_indentRefs.back();
|
||||
}
|
||||
|
||||
// PopIndentToHere
|
||||
// . Pops indentations off the stack until we reach the current indentation
|
||||
// level,
|
||||
// and enqueues the proper token each time.
|
||||
// . Then pops all invalid indentations off.
|
||||
void Scanner::PopIndentToHere() {
|
||||
// are we in flow?
|
||||
if (InFlowContext())
|
||||
if (InFlowContext()) {
|
||||
return;
|
||||
}
|
||||
|
||||
// now pop away
|
||||
while (!m_indents.empty()) {
|
||||
const IndentMarker& indent = *m_indents.top();
|
||||
if (indent.column < INPUT.column())
|
||||
if (indent.column < INPUT.column()) {
|
||||
break;
|
||||
}
|
||||
if (indent.column == INPUT.column() &&
|
||||
!(indent.type == IndentMarker::SEQ &&
|
||||
!Exp::BlockEntry().Matches(INPUT)))
|
||||
!Exp::BlockEntry().Matches(INPUT))) {
|
||||
break;
|
||||
}
|
||||
|
||||
PopIndent();
|
||||
}
|
||||
|
||||
while (!m_indents.empty() && m_indents.top()->status == IndentMarker::INVALID)
|
||||
while (!m_indents.empty() &&
|
||||
m_indents.top()->status == IndentMarker::INVALID) {
|
||||
PopIndent();
|
||||
}
|
||||
}
|
||||
|
||||
// PopAllIndents
|
||||
// . Pops all indentations (except for the base empty one) off the stack,
|
||||
// and enqueues the proper token each time.
|
||||
void Scanner::PopAllIndents() {
|
||||
// are we in flow?
|
||||
if (InFlowContext())
|
||||
if (InFlowContext()) {
|
||||
return;
|
||||
}
|
||||
|
||||
// now pop away
|
||||
while (!m_indents.empty()) {
|
||||
const IndentMarker& indent = *m_indents.top();
|
||||
if (indent.type == IndentMarker::NONE)
|
||||
if (indent.type == IndentMarker::NONE) {
|
||||
break;
|
||||
}
|
||||
|
||||
PopIndent();
|
||||
}
|
||||
}
|
||||
|
||||
// PopIndent
|
||||
// . Pops a single indent, pushing the proper token
|
||||
void Scanner::PopIndent() {
|
||||
const IndentMarker& indent = *m_indents.top();
|
||||
m_indents.pop();
|
||||
@@ -358,23 +361,20 @@ void Scanner::PopIndent() {
|
||||
return;
|
||||
}
|
||||
|
||||
if (indent.type == IndentMarker::SEQ)
|
||||
if (indent.type == IndentMarker::SEQ) {
|
||||
m_tokens.push(Token(Token::BLOCK_SEQ_END, INPUT.mark()));
|
||||
else if (indent.type == IndentMarker::MAP)
|
||||
} else if (indent.type == IndentMarker::MAP) {
|
||||
m_tokens.push(Token(Token::BLOCK_MAP_END, INPUT.mark()));
|
||||
}
|
||||
}
|
||||
|
||||
// GetTopIndent
|
||||
int Scanner::GetTopIndent() const {
|
||||
if (m_indents.empty())
|
||||
if (m_indents.empty()) {
|
||||
return 0;
|
||||
}
|
||||
return m_indents.top()->column;
|
||||
}
|
||||
|
||||
// ThrowParserException
|
||||
// . Throws a ParserException with the current token location
|
||||
// (if available).
|
||||
// . Does not parse any more tokens.
|
||||
void Scanner::ThrowParserException(const std::string& msg) const {
|
||||
Mark mark = Mark::null_mark();
|
||||
if (!m_tokens.empty()) {
|
||||
@@ -383,4 +383,4 @@ void Scanner::ThrowParserException(const std::string& msg) const {
|
||||
}
|
||||
throw ParserException(mark, msg);
|
||||
}
|
||||
}
|
||||
} // namespace YAML
|
||||
|
@@ -24,15 +24,24 @@ namespace YAML {
|
||||
class Node;
|
||||
class RegEx;
|
||||
|
||||
/**
|
||||
* A scanner transforms a stream of characters into a stream of tokens.
|
||||
*/
|
||||
class Scanner {
|
||||
public:
|
||||
Scanner(std::istream &in);
|
||||
explicit Scanner(std::istream &in);
|
||||
~Scanner();
|
||||
|
||||
// token queue management (hopefully this looks kinda stl-ish)
|
||||
/** Returns true if there are no more tokens to be read. */
|
||||
bool empty();
|
||||
|
||||
/** Removes the next token in the queue. */
|
||||
void pop();
|
||||
|
||||
/** Returns, but does not remove, the next token in the queue. */
|
||||
Token &peek();
|
||||
|
||||
/** Returns the current mark in the input stream. */
|
||||
Mark mark() const;
|
||||
|
||||
private:
|
||||
@@ -52,11 +61,29 @@ class Scanner {
|
||||
|
||||
private:
|
||||
// scanning
|
||||
|
||||
/**
|
||||
* Scans until there's a valid token at the front of the queue, or the queue
|
||||
* is empty. The state can be checked by {@link #empty}, and the next token
|
||||
* retrieved by {@link #peek}.
|
||||
*/
|
||||
void EnsureTokensInQueue();
|
||||
|
||||
/**
|
||||
* The main scanning function; this method branches out to scan whatever the
|
||||
* next token should be.
|
||||
*/
|
||||
void ScanNextToken();
|
||||
|
||||
/** Eats the input stream until it reaches the next token-like thing. */
|
||||
void ScanToNextToken();
|
||||
|
||||
/** Sets the initial conditions for starting a stream. */
|
||||
void StartStream();
|
||||
|
||||
/** Closes out the stream, finish up, etc. */
|
||||
void EndStream();
|
||||
|
||||
Token *PushToken(Token::TYPE type);
|
||||
|
||||
bool InFlowContext() const { return !m_flows.empty(); }
|
||||
@@ -64,9 +91,29 @@ class Scanner {
|
||||
std::size_t GetFlowLevel() const { return m_flows.size(); }
|
||||
|
||||
Token::TYPE GetStartTokenFor(IndentMarker::INDENT_TYPE type) const;
|
||||
|
||||
/**
|
||||
* Pushes an indentation onto the stack, and enqueues the proper token
|
||||
* (sequence start or mapping start).
|
||||
*
|
||||
* @return the indent marker it generates (if any).
|
||||
*/
|
||||
IndentMarker *PushIndentTo(int column, IndentMarker::INDENT_TYPE type);
|
||||
|
||||
/**
|
||||
* Pops indentations off the stack until it reaches the current indentation
|
||||
* level, and enqueues the proper token each time. Then pops all invalid
|
||||
* indentations off.
|
||||
*/
|
||||
void PopIndentToHere();
|
||||
|
||||
/**
|
||||
* Pops all indentations (except for the base empty one) off the stack, and
|
||||
* enqueues the proper token each time.
|
||||
*/
|
||||
void PopAllIndents();
|
||||
|
||||
/** Pops a single indent, pushing the proper token. */
|
||||
void PopIndent();
|
||||
int GetTopIndent() const;
|
||||
|
||||
@@ -78,9 +125,17 @@ class Scanner {
|
||||
bool VerifySimpleKey();
|
||||
void PopAllSimpleKeys();
|
||||
|
||||
/**
|
||||
* Throws a ParserException with the current token location (if available),
|
||||
* and does not parse any more tokens.
|
||||
*/
|
||||
void ThrowParserException(const std::string &msg) const;
|
||||
|
||||
bool IsWhitespaceToBeEaten(char ch);
|
||||
|
||||
/**
|
||||
* Returns the appropriate regex to check if the next token is a value token.
|
||||
*/
|
||||
const RegEx &GetValueRegex() const;
|
||||
|
||||
struct SimpleKey {
|
||||
|
Reference in New Issue
Block a user