// Copyright 2007-2011 Baptiste Lepilleur
// Distributed under MIT license, or public domain if desired and
// recognized in your jurisdiction.
// See file LICENSE for detail or copy at http://jsoncpp.sourceforge.net/LICENSE
#if !defined(JSON_IS_AMALGAMATION)
# include <json/assertions.h>
# include <json/reader.h>
# include <json/value.h>
# include "json_tool.h"
#endif // if !defined(JSON_IS_AMALGAMATION)
#include <utility>
#include <cstdio>
#include <cassert>
#include <cstring>
#include <stdexcept>
#if _MSC_VER >= 1400 // VC++ 8.0
#pragma warning( disable : 4996 ) // disable warning about strdup being deprecated.
#endif
namespace Json {
// Implementation of class Features
// ////////////////////////////////
Features::Features()
: allowComments_( true )
, strictRoot_( false )
{
}
Features
Features::all()
{
return Features();
}
Features
Features::strictMode()
{
Features features;
features.allowComments_ = false;
features.strictRoot_ = true;
return features;
}
// Implementation of class Reader
// ////////////////////////////////
static inline bool
in( Reader::Char c, Reader::Char c1, Reader::Char c2, Reader::Char c3, Reader::Char c4 )
{
return c == c1 || c == c2 || c == c3 || c == c4;
}
static inline bool
in( Reader::Char c, Reader::Char c1, Reader::Char c2, Reader::Char c3, Reader::Char c4, Reader::Char c5 )
{
return c == c1 || c == c2 || c == c3 || c == c4 || c == c5;
}
static bool
containsNewLine( Reader::Location begin,
Reader::Location end )
{
for ( ;begin < end; ++begin )
if ( *begin == '\n' || *begin == '\r' )
return true;
return false;
}
// Class Reader
// //////////////////////////////////////////////////////////////////
Reader::Reader()
: errors_(),
document_(),
begin_(),
end_(),
current_(),
lastValueEnd_(),
lastValue_(),
commentsBefore_(),
features_( Features::all() ),
collectComments_()
{
}
Reader::Reader( const Features &features )
: errors_(),
document_(),
begin_(),
end_(),
current_(),
lastValueEnd_(),
lastValue_(),
commentsBefore_(),
features_( features ),
collectComments_()
{
}
bool
Reader::parse( const std::string &document,
Value &root,
bool collectComments )
{
document_ = document;
const char *begin = document_.c_str();
const char *end = begin + document_.length();
return parse( begin, end, root, collectComments );
}
bool
Reader::parse( std::istream& sin,
Value &root,
bool collectComments )
{
//std::istream_iterator<char> begin(sin);
//std::istream_iterator<char> end;
// Those would allow streamed input from a file, if parse() were a
// template function.
// Since std::string is reference-counted, this at least does not
// create an extra copy.
std::string doc;
std::getline(sin, doc, (char)EOF);
return parse( doc, root, collectComments );
}
bool
Reader::parse( const char *beginDoc, const char *endDoc,
Value &root,
bool collectComments )
{
if ( !features_.allowComments_ )
{
collectComments = false;
}
begin_ = beginDoc;
end_ = endDoc;
collectComments_ = collectComments;
current_ = begin_;
lastValueEnd_ = 0;
lastValue_ = 0;
commentsBefore_ = "";
errors_.clear();
while ( !nodes_.empty() )
nodes_.pop();
nodes_.push( &root );
bool successful = readValue();
Token token;
skipCommentTokens( token );
if ( collectComments_ && !commentsBefore_.empty() )
root.setComment( commentsBefore_, commentAfter );
if ( features_.strictRoot_ )
{
if ( !root.isArray() && !root.isObject() )
{
// Set error location to start of doc, ideally should be first token found in doc
token.type_ = tokenError;
token.start_ = beginDoc;
token.end_ = endDoc;
addError( "A valid JSON document must be either an array or an object value.",
token );
return false;
}
}
return successful;
}
bool
Reader::readValue()
{
Token token;
skipCommentTokens( token );
bool successful = true;
if ( collectComments_ && !commentsBefore_.empty() )
{
currentValue().setComment( commentsBefore_, commentBefore );
commentsBefore_ = "";
}
switch ( token.type_ )
{
case tokenObjectBegin:
successful = readObject( token );
break;
case tokenArrayBegin:
successful = readArray( token );
break;
case tokenNumber:
successful = decodeNumber( token );
break;
case tokenString:
successful = decodeString( token );
break;
case tokenTrue:
currentValue() = true;
break;
case tokenFalse:
currentValue() = false;
break;
case tokenNull:
currentValue() = Value();
break;
default:
return addError( "Syntax error: value, object or array expected.", token );
}
if ( collectComments_ )
{
lastValueEnd_ = current_;
lastValue_ = ¤tValue();
}
return successful;
}
void
Reader::skipCommentTokens( Token &token )
{
if ( features_.allowComments_ )
{
do
{
readToken( token );
}
while ( token.type_ == tokenComment );
}
else
{
readToken( token );
}
}
bool
Reader::expectToken( TokenType type, Token &token, const char *message )
{
readToken( token );
if ( token.type_ != type )
return addError( message, token );
return true;
}
bool
Reader::readToken( Token &token )
{
skipSpaces();
token.start_ = current_;
Char c = getNextChar();
bool ok = true;
switch ( c )
{
case '{':
token.type_ = tokenObjectBegin;
break;
case '}':
token.type_ = tokenObjectEnd;
break;
case '[':
token.type_ = tokenArrayBegin;
break;
case ']':
token.type_ = tokenArrayEnd;
break;
case '"':
token.type_ = tokenString;
ok = readString();
break;
case '/':
token.type_ = tokenComment;
ok = readComment();
break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
case '8':
case '9':
case '-':
token.type_ = tokenNumber;
readNumber();
break;
case 't':
token.type_ = tokenTrue;
ok = match( "rue", 3 );
break;
case 'f':
token.type_ = tokenFalse;
ok = match( "alse", 4 );
break;
case 'n':
token.type_ = tokenNull;
ok = match( "ull", 3 );
break;
case ',':
token.type_ = tokenArraySeparator;
break;
case ':':
token.type_ = tokenMemberSeparator;
break;
case 0:
token.type_ = tokenEndOfStream;
break;
default:
ok = false;
break;
}
if ( !ok )
token.type_ = tokenError;
token.end_ = current_;
return true;
}
void
Reader::skipSpaces()
{
while ( current_ != end_ )
{
Char c = *current_;
if ( c == ' ' || c == '\t' || c == '\r' || c == '\n' )
++current_;
else
break;
}
}
bool
Reader::match( Location pattern,
int patternLength )
{
if ( end_ - current_ < patternLength )
return false;
int index = patternLength;
while ( index-- )
if ( current_[index] != pattern[index] )
return false;
current_ += patternLength;
return true;
}
bool
Reader::readComment()
{
Location commentBegin = current_ - 1;
Char c = getNextChar();
bool successful = false;
if ( c == '*' )
successful = readCStyleComment();
else if ( c == '/' )
successful = readCppStyleComment();
if ( !successful )
return false;
if ( collectComments_ )
{
CommentPlacement placement = commentBefore;
if ( lastValueEnd_ && !containsNewLine( lastValueEnd_, commentBegin ) )
{
if ( c != '*' || !containsNewLine( commentBegin, current_ ) )
placement = commentAfterOnSameLine;
}
addComment( commentBegin, current_, placement );
}
return true;
}
void
Reader::addComment( Location begin,
Location end,
CommentPlacement placement )
{
assert( collectComments_ );
if ( placement == commentAfterOnSameLine )
{
assert( lastValue_ != 0 );
lastValue_->setComment( std::string( begin, end ), placement );
}
else
{
if ( !commentsBefore_.empty() )
commentsBefore_ += "\n";
commentsBefore_ += std::string( begin, end );
}
}
bool
Reader::readCStyleComment()
{
while ( current_ != end_ )
{
Char c = getNextChar();
if ( c == '*' && *current_ == '/' )
break;
}
return getNextChar() == '/';
}
bool
Reader::readCppStyleComment()
{
while ( current_ != end_ )
{
Char c = getNextChar();
if ( c == '\r' || c == '\n' )
break;
}
return true;
}
void
Reader::readNumber()
{
while ( current_ != end_ )
{
if ( !(*current_ >= '0' && *current_ <= '9') &&
!in( *current_, '.', 'e', 'E', '+', '-' ) )
break;
++current_;
}
}
bool
Reader::readString()
{
Char c = 0;
while ( current_ != end_ )
{
c = getNextChar();
if ( c == '\\' )
getNextChar();
else if ( c == '"' )
break;
}
return c == '"';
}
bool
Reader::readObject( Token &/*tokenStart*/ )
{
Token tokenName;
std::string name;
currentValue() = Value( objectValue );
while ( readToken( tokenName ) )
{
bool initialTokenOk = true;
while ( tokenName.type_ == tokenComment && initialTokenOk )
initialTokenOk = readToken( tokenName );
if ( !initialTokenOk )
break;
if ( tokenName.type_ == tokenObjectEnd && name.empty() ) // empty object
return true;
if ( tokenName.type_ != tokenString )
break;
name = "";
if ( !decodeString( tokenName, name ) )
return recoverFromError( tokenObjectEnd );
Token colon;
if ( !readToken( colon ) || colon.type_ != tokenMemberSeparator )
{
return addErrorAndRecover( "Missing ':' after object member name",
colon,
tokenObjectEnd );
}
Value &value = currentValue()[ name ];
nodes_.push( &value );
bool ok = readValue();
nodes_.pop();
if ( !ok ) // error already set
return recoverFromError( tokenObjectEnd );
Token comma;
if ( !readToken( comma )
|| ( comma.type_ != tokenObjectEnd &&
comma.type_ != tokenArraySeparator &&
comma.type_ != tokenComment ) )
{
return addErrorAndRecover( "Missing ',' or '}' in object declaration",
comma,
tokenObjectEnd );
}
bool finalizeTokenOk = true;
while ( comma.type_ == tokenComment &&
finalizeTokenOk )
finalizeTokenOk = readToken( comma );
if ( comma.type_ == tokenObjectEnd )
return true;
}
return addErrorAndRecover( "Missing '}' or object member name",
tokenName,
tokenObjectEnd );
}
bool
Reader::readArray( Token &/*tokenStart*/ )
{
currentValue() = Value( arrayValue );
skipSpaces();
if ( *current_ == ']' ) // empty array
{
Token endArray;
readToken( endArray );
return true;
}
int index = 0;
for (;;)
{
Value &value = currentValue()[ index++ ];
nodes_.push( &value );
bool ok = readValue();
nodes_.pop();
if ( !ok ) // error already set
return recoverFromError( tokenArrayEnd );
Token token;
// Accept Comment after last item in the array.
ok = readToken( token );
while ( token.type_ == tokenComment && ok )
{
ok = readToken( token );
}
bool badTokenType = ( token.type_ != tokenArraySeparator &&
token.type_ != tokenArrayEnd );
if ( !ok || badTokenType )
{
return addErrorAndRecover( "Missing ',' or ']' in array declaration",
token,
tokenArrayEnd );
}
if ( token.type_ == tokenArrayEnd )
break;
}
return true;
}
bool
Reader::decodeNumber( Token &token )
{
bool isDouble = false;
for ( Location inspect = token.start_; inspect != token.end_; ++inspect )
{
isDouble = isDouble
|| in( *inspect, '.', 'e', 'E', '+' )
|| ( *inspect == '-' && inspect != token.start_ );
}
if ( isDouble )
return decodeDouble( token );
// Attempts to parse the number as an integer. If the number is
// larger than the maximum supported value of an integer then
// we decode the number as a double.
Location current = token.start_;
bool isNegative = *current == '-';
if ( isNegative )
++current;
Value::LargestUInt maxIntegerValue = isNegative ? Value::LargestUInt(-Value::minLargestInt)
: Value::maxLargestUInt;
Value::LargestUInt threshold = maxIntegerValue / 10;
Value::LargestUInt value = 0;
while ( current < token.end_ )
{
Char c = *current++;
if ( c < '0' || c > '9' )
return addError( "'" + std::string( token.start_, token.end_ ) + "' is not a number.", token );
Value::UInt digit(c - '0');
if ( value >= threshold )
{
// We've hit or exceeded the max value divided by 10 (rounded down). If
// a) we've only just touched the limit, b) this is the last digit, and
// c) it's small enough to fit in that rounding delta, we're okay.
// Otherwise treat this number as a double to avoid overflow.
if (value > threshold ||
current != token.end_ ||
digit > maxIntegerValue % 10)
{
return decodeDouble( token );
}
}
value = value * 10 + digit;
}
if ( isNegative )
currentValue() = -Value::LargestInt( value );
else if ( value <= Value::LargestUInt(Value::maxInt) )
currentValue() = Value::LargestInt( value );
else
currentValue() = value;
return true;
}
bool
Reader::decodeDouble( Token &token )
{
double value = 0;
const int bufferSize = 32;
int count;
int length = int(token.end_ - token.start_);
// Sanity check to avoid buffer overflow exploits.
if (length < 0) {
return addError( "Unable to parse token length", token );
}
// Avoid using a string constant for the format control string given to
// sscanf, as this can cause hard to debug crashes on OS X. See here for more
// info:
//
// http://developer.apple.com/library/mac/#DOCUMENTATION/DeveloperTools/gcc-4.0.1/gcc/Incompatibilities.html
char format[] = "%lf";
if ( length <= bufferSize )
{
Char buffer[bufferSize+1];
memcpy( buffer, token.start_, length );
buffer[length] = 0;
count = sscanf( buffer, format, &value );
}
else
{
std::string buffer( token.start_, token.end_ );
count = sscanf( buffer.c_str(), format, &value );
}
if ( count != 1 )
return addError( "'" + std::string( token.start_, token.end_ ) + "' is not a number.", token );
currentValue() = value;
return true;
}
bool
Reader::decodeString( Token &token )
{
std::string decoded;
if ( !decodeString( token, decoded ) )
return false;
currentValue() = decoded;
return true;
}
bool
Reader::decodeString( Token &token, std::string &decoded )
{
decoded.reserve( token.end_ - token.start_ - 2 );
Location current = token.start_ + 1; // skip '"'
Location end = token.end_ - 1; // do not include '"'
while ( current != end )
{
Char c = *current++;
if ( c == '"' )
break;
else if ( c == '\\' )
{
if ( current == end )
return addError( "Empty escape sequence in string", token, current );
Char escape = *current++;
switch ( escape )
{
case '"': decoded += '"'; break;
case '/': decoded += '/'; break;
case '\\': decoded += '\\'; break;
case 'b': decoded += '\b'; break;
case 'f': decoded += '\f'; break;
case 'n': decoded += '\n'; break;
case 'r': decoded += '\r'; break;
case 't': decoded += '\t'; break;
case 'u':
{
unsigned int unicode;
if ( !decodeUnicodeCodePoint( token, current, end, unicode ) )
return false;
decoded += codePointToUTF8(unicode);
}
break;
default:
return addError( "Bad escape sequence in string", token, current );
}
}
else
{
decoded += c;
}
}
return true;
}
bool
Reader::decodeUnicodeCodePoint( Token &token,
Location ¤t,
Location end,
unsigned int &unicode )
{
if ( !decodeUnicodeEscapeSequence( token, current, end, unicode ) )
return false;
if (unicode >= 0xD800 && unicode <= 0xDBFF)
{
// surrogate pairs
if (end - current < 6)
return addError( "additional six characters expected to parse unicode surrogate pair.", token, current );
unsigned int surrogatePair;
if (*(current++) == '\\' && *(current++)== 'u')
{
if (decodeUnicodeEscapeSequence( token, current, end, surrogatePair ))
{
unicode = 0x10000 + ((unicode & 0x3FF) << 10) + (surrogatePair & 0x3FF);
}
else
return false;
}
else
return addError( "expecting another \\u token to begin the second half of a unicode surrogate pair", token, current );
}
return true;
}
bool
Reader::decodeUnicodeEscapeSequence( Token &token,
Location ¤t,
Location end,
unsigned int &unicode )
{
if ( end - current < 4 )
return addError( "Bad unicode escape sequence in string: four digits expected.", token, current );
unicode = 0;
for ( int index =0; index < 4; ++index )
{
Char c = *current++;
unicode *= 16;
if ( c >= '0' && c <= '9' )
unicode += c - '0';
else if ( c >= 'a' && c <= 'f' )
unicode += c - 'a' + 10;
else if ( c >= 'A' && c <= 'F' )
unicode += c - 'A' + 10;
else
return addError( "Bad unicode escape sequence in string: hexadecimal digit expected.", token, current );
}
return true;
}
bool
Reader::addError( const std::string &message,
Token &token,
Location extra )
{
ErrorInfo info;
info.token_ = token;
info.message_ = message;
info.extra_ = extra;
errors_.push_back( info );
return false;
}
bool
Reader::recoverFromError( TokenType skipUntilToken )
{
int errorCount = int(errors_.size());
Token skip;
for (;;)
{
if ( !readToken(skip) )
errors_.resize( errorCount ); // discard errors caused by recovery
if ( skip.type_ == skipUntilToken || skip.type_ == tokenEndOfStream )
break;
}
errors_.resize( errorCount );
return false;
}
bool
Reader::addErrorAndRecover( const std::string &message,
Token &token,
TokenType skipUntilToken )
{
addError( message, token );
return recoverFromError( skipUntilToken );
}
Value &
Reader::currentValue()
{
return *(nodes_.top());
}
Reader::Char
Reader::getNextChar()
{
if ( current_ == end_ )
return 0;
return *current_++;
}
void
Reader::getLocationLineAndColumn( Location location,
int &line,
int &column ) const
{
Location current = begin_;
Location lastLineStart = current;
line = 0;
while ( current < location && current != end_ )
{
Char c = *current++;
if ( c == '\r' )
{
if ( *current == '\n' )
++current;
lastLineStart = current;
++line;
}
else if ( c == '\n' )
{
lastLineStart = current;
++line;
}
}
// column & line start at 1
column = int(location - lastLineStart) + 1;
++line;
}
std::string
Reader::getLocationLineAndColumn( Location location ) const
{
int line, column;
getLocationLineAndColumn( location, line, column );
char buffer[18+16+16+1];
sprintf( buffer, "Line %d, Column %d", line, column );
return buffer;
}
// Deprecated. Preserved for backward compatibility
std::string
Reader::getFormatedErrorMessages() const
{
return getFormattedErrorMessages();
}
std::string
Reader::getFormattedErrorMessages() const
{
std::string formattedMessage;
for ( Errors::const_iterator itError = errors_.begin();
itError != errors_.end();
++itError )
{
const ErrorInfo &error = *itError;
formattedMessage += "* " + getLocationLineAndColumn( error.token_.start_ ) + "\n";
formattedMessage += " " + error.message_ + "\n";
if ( error.extra_ )
formattedMessage += "See " + getLocationLineAndColumn( error.extra_ ) + " for detail.\n";
}
return formattedMessage;
}
std::istream& operator>>( std::istream &sin, Value &root )
{
Json::Reader reader;
bool ok = reader.parse(sin, root, true);
if (!ok) {
fprintf(
stderr,
"Error from reader: %s",
reader.getFormattedErrorMessages().c_str());
JSON_FAIL_MESSAGE("reader error");
}
return sin;
}
} // namespace Json