bool Parser::Impl::parseTest() {
// test := identifier arguments
// arguments := *argument [ test / test-list ]
//
// identifier
//
if ( !obtainToken() || atEnd() )
return false;
if ( token() != Lexer::Identifier )
return false;
if ( scriptBuilder() )
scriptBuilder()->testStart( tokenValue() );
consumeToken();
//
// *argument
//
if ( !obtainToken() )
return false;
if ( atEnd() ) // a test w/o args
goto TestEnd;
if ( isArgumentToken() && !parseArgumentList() ) {
assert( error() );
return false;
}
//
// test / test-list
//
if ( !obtainToken() )
return false;
if ( atEnd() ) // a test w/o nested tests
goto TestEnd;
if ( token() == Lexer::Special && tokenValue() == "(" ) { // test-list
if ( !parseTestList() ) {
assert( error() );
return false;
}
} else if ( token() == Lexer::Identifier ) { // should be test:
if ( !parseTest() ) {
assert( error() );
return false;
}
}
TestEnd:
if ( scriptBuilder() )
scriptBuilder()->testEnd();
return true;
}
bool Parser::Impl::parseArgument() {
// argument := string-list / number / tag
if ( !obtainToken() || atEnd() )
return false;
if ( token() == Lexer::Number ) {
if ( !parseNumber() ) {
assert( error() );
return false;
}
return true;
} else if ( token() == Lexer::Tag ) {
if ( scriptBuilder() )
scriptBuilder()->taggedArgument( tokenValue() );
consumeToken();
return true;
} else if ( isStringToken() ) {
if ( scriptBuilder() )
scriptBuilder()->stringArgument( tokenValue(), token() == Lexer::MultiLineString, QString() );
consumeToken();
return true;
} else if ( token() == Lexer::Special && tokenValue() == "[" ) {
if ( !parseStringList() ) {
assert( error() );
return false;
}
return true;
}
return false;
}
bool Parser::Impl::obtainToken() {
while ( !mToken && !lexer.atEnd() && !lexer.error() ) {
mToken = lexer.nextToken( mTokenValue );
if ( lexer.error() )
break;
// comments and line feeds are semantically invisible and may
// appear anywhere, so we handle them here centrally:
switch ( token() ) {
case Lexer::HashComment:
if ( scriptBuilder() )
scriptBuilder()->hashComment( tokenValue() );
consumeToken();
break;
case Lexer::BracketComment:
if ( scriptBuilder() )
scriptBuilder()->bracketComment( tokenValue() );
consumeToken();
break;
case Lexer::LineFeeds:
for ( unsigned int i = 0, end = tokenValue().toUInt() ; i < end ; ++i )
if ( scriptBuilder() ) // better check every iteration, b/c
// we call out to ScriptBuilder,
// where nasty things might happen!
scriptBuilder()->lineFeed();
consumeToken();
break;
default: ; // make compiler happy
}
}
if ( lexer.error() && scriptBuilder() )
scriptBuilder()->error( lexer.error() );
return !lexer.error();
}
bool Parser::Impl::parseBlock()
{
// our ABNF:
// block := "{" [ command-list ] "}"
if(!obtainToken() || atEnd())
return false;
if(token() != Lexer::Special || tokenValue() != "{")
return false;
if(scriptBuilder())
scriptBuilder()->blockStart();
consumeToken();
if(!obtainToken())
return false;
if(atEnd())
{
makeError(Error::PrematureEndOfBlock);
return false;
}
if(token() == Lexer::Identifier)
{
if(!parseCommandList())
{
assert(error());
return false;
}
}
if(!obtainToken())
return false;
if(atEnd())
{
makeError(Error::PrematureEndOfBlock);
return false;
}
if(token() != Lexer::Special || tokenValue() != "}")
{
makeError(Error::NonCommandInCommandList);
return false;
}
if(scriptBuilder())
scriptBuilder()->blockEnd();
consumeToken();
return true;
}
ListNode addToList(int lineNo, int tokenId, char* tokenString,
Token tokenType, ListNode head){
ListNode this = head;
ListNode newNode;
newNode = (ListNode) malloc(sizeof(struct node));
newNode -> lineNo = lineNo;
newNode -> tokenId = tokenId;
newNode -> tokenStr = strdup(tokenString);
newNode -> tokenType = tokenType;
newNode -> token = tokenValue(newNode -> tokenType, newNode -> tokenStr);
newNode -> next = NULL;
if(head == NULL){
head = newNode;
return head;
}
while(this -> next != NULL){
this = this -> next;
}
this -> next = newNode;
return head;
}
bool Parser::Impl::parseNumber()
{
// The lexer returns the number including the quantifier as a
// single token value. Here, we split is an check that the number
// is not out of range:
if(!obtainToken() || atEnd())
return false;
if(token() != Lexer::Number)
return false;
// number:
unsigned long result = 0;
unsigned int i = 0;
const QCString s = tokenValue().latin1();
for(const unsigned int len = s.length() ; i < len && isdigit(s[i]) ; ++i)
{
const unsigned long digitValue = s[i] - '0' ;
if(willOverflowULong(result, digitValue))
{
makeError(Error::NumberOutOfRange);
return false;
}
else
{
result *= 10 ;
result += digitValue ;
}
}
// optional quantifier:
char quantifier = '\0';
if(i < s.length())
{
assert(i + 1 == s.length());
quantifier = s[i];
const unsigned long factor = factorForQuantifier(quantifier);
if(result > double(ULONG_MAX) / double(factor))
{
makeError(Error::NumberOutOfRange);
return false;
}
result *= factor;
}
if(scriptBuilder())
scriptBuilder()->numberArgument(result, quantifier);
consumeToken();
return true;
}
short Scanner::getToken( YYSTYPE & lval )
//---------------------------------------
// get the next token from the input stream.
// at each call, the next character should be
// in _current.
{
const int MaxBufLen = 512;
char buffer[ MaxBufLen ];
int bufPos; // position in buffer
char special;
gobble();
if( isEOF() ) {
return 0;
}
if( isSpecial() ) {
special = (char) _current;
get(); // set up for next call
return special; // <-------- early return
}
if( isQuote() ) {
readQuotedString( lval );
return T_String; // <-------- early return
}
if( _current == '0' ) {
get();
if( toupper( _current ) == 'X' ) {
readHex( lval );
} else {
if( isDigit() ) {
readDecimal( lval );
} else {
lval = 0;
}
}
return T_Number; // <-------- early return
}
if( isDigit() ) {
readDecimal( lval );
return T_Number; // <-------- early return
}
for( bufPos = 0; bufPos < MaxBufLen; bufPos += 1 ) {
buffer[ bufPos ] = (char) _current;
get();
if( isEOF() || isSpace() || isSpecial() ) break;
}
bufPos += 1;
assert( bufPos < MaxBufLen );
buffer[ bufPos ] = '\0';
return tokenValue( buffer, lval );
}
bool Parser::Impl::parseStringList() {
// string-list := "[" string *("," string) "]" / string
// ;; if there is only a single string, the brackets are optional
//
// However, since strings are already handled separately from
// string lists in parseArgument(), our ABNF is modified to:
// string-list := "[" string *("," string) "]"
if ( !obtainToken() || atEnd() )
return false;
if ( token() != Lexer::Special || tokenValue() != "[" )
return false;
if ( scriptBuilder() )
scriptBuilder()->stringListArgumentStart();
consumeToken();
// generic while/switch construct for comma-separated lists. See
// parseTestList() for another one. Any fix here is like to apply there, too.
bool lastWasComma = true;
while ( !atEnd() ) {
if ( !obtainToken() )
return false;
switch ( token() ) {
case Lexer::None:
break;
case Lexer::Special:
assert( tokenValue().length() == 1 );
switch ( tokenValue()[0].toLatin1() ) {
case ']':
consumeToken();
if ( lastWasComma ) {
makeError( Error::ConsecutiveCommasInStringList );
return false;
}
if ( scriptBuilder() )
scriptBuilder()->stringListArgumentEnd();
return true;
case ',':
consumeToken();
if ( lastWasComma ) {
makeError( Error::ConsecutiveCommasInStringList );
return false;
}
lastWasComma = true;
break;
default:
makeError( Error::NonStringInStringList );
return false;
}
break;
case Lexer::QuotedString:
case Lexer::MultiLineString:
if ( !lastWasComma ) {
makeError( Error::MissingCommaInStringList );
return false;
}
lastWasComma = false;
if ( scriptBuilder() )
scriptBuilder()->stringListEntry( tokenValue(), token() == Lexer::MultiLineString, QString() );
consumeToken();
break;
default:
makeError( Error::NonStringInStringList );
return false;
}
}
makeError( Error::PrematureEndOfStringList );
return false;
}
bool Parser::Impl::parseTestList() {
// test-list := "(" test *("," test) ")"
if ( !obtainToken() || atEnd() )
return false;
if ( token() != Lexer::Special || tokenValue() != "(" )
return false;
if ( scriptBuilder() )
scriptBuilder()->testListStart();
consumeToken();
// generic while/switch construct for comma-separated lists. See
// parseStringList() for another one. Any fix here is like to apply there, too.
bool lastWasComma = true;
while ( !atEnd() ) {
if ( !obtainToken() )
return false;
switch ( token() ) {
case Lexer::None:
break;
case Lexer::Special:
assert( tokenValue().length() == 1 );
assert( tokenValue()[0].toLatin1() );
switch ( tokenValue()[0].toLatin1() ) {
case ')':
consumeToken();
if ( lastWasComma ) {
makeError( Error::ConsecutiveCommasInTestList );
return false;
}
if ( scriptBuilder() )
scriptBuilder()->testListEnd();
return true;
case ',':
consumeToken();
if( lastWasComma ) {
makeError( Error::ConsecutiveCommasInTestList );
return false;
}
lastWasComma = true;
break;
default:
makeError( Error::NonStringInStringList );
return false;
}
break;
case Lexer::Identifier:
if ( !lastWasComma ) {
makeError( Error::MissingCommaInTestList );
return false;
} else {
lastWasComma = false;
if ( !parseTest() ) {
assert( error() );
return false;
}
}
break;
default:
makeUnexpectedTokenError( Error::NonTestInTestList );
return false;
}
}
makeError( Error::PrematureEndOfTestList );
return false;
}
bool Parser::Impl::parseCommand() {
// command := identifier arguments ( ";" / block )
// arguments := *argument [ test / test-list ]
// block := "{" *command "}"
// our ABNF:
// block := "{" [ command-list ] "}"
if ( atEnd() )
return false;
//
// identifier
//
if ( !obtainToken() || token() != Lexer::Identifier )
return false;
if ( scriptBuilder() )
scriptBuilder()->commandStart( tokenValue() );
consumeToken();
//
// *argument
//
if ( !obtainToken() )
return false;
if ( atEnd() ) {
makeError( Error::MissingSemicolonOrBlock );
return false;
}
if ( isArgumentToken() && !parseArgumentList() ) {
assert( error() );
return false;
}
//
// test / test-list
//
if ( !obtainToken() )
return false;
if ( atEnd() ) {
makeError( Error::MissingSemicolonOrBlock );
return false;
}
if ( token() == Lexer::Special && tokenValue() == "(" ) { // test-list
if ( !parseTestList() ) {
assert( error() );
return false;
}
} else if ( token() == Lexer::Identifier ) { // should be test:
if ( !parseTest() ) {
assert( error() );
return false;
}
}
//
// ";" / block
//
if ( !obtainToken() )
return false;
if ( atEnd() ) {
makeError( Error::MissingSemicolonOrBlock );
return false;
}
if ( token() != Lexer::Special ) {
makeUnexpectedTokenError( Error::ExpectedBlockOrSemicolon );
return false;
}
if ( tokenValue() == ";" )
consumeToken();
else if ( tokenValue() == "{" ) { // block
if ( !parseBlock() )
return false; // it's an error since we saw '{'
} else {
makeError( Error::MissingSemicolonOrBlock );
return false;
}
if ( scriptBuilder() )
scriptBuilder()->commandEnd();
return true;
}