static void identifier(void) { const RTYPE *rptr; /* Pointer to reserved word */ tknSubType = txNONE; /* Initialize */ /* Concatenate identifier */ do { *stringSP++ = toupper(inChar); /* concatenate char */ getCharacter(); /* get next character */ } while ((isalnum(inChar)) || (inChar == '_')); *stringSP++ = '\0'; /* make ASCIIZ string */ /* Check if the identifier is a reserved word */ rptr = findReservedWord(tkn_strt); if (rptr) { token = rptr->rtype; /* get type from rsw table */ tknSubType = rptr->subtype; /* get subtype from rsw table */ stringSP = tkn_strt; /* pop token from stack */ } /* End if */ /* Check if the identifier is a symbol */ else { tknPtr = findSymbol(tkn_strt); if (tknPtr) { token = tknPtr->sKind; /* get type from symbol table */ stringSP = tkn_strt; /* pop token from stack */ /* The following assignments only apply to constants. However it * is simpler just to make the assignments than it is to determine * if is appropriate to do so */ if (token == tREAL_CONST) tknReal = tknPtr->sParm.c.val.f; else tknInt = tknPtr->sParm.c.val.i; } /* End if */ /* Otherwise, the token is an identifier */ else token = tIDENT; } /* end else */ } /* End identifier */
int QScript::Lexer::lex() { int token = 0; state = Start; ushort stringType = 0; // either single or double quotes pos8 = pos16 = 0; done = false; terminator = false; // did we push a token on the stack previously ? // (after an automatic semicolon insertion) if (stackToken >= 0) { setDone(Other); token = stackToken; stackToken = -1; } while (!done) { switch (state) { case Start: if (isWhiteSpace()) { // do nothing } else if (current == '/' && next1 == '/') { recordStartPos(); shift(1); state = InSingleLineComment; } else if (current == '/' && next1 == '*') { recordStartPos(); shift(1); state = InMultiLineComment; } else if (current == 0) { syncProhibitAutomaticSemicolon(); if (!terminator && !delimited && !prohibitAutomaticSemicolon) { // automatic semicolon insertion if program incomplete token = QScriptGrammar::T_SEMICOLON; stackToken = 0; setDone(Other); } else { setDone(Eof); } } else if (isLineTerminator()) { shiftWindowsLineBreak(); yylineno++; yycolumn = 0; bol = true; terminator = true; syncProhibitAutomaticSemicolon(); if (restrKeyword) { token = QScriptGrammar::T_SEMICOLON; setDone(Other); } } else if (current == '"' || current == '\'') { recordStartPos(); state = InString; stringType = current; } else if (isIdentLetter(current)) { recordStartPos(); record16(current); state = InIdentifier; } else if (current == '0') { recordStartPos(); record8(current); state = InNum0; } else if (isDecimalDigit(current)) { recordStartPos(); record8(current); state = InNum; } else if (current == '.' && isDecimalDigit(next1)) { recordStartPos(); record8(current); state = InDecimal; } else { recordStartPos(); token = matchPunctuator(current, next1, next2, next3); if (token != -1) { if (terminator && !delimited && !prohibitAutomaticSemicolon && (token == QScriptGrammar::T_PLUS_PLUS || token == QScriptGrammar::T_MINUS_MINUS)) { // automatic semicolon insertion stackToken = token; token = QScriptGrammar::T_SEMICOLON; } setDone(Other); } else { setDone(Bad); err = IllegalCharacter; errmsg = QLatin1String("Illegal character"); } } break; case InString: if (current == stringType) { shift(1); setDone(String); } else if (current == 0 || isLineTerminator()) { setDone(Bad); err = UnclosedStringLiteral; errmsg = QLatin1String("Unclosed string at end of line"); } else if (current == '\\') { state = InEscapeSequence; } else { record16(current); } break; // Escape Sequences inside of strings case InEscapeSequence: if (isOctalDigit(current)) { if (current >= '0' && current <= '3' && isOctalDigit(next1) && isOctalDigit(next2)) { record16(convertOctal(current, next1, next2)); shift(2); state = InString; } else if (isOctalDigit(current) && isOctalDigit(next1)) { record16(convertOctal('0', current, next1)); shift(1); state = InString; } else if (isOctalDigit(current)) { record16(convertOctal('0', '0', current)); state = InString; } else { setDone(Bad); err = IllegalEscapeSequence; errmsg = QLatin1String("Illegal escape squence"); } } else if (current == 'x') state = InHexEscape; else if (current == 'u') state = InUnicodeEscape; else { if (isLineTerminator()) { shiftWindowsLineBreak(); yylineno++; yycolumn = 0; bol = true; } else { record16(singleEscape(current)); } state = InString; } break; case InHexEscape: if (isHexDigit(current) && isHexDigit(next1)) { state = InString; record16(QLatin1Char(convertHex(current, next1))); shift(1); } else if (current == stringType) { record16(QLatin1Char('x')); shift(1); setDone(String); } else { record16(QLatin1Char('x')); record16(current); state = InString; } break; case InUnicodeEscape: if (isHexDigit(current) && isHexDigit(next1) && isHexDigit(next2) && isHexDigit(next3)) { record16(convertUnicode(current, next1, next2, next3)); shift(3); state = InString; } else if (current == stringType) { record16(QLatin1Char('u')); shift(1); setDone(String); } else { setDone(Bad); err = IllegalUnicodeEscapeSequence; errmsg = QLatin1String("Illegal unicode escape sequence"); } break; case InSingleLineComment: if (isLineTerminator()) { shiftWindowsLineBreak(); yylineno++; yycolumn = 0; terminator = true; bol = true; if (restrKeyword) { token = QScriptGrammar::T_SEMICOLON; setDone(Other); } else state = Start; } else if (current == 0) { setDone(Eof); } break; case InMultiLineComment: if (current == 0) { setDone(Bad); err = UnclosedComment; errmsg = QLatin1String("Unclosed comment at end of file"); } else if (isLineTerminator()) { shiftWindowsLineBreak(); yylineno++; } else if (current == '*' && next1 == '/') { state = Start; shift(1); } break; case InIdentifier: if (isIdentLetter(current) || isDecimalDigit(current)) { record16(current); break; } setDone(Identifier); break; case InNum0: if (current == 'x' || current == 'X') { record8(current); state = InHex; } else if (current == '.') { record8(current); state = InDecimal; } else if (current == 'e' || current == 'E') { record8(current); state = InExponentIndicator; } else if (isOctalDigit(current)) { record8(current); state = InOctal; } else if (isDecimalDigit(current)) { record8(current); state = InDecimal; } else { setDone(Number); } break; case InHex: if (isHexDigit(current)) record8(current); else setDone(Hex); break; case InOctal: if (isOctalDigit(current)) { record8(current); } else if (isDecimalDigit(current)) { record8(current); state = InDecimal; } else { setDone(Octal); } break; case InNum: if (isDecimalDigit(current)) { record8(current); } else if (current == '.') { record8(current); state = InDecimal; } else if (current == 'e' || current == 'E') { record8(current); state = InExponentIndicator; } else { setDone(Number); } break; case InDecimal: if (isDecimalDigit(current)) { record8(current); } else if (current == 'e' || current == 'E') { record8(current); state = InExponentIndicator; } else { setDone(Number); } break; case InExponentIndicator: if (current == '+' || current == '-') { record8(current); } else if (isDecimalDigit(current)) { record8(current); state = InExponent; } else { setDone(Bad); err = IllegalExponentIndicator; errmsg = QLatin1String("Illegal syntax for exponential number"); } break; case InExponent: if (isDecimalDigit(current)) { record8(current); } else { setDone(Number); } break; default: Q_ASSERT_X(0, "Lexer::lex", "Unhandled state in switch statement"); } // move on to the next character if (!done) shift(1); if (state != Start && state != InSingleLineComment) bol = false; } // no identifiers allowed directly after numeric literal, e.g. "3in" is bad if ((state == Number || state == Octal || state == Hex) && isIdentLetter(current)) { state = Bad; err = IllegalIdentifier; errmsg = QLatin1String("Identifier cannot start with numeric literal"); } // terminate string buffer8[pos8] = '\0'; double dval = 0; if (state == Number) { dval = qstrtod(buffer8, 0, 0); } else if (state == Hex) { // scan hex numbers dval = QScript::integerFromString(buffer8, pos8, 16); state = Number; } else if (state == Octal) { // scan octal number dval = QScript::integerFromString(buffer8, pos8, 8); state = Number; } restrKeyword = false; delimited = false; switch (parenthesesState) { case IgnoreParentheses: break; case CountParentheses: if (token == QScriptGrammar::T_RPAREN) { --parenthesesCount; if (parenthesesCount == 0) parenthesesState = BalancedParentheses; } else if (token == QScriptGrammar::T_LPAREN) { ++parenthesesCount; } break; case BalancedParentheses: parenthesesState = IgnoreParentheses; break; } switch (state) { case Eof: return 0; case Other: if(token == QScriptGrammar::T_RBRACE || token == QScriptGrammar::T_SEMICOLON) delimited = true; return token; case Identifier: if ((token = findReservedWord(buffer16, pos16)) < 0) { /* TODO: close leak on parse error. same holds true for String */ if (driver) { Q_ASSERT_X(false, Q_FUNC_INFO, "not implemented"); qsyylval.ustr = 0; // driver->intern(buffer16, pos16); } else qsyylval.ustr = 0; return QScriptGrammar::T_IDENTIFIER; } if (token == QScriptGrammar::T_CONTINUE || token == QScriptGrammar::T_BREAK || token == QScriptGrammar::T_RETURN || token == QScriptGrammar::T_THROW) { restrKeyword = true; } else if (token == QScriptGrammar::T_IF || token == QScriptGrammar::T_FOR || token == QScriptGrammar::T_WHILE || token == QScriptGrammar::T_WITH) { parenthesesState = CountParentheses; parenthesesCount = 0; } else if (token == QScriptGrammar::T_DO) { parenthesesState = BalancedParentheses; } return token; case String: if (driver) { Q_ASSERT_X(false, Q_FUNC_INFO, "not implemented"); qsyylval.ustr = 0; // driver->intern(buffer16, pos16); } else qsyylval.ustr = 0; return QScriptGrammar::T_STRING_LITERAL; case Number: qsyylval.dval = dval; return QScriptGrammar::T_NUMERIC_LITERAL; case Bad: return -1; default: Q_ASSERT(!"unhandled numeration value in switch"); return -1; } }