squilu/squirrel/sqlexer.cpp

/*
	see copyright notice in squirrel.h
*/
#include "sqpcheader.h"
#include <ctype.h>
#include <stdlib.h>
#include <limits.h>
#include "sqtable.h"
#include "sqstring.h"
#include "sqcompiler.h"
#include "sqlexer.h"

#define CUR_CHAR (_currdata)
#define RETURN_TOKEN(t) { _prevtoken = _curtoken; _curtoken = t; return t;}
#define IS_EOB() (CUR_CHAR <= SQUIRREL_EOB)
#define NEXT() {Next();_currentcolumn++;}
#define INIT_TEMP_STRING() { _longstr.resize(0);}
#define APPEND_CHAR(c) { _longstr.push_back(c);}
#define TERMINATE_BUFFER() {_longstr.push_back(_SC('\0'));}
#define ADD_KEYWORD(key,id) _keywords->NewSlot( SQString::Create(ss, _SC(#key)) ,SQInteger(id))

SQLexer::SQLexer(){}
SQLexer::~SQLexer()
{
	_keywords->Release();
}

void SQLexer::Init(SQSharedState *ss, SQLEXREADFUNC rg, SQUserPointer up,CompilerErrorFunc efunc,void *ed)
{
	_errfunc = efunc;
	_errtarget = ed;
	_sharedstate = ss;
	_keywords = SQTable::Create(ss, 26);
	ADD_KEYWORD(while, TK_WHILE);
	ADD_KEYWORD(do, TK_DO);
	ADD_KEYWORD(if, TK_IF);
	ADD_KEYWORD(else, TK_ELSE);
	ADD_KEYWORD(break, TK_BREAK);
	ADD_KEYWORD(continue, TK_CONTINUE);
	ADD_KEYWORD(return, TK_RETURN);
	ADD_KEYWORD(null, TK_NULL);
	ADD_KEYWORD(NULL, TK_NULL);
	ADD_KEYWORD(function, TK_FUNCTION);
	ADD_KEYWORD(local, TK_LOCAL);
	ADD_KEYWORD(var, TK_LOCAL);

	ADD_KEYWORD(char_t, TK_LOCAL_CHAR_T);
	ADD_KEYWORD(wchar_t, TK_LOCAL_WCHAR_T);

	ADD_KEYWORD(bool_t, TK_LOCAL_BOOL_T);
	ADD_KEYWORD(table_t, TK_LOCAL_TABLE_T);
	ADD_KEYWORD(array_t, TK_LOCAL_ARRAY_T);

	ADD_KEYWORD(int8_t, TK_LOCAL_INT8_T);
	ADD_KEYWORD(int16_t, TK_LOCAL_INT16_T);
	ADD_KEYWORD(int32_t, TK_LOCAL_INT32_T);
	ADD_KEYWORD(int64_t, TK_LOCAL_INT64_T);
	ADD_KEYWORD(int_t, TK_LOCAL_INT_T);

	ADD_KEYWORD(uint8_t, TK_LOCAL_UINT8_T);
	ADD_KEYWORD(uint16_t, TK_LOCAL_UINT16_T);
	ADD_KEYWORD(uint32_t, TK_LOCAL_UINT32_T);
	ADD_KEYWORD(uint64_t, TK_LOCAL_UINT64_T);
	ADD_KEYWORD(uint_t, TK_LOCAL_UINT_T);

	ADD_KEYWORD(float_t, TK_LOCAL_FLOAT_T);
	ADD_KEYWORD(double_t, TK_LOCAL_DOUBLE_T);
	ADD_KEYWORD(long_double_t, TK_LOCAL_LONG_DOUBLE_T);

	ADD_KEYWORD(for, TK_FOR);
	ADD_KEYWORD(foreach, TK_FOREACH);
	ADD_KEYWORD(in, TK_IN);
	ADD_KEYWORD(typeof, TK_TYPEOF);
	ADD_KEYWORD(base, TK_BASE);
	ADD_KEYWORD(delete, TK_DELETE);
	ADD_KEYWORD(try, TK_TRY);
	ADD_KEYWORD(catch, TK_CATCH);
	ADD_KEYWORD(throw, TK_THROW);
	ADD_KEYWORD(clone, TK_CLONE);
	ADD_KEYWORD(yield, TK_YIELD);
	ADD_KEYWORD(resume, TK_RESUME);
	ADD_KEYWORD(switch, TK_SWITCH);
	ADD_KEYWORD(case, TK_CASE);
	ADD_KEYWORD(default, TK_DEFAULT);
	ADD_KEYWORD(this, TK_THIS);
	ADD_KEYWORD(class,TK_CLASS);
	ADD_KEYWORD(extends,TK_EXTENDS);
	ADD_KEYWORD(constructor,TK_CONSTRUCTOR);
	ADD_KEYWORD(instanceof,TK_INSTANCEOF);
	ADD_KEYWORD(true,TK_TRUE);
	ADD_KEYWORD(false,TK_FALSE);
	ADD_KEYWORD(static,TK_STATIC);
	ADD_KEYWORD(enum,TK_ENUM);
	ADD_KEYWORD(const,TK_CONST);
	ADD_KEYWORD(__LINE__,TK___LINE__);
	ADD_KEYWORD(__FILE__,TK___FILE__);
	ADD_KEYWORD(new,TK_IGNORE);

	_readf = rg;
	_up = up;
	_lasttokenline = _currentline = 1;
	_currentcolumn = 0;
	_prevtoken = -1;
	_reached_eof = SQFalse;
	Next();
}

void SQLexer::Error(const SQChar *err)
{
	_errfunc(_errtarget,err);
}

void SQLexer::Next()
{
	SQInteger t = _readf(_up);
	if(t > MAX_CHAR) Error(_SC("Invalid character"));
	if(t != 0) {
		_currdata = (LexChar)t;
		return;
	}
	_currdata = SQUIRREL_EOB;
	_reached_eof = SQTrue;
}

const SQChar *SQLexer::Tok2Str(SQInteger tok)
{
	SQObjectPtr itr, key, val;
	SQInteger nitr;
	while((nitr = _keywords->Next(false,itr, key, val)) != -1) {
		itr = (SQInteger)nitr;
		if(((SQInteger)_integer(val)) == tok)
			return _stringval(key);
	}
	return NULL;
}

void SQLexer::LexBlockComment()
{
	bool done = false;
	while(!done) {
		switch(CUR_CHAR) {
			case _SC('*'): { NEXT(); if(CUR_CHAR == _SC('/')) { done = true; NEXT(); }}; continue;
			case _SC('\n'): _currentline++; NEXT(); continue;
			case SQUIRREL_EOB: Error(_SC("missing \"*/\" in comment"));
			default: NEXT();
		}
	}
}
void SQLexer::LexLineComment()
{
	do { NEXT(); } while (CUR_CHAR != _SC('\n') && (!IS_EOB()));
}

SQInteger SQLexer::Lex()
{
	_lasttokenline = _currentline;
	while(CUR_CHAR != SQUIRREL_EOB) {
		switch(CUR_CHAR){
		case _SC('\t'): case _SC('\r'): case _SC(' '): NEXT(); continue;
		case _SC('\n'):
			_currentline++;
			_prevtoken=_curtoken;
			_curtoken=_SC('\n');
			NEXT();
			_currentcolumn=1;
			continue;
		case _SC('#'): LexLineComment(); continue;
		case _SC('/'):
			NEXT();
			switch(CUR_CHAR){
			case _SC('*'):
				NEXT();
				LexBlockComment();
				continue;
			case _SC('/'):
				LexLineComment();
				continue;
			case _SC('='):
				NEXT();
				RETURN_TOKEN(TK_DIVEQ);
				continue;
			case _SC('>'):
				NEXT();
				RETURN_TOKEN(TK_ATTR_CLOSE);
				continue;
			default:
				RETURN_TOKEN('/');
			}
		case _SC('='):
			NEXT();
			if (CUR_CHAR != _SC('=')){ RETURN_TOKEN('=') }
			else { NEXT(); RETURN_TOKEN(TK_EQ); }
		case _SC('<'):
			NEXT();
			switch(CUR_CHAR) {
			case _SC('='):
				NEXT();
				if(CUR_CHAR == _SC('>')) {
					NEXT();
					RETURN_TOKEN(TK_3WAYSCMP);
				}
				RETURN_TOKEN(TK_LE)
				break;
			case _SC('-'): NEXT(); RETURN_TOKEN(TK_NEWSLOT); break;
			case _SC('<'): NEXT(); RETURN_TOKEN(TK_SHIFTL); break;
			case _SC('/'): NEXT(); RETURN_TOKEN(TK_ATTR_OPEN); break;
			}
			RETURN_TOKEN('<');
		case _SC('>'):
			NEXT();
			if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_GE);}
			else if(CUR_CHAR == _SC('>')){
				NEXT();
				if(CUR_CHAR == _SC('>')){
					NEXT();
					RETURN_TOKEN(TK_USHIFTR);
				}
				RETURN_TOKEN(TK_SHIFTR);
			}
			else { RETURN_TOKEN('>') }
		case _SC('!'):
			NEXT();
			if (CUR_CHAR != _SC('=')){ RETURN_TOKEN('!')}
			else { NEXT(); RETURN_TOKEN(TK_NE); }
		case _SC('@'): {
			SQInteger stype;
			NEXT();
			if(CUR_CHAR != _SC('"')) {
				RETURN_TOKEN('@');
			}
			if((stype=ReadString('"',true))!=-1) {
				RETURN_TOKEN(stype);
			}
			Error(_SC("error parsing the string"));
					   }
		case _SC('"'):
		case _SC('\''): {
			SQInteger stype;
			if((stype=ReadString(CUR_CHAR,false))!=-1){
				RETURN_TOKEN(stype);
			}
			Error(_SC("error parsing the string"));
			}
		case _SC('{'): case _SC('}'): case _SC('('): case _SC(')'): case _SC('['): case _SC(']'):
		case _SC(';'): case _SC(','): case _SC('?'): case _SC('~'):
			{
			    SQInteger ret = CUR_CHAR;
                NEXT();
                if(ret == _SC('[') && CUR_CHAR == _SC('=')){
                    //lets try lua literal delimiters
                    SQInteger stype;
                    if((stype=ReadString(CUR_CHAR,true))!=-1){
                        RETURN_TOKEN(stype);
                    }
                    Error(_SC("error parsing the string"));
                }
                else RETURN_TOKEN(ret);
            }
		case _SC('.'):
			NEXT();
			if (CUR_CHAR != _SC('.')){ RETURN_TOKEN('.') }
			NEXT();
			if (CUR_CHAR != _SC('.')){ Error(_SC("invalid token '..'")); }
			NEXT();
			RETURN_TOKEN(TK_VARPARAMS);
		case _SC('^'):
			NEXT();
			//if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_BIT_XOR_EQ);}
			RETURN_TOKEN('^');
		case _SC('&'):
			NEXT();
			//if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_BIT_AND_EQ);}
			if (CUR_CHAR != _SC('&')){ RETURN_TOKEN('&') }
			else { NEXT(); RETURN_TOKEN(TK_AND); }
		case _SC('|'):
			NEXT();
			//if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_BIT_OR_EQ);}
			if (CUR_CHAR != _SC('|')){ RETURN_TOKEN('|') }
			else { NEXT(); RETURN_TOKEN(TK_OR); }
		case _SC(':'):
			NEXT();
			if (CUR_CHAR != _SC(':')){ RETURN_TOKEN(':') }
			else { NEXT(); RETURN_TOKEN(TK_DOUBLE_COLON); }
		case _SC('*'):
			NEXT();
			if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_MULEQ);}
			else RETURN_TOKEN('*');
		case _SC('%'):
			NEXT();
			if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_MODEQ);}
			else RETURN_TOKEN('%');
		case _SC('-'):
			NEXT();
			if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_MINUSEQ);}
			else if  (CUR_CHAR == _SC('-')){ NEXT(); RETURN_TOKEN(TK_MINUSMINUS);}
			else if  (CUR_CHAR == _SC('>')){ NEXT(); RETURN_TOKEN('.');} //accept C/C++ like pointers
			else RETURN_TOKEN('-');
		case _SC('+'):
			NEXT();
			if (CUR_CHAR == _SC('=')){ NEXT(); RETURN_TOKEN(TK_PLUSEQ);}
			else if (CUR_CHAR == _SC('+')){ NEXT(); RETURN_TOKEN(TK_PLUSPLUS);}
			else RETURN_TOKEN('+');
		case SQUIRREL_EOB:
			return 0;
		default:{
				if (scisdigit(CUR_CHAR)) {
					SQInteger ret = ReadNumber();
					RETURN_TOKEN(ret);
				}
				else if (scisalpha(CUR_CHAR) || CUR_CHAR == _SC('_')) {
					SQInteger t = ReadID();
					RETURN_TOKEN(t);
				}
				else {
					SQInteger c = CUR_CHAR;
					if (sciscntrl((int)c)) Error(_SC("unexpected character(control)"));
					NEXT();
					RETURN_TOKEN(c);
				}
				RETURN_TOKEN(0);
			}
		}
	}
	return 0;
}

SQInteger SQLexer::GetIDType(SQChar *s)
{
	SQObjectPtr t;
	if(_keywords->Get(SQString::Create(_sharedstate, s), t)) {
		return SQInteger(_integer(t));
	}
	return TK_IDENTIFIER;
}

SQInteger SQLexer::ReadString(SQInteger ndelim,bool verbatim)
{
	INIT_TEMP_STRING();
	SQInteger start_equals = 0;
	if(ndelim == _SC('=')){
	    //lua like literal
	    while(!IS_EOB() && CUR_CHAR == _SC('=')) {
	        ++start_equals;
	        NEXT();
	    }
	    if(CUR_CHAR != _SC('[')){
	        //it's not a lua literal delimiter
	        Error(_SC("expect '[' on literal delimiter"));
	        return -1;
	    }
	    ndelim = _SC(']');
	}
	NEXT();
	if(IS_EOB()) return -1;
	if(start_equals && CUR_CHAR == _SC('\n')) {
	    ++_currentline;
	    NEXT(); //if a new line follows the start of delimiter drop it
        if(IS_EOB()) return -1;
	}
	for(;;) {
		while(CUR_CHAR != ndelim) {
			switch(CUR_CHAR) {
			case SQUIRREL_EOB:
				Error(_SC("unfinished string"));
				return -1;
			case _SC('\n'):
				if(!verbatim) Error(_SC("newline in a constant"));
				APPEND_CHAR(CUR_CHAR); NEXT();
				_currentline++;
				break;
			case _SC('\\'):
				if(verbatim) {
					APPEND_CHAR('\\'); NEXT();
				}
				else {
					NEXT();
					switch(CUR_CHAR) {
					case _SC('x'): NEXT(); {
						if(!isxdigit(CUR_CHAR)) Error(_SC("hexadecimal number expected"));
						const SQInteger maxdigits = 4;
						SQChar temp[maxdigits+1];
						SQInteger n = 0;
						while(isxdigit(CUR_CHAR) && n < maxdigits) {
							temp[n] = CUR_CHAR;
							n++;
							NEXT();
						}
						temp[n] = 0;
						SQChar *sTemp;
						APPEND_CHAR((SQChar)scstrtoul(temp,&sTemp,16));
					}
				    break;
					case _SC('t'): APPEND_CHAR(_SC('\t')); NEXT(); break;
					case _SC('a'): APPEND_CHAR(_SC('\a')); NEXT(); break;
					case _SC('b'): APPEND_CHAR(_SC('\b')); NEXT(); break;
					case _SC('n'): APPEND_CHAR(_SC('\n')); NEXT(); break;
					case _SC('r'): APPEND_CHAR(_SC('\r')); NEXT(); break;
					case _SC('v'): APPEND_CHAR(_SC('\v')); NEXT(); break;
					case _SC('f'): APPEND_CHAR(_SC('\f')); NEXT(); break;
					case _SC('0'): APPEND_CHAR(_SC('\0')); NEXT(); break;
					case _SC('\\'): APPEND_CHAR(_SC('\\')); NEXT(); break;
					case _SC('"'): APPEND_CHAR(_SC('"')); NEXT(); break;
					case _SC('\''): APPEND_CHAR(_SC('\'')); NEXT(); break;
					default:
						Error(_SC("unrecognised escaper char"));
					break;
					}
				}
				break;
			default:
				APPEND_CHAR(CUR_CHAR);
				NEXT();
			}
		}
		NEXT();
		if(start_equals){
		    bool lastBraceAdded = false;
		    if(CUR_CHAR == _SC('=')){
                SQInteger end_equals = start_equals;
                NEXT();
                if(CUR_CHAR == _SC('=') || CUR_CHAR == _SC(']')){
                    --end_equals;
                    while(!IS_EOB() && CUR_CHAR == _SC('=')) {
                        --end_equals;
                        NEXT();
                    }
                    if(end_equals) Error(_SC("expect same number of '=' on literal delimiter"));
                    if(CUR_CHAR != _SC(']')) Error(_SC("expect ']' to close literal delimiter"));
                    NEXT();
                    break;
                }
                APPEND_CHAR(_SC(']')); //the first NEXT() after break the while loop
                APPEND_CHAR(_SC('='));
                lastBraceAdded = true;
		    }
		    if(!lastBraceAdded) APPEND_CHAR(_SC(']')); //the first NEXT() after break the while loop
		    APPEND_CHAR(CUR_CHAR);
		    NEXT();
		}
		else if(verbatim && CUR_CHAR == '"') { //double quotation
			APPEND_CHAR(CUR_CHAR);
			NEXT();
		}
		else {
			break;
		}
	}
	TERMINATE_BUFFER();
	SQInteger len = _longstr.size()-1;
	if(ndelim == _SC('\'')) {
		if(len == 0) Error(_SC("empty constant"));
		if(len > 1) Error(_SC("constant too long"));
		_nvalue = _longstr[0];
		return TK_INTEGER;
	}
	_svalue = &_longstr[0];
	return TK_STRING_LITERAL;
}

void LexHexadecimal(const SQChar *s,SQUnsignedInteger *res)
{
	*res = 0;
	while(*s != 0)
	{
		if(scisdigit(*s)) *res = (*res)*16+((*s++)-'0');
		else if(scisxdigit(*s)) *res = (*res)*16+(toupper(*s++)-'A'+10);
		else { assert(0); }
	}
}

void LexInteger(const SQChar *s,SQUnsignedInteger *res)
{
	*res = 0;
	while(*s != 0)
	{
		*res = (*res)*10+((*s++)-'0');
	}
}

SQInteger scisodigit(SQInteger c) { return c >= _SC('0') && c <= _SC('7'); }

void LexOctal(const SQChar *s,SQUnsignedInteger *res)
{
	*res = 0;
	while(*s != 0)
	{
		if(scisodigit(*s)) *res = (*res)*8+((*s++)-'0');
		else { assert(0); }
	}
}

SQInteger isexponent(SQInteger c) { return c == 'e' || c=='E'; }


#define MAX_HEX_DIGITS (sizeof(SQInteger)*2)
SQInteger SQLexer::ReadNumber()
{
#define TINT 1
#define TFLOAT 2
#define THEX 3
#define TSCIENTIFIC 4
#define TOCTAL 5
	SQInteger type = TINT, firstchar = CUR_CHAR;
	SQUnsignedInteger itmp;
	SQChar *sTemp;
	INIT_TEMP_STRING();
	NEXT();
	if(firstchar == _SC('0') && (toupper(CUR_CHAR) == _SC('X') || scisodigit(CUR_CHAR)) ) {
		if(scisodigit(CUR_CHAR)) {
			type = TOCTAL;
			while(scisodigit(CUR_CHAR)) {
				APPEND_CHAR(CUR_CHAR);
				NEXT();
			}
			if(scisdigit(CUR_CHAR)) Error(_SC("invalid octal number"));
		}
		else {
			NEXT();
			type = THEX;
			while(isxdigit(CUR_CHAR)) {
				APPEND_CHAR(CUR_CHAR);
				NEXT();
			}
			if(_longstr.size() > MAX_HEX_DIGITS) Error(_SC("too many digits for an Hex number"));
		}
	}
	else {
		APPEND_CHAR((int)firstchar);
		while (CUR_CHAR == _SC('.') || scisdigit(CUR_CHAR) || isexponent(CUR_CHAR)) {
            if(CUR_CHAR == _SC('.') || isexponent(CUR_CHAR)) type = TFLOAT;
			if(isexponent(CUR_CHAR)) {
				if(type != TFLOAT) Error(_SC("invalid numeric format"));
				type = TSCIENTIFIC;
				APPEND_CHAR(CUR_CHAR);
				NEXT();
				if(CUR_CHAR == '+' || CUR_CHAR == '-'){
					APPEND_CHAR(CUR_CHAR);
					NEXT();
				}
				if(!scisdigit(CUR_CHAR)) Error(_SC("exponent expected"));
			}

			APPEND_CHAR(CUR_CHAR);
			NEXT();
		}
	}
	TERMINATE_BUFFER();
	switch(type) {
	case TSCIENTIFIC:
	case TFLOAT:
		_fvalue = (SQFloat)scstrtod(&_longstr[0],&sTemp);
		return TK_FLOAT;
	case TINT:
		LexInteger(&_longstr[0],&itmp);
		break;
	case THEX:
		LexHexadecimal(&_longstr[0],&itmp);
		break;
	case TOCTAL:
		LexOctal(&_longstr[0],&itmp);
		break;
	}
	switch(type) {
	case TINT:
	case THEX:
	case TOCTAL:
        if(itmp > INT_MAX) Error(_SC("integer overflow"));
        _nvalue = (SQInteger) itmp;
		return TK_INTEGER;
	}
	return 0;
}

SQInteger SQLexer::ReadID()
{
	SQInteger res;
	INIT_TEMP_STRING();
	do {
		APPEND_CHAR(CUR_CHAR);
		NEXT();
	} while(scisalnum(CUR_CHAR) || CUR_CHAR == _SC('_'));
	TERMINATE_BUFFER();
	res = GetIDType(&_longstr[0]);
	if(res == TK_IDENTIFIER || res == TK_CONSTRUCTOR) {
		_svalue = &_longstr[0];
	}
	return res;
}