lex.c

CFront1.0的源代码,第一代C++编译器的思想...

			get(lxchar);

		switch( (p=lxcp[lxchar+1])->lxact ){

		case A_1C:	// eat up a single character, and return an opcode 
			reti(p->lxtok,p->lxtok);

		case A_EOF:
			if (br_level || bl_level)
				error("'%s' missing at end of input",(bl_level) ? "}" : ")");
			reti(EOFTOK,0);

		case A_ERR:
			if (lxchar == '#') {	// cope with header file not ended
						// with '\n'
				unget('#');
				saved = lxtitle();
				continue;
			}
			if (' '<=lxchar && lxchar<='~')	// ASCII printable
				error("illegal character '%c' (ignored)",lxchar);
			else
				error("illegal character '0%o' (ignored)",lxchar);
			continue;

		case A_LET:
			/* collect an identifier, check for reserved word, and return */
			lxget( lxchar, LEXLET|LEXDIG );

			if (n = ktbl->look(txtstart,0)) {
				TOK x;
				del_txt();
				switch (x=n->base) {
				case TNAME:
					retn(TNAME,n);
					break;
				case LOC:
					retl(n->syn_class);
				default:
					reti(n->syn_class,x);
				}
			}
			else {
				rets(ID,txtstart);
			}

		case A_DIG:

			ret = ICON;

			if (lxchar=='0') {	/* octal or hexadecimal number */
				pch('0');
				switch (get(lxchar)) {
				case 'l':
				case 'L':
					pch('L');
					pch(0);
					rets(ICON,txtstart);
				case 'x':
				case 'X':
					lxget('X',LEXHEX);
					if (txtfree-txtstart<4) // minimum "0Xd\0"
						error("hexadecimal digitX after \"0x\"");
					switch (get(lxchar)) {
					case 'l':
					case 'L':
						txtfree--;
						pch('L');
						pch(0);
						break;
					default:
						saved = lxchar;
					}
					rets(ICON,txtstart);
				case '8':
				case '9':
					error("8 or 9 used as octal digit");
				case '0':
				case '1':
				case '2':
				case '3':
				case '4':
				case '5':
				case '6':
				case '7':
					pch(lxchar);
				ox:
					switch (get(lxchar)) {
					case '8':
					case '9':
						error("8 or 9 used as octal digit");
					case '0':
					case '1':
					case '2':
					case '3':
					case '4':
					case '5':
					case '6':
					case '7':
						pch(lxchar);
						goto ox;
					case 'l':
					case 'L':
						pch('L');
						pch(0);
						break;
					default:
						pch(0);
						saved = lxchar;
					}
					rets(ICON,txtstart);
				case '.':
					lxget('.',LEXDIG);
					goto getfp;
				default:
					saved = lxchar;
					reti(ZERO,0);
				}
			}
			else
				lxget(lxchar,LEXDIG);

			if (get(lxchar) == '.') {
				txtfree--;
				lxget('.', LEXDIG );
		getfp:
				ret = FCON;
				get(lxchar);
			};

			switch (lxchar) {
			case 'e':
			case 'E':
				txtfree--;
				switch (get(lxchar)) {
				case '-':
				case '+':
					pch('e');
					break;
				default:
					unget(lxchar);
					lxchar = 'e';
				};
				lxget( lxchar, LEXDIG );
				ret = FCON;
				break;
			case 'l':
			case 'L':
				txtfree--;
				pch('L');
				break;
			default:
				saved = lxchar;
			};

			pch(0);
			rets(ret,txtstart);

		case A_DOT:
			if (get(lxchar) == '.') {	/* look for ellipsis */
				if (get(lxchar) != '.') {
					error("token .. ?");
					saved = lxchar;
				}
				reti(ELLIPSIS,0);
			}
			if( lxmask[lxchar+1] & LEXDIG ){/* look for floating constant */
				unget(lxchar);
				lxget( '.', LEXDIG );
				goto getfp;
			}
			saved = lxchar;
			reti(DOT,0);

		case A_STR:
			/* save string constant in buffer */
			forever
			switch (get(lxchar)) {
			case '\\':
				pch('\\');
				get(lxchar);
				if (lxchar == 'x')
					hex_to_oct();
				else
					pch(lxchar);
				break;
			case '"':
				pch(0);
				rets(STRING,txtstart);
			case '\n':
				error("newline in string");
				pch(0);
				rets(STRING,txtstart);
			case EOF:
				error("eof in string");
				pch(0);
				rets(STRING,txtstart);
			default:
				pch(lxchar);
			}

		case A_CC:
			/* character constant */
			rets(CCON,chconst());

		case A_BCD:
			{
				register i;
				int j;
	
				pch('`');
	
				for (i=0; i<7; ++i) {
					pch(get(j));
					if (j == '`' ) break;
				}
				pch(0);
				if (6<i)
					error("bcd constant exceeds 6 characters" );
				rets(CCON,txtstart);
			}

		case A_SL:	/* / */
			switch (get(lxchar))  {
			case '*':
				lxcom();
				break;
			case '/':
				linecom();
				break;
			case '=':
				reti(ASOP,ASDIV);
			default:
				saved = lxchar;
				reti(DIVOP,DIV);
			}

		case A_WS:
			continue;

		case A_NL:
			++curloc.line;
			Nline++;
			saved = lxtitle();
			continue;

		case A_LC:
			if (BLMAX <= bl_level++) {
				error('s',"blocks too deaply nested");
				ext(3);
			}
			retl(LC);

		case A_RC:
			if (bl_level-- <= 0) {
				error("unX '}'");
				bl_level = 0;
			}
			retl(RC);

		case A_L:
			br_level++;
			reti(LP,0);

		case A_R:
			if (br_level-- <= 0) {
				error("unX ')'");
				br_level = 0;
			}
			reti(RP,0);

		case A_ASS:
			switch (get(lxchar)) {
			case '=':
				reti(EQUOP,EQ);
			default:
				saved = lxchar;
				reti(ASSIGN,ASSIGN);
			}

		case A_COL:
			switch (get(lxchar)) {
			case ':':
				reti(MEM,0);
			case '=':
				error("':=' is not a c++ operator");
				reti(ASSIGN,ASSIGN);
			default:
				saved = lxchar;
				reti(COLON,COLON);
			}
		case A_NOT:
			switch (get(lxchar)) {
			case '=':
				reti(EQUOP,NE);
			default:
				saved = lxchar;
				reti(NOT,NOT);
			}
		case A_GT:
			switch(get(lxchar)) {
			case '>':
				switch (get(lxchar)) {
				case '=':
					reti(ASOP,ASRS);
					break;
				default:
					saved = lxchar;
					reti(SHIFTOP,RS);
				}
			case '=':
				reti(RELOP,GE);
			default:
				saved = lxchar;
				reti(RELOP,GT);
			}
		case A_LT:
			switch (get(lxchar)) {
			case '<':
				switch (get(lxchar)) {
				case '=':
					reti(ASOP,ASLS);
				default:
					saved = lxchar;
					reti(SHIFTOP,LS);
				}
			case '=':
				reti(RELOP,LE);
			default:
				saved = lxchar;
				reti(RELOP,LT);
			}
		case A_AND:
			switch (get(lxchar)) {
			case '&':
				reti(ANDAND,ANDAND);
			case '=':
				reti(ASOP,ASAND);
			default:
				saved = lxchar;
				reti(AND,AND);
			}
		case A_OR:
			switch (get(lxchar)) {
			case '|':
				reti(OROR,OROR);
			case '=':
				reti(ASOP,ASOR);
			default:
				saved = lxchar;
				reti(OR,OR);
			}
		case A_ER:
			switch (get(lxchar)) {
			case '=':
				reti(ASOP,ASER);
			default:
				saved = lxchar;
				reti(ER,ER);
			}
		case A_PL:
			switch (get(lxchar)) {
			case '=':
				reti(ASOP,ASPLUS);
			case '+':
				reti(ICOP,INCR);
			default:
				saved = lxchar;
				reti(PLUS,PLUS);
			}
		case A_MIN:
			switch (get(lxchar)) {
			case '=':
				reti(ASOP,ASMINUS);
			case '-':
				reti(ICOP,DECR);
			case '>':
				reti(REF,REF);
			default:
				saved = lxchar;
				reti(MINUS,MINUS);
			}
		case A_MUL:
			switch (get(lxchar)) {
			case '=':
				reti(ASOP,ASMUL);
			case '/':
				error('w',"*/ not as end of comment");
			default:
				saved = lxchar;
				reti(MUL,MUL);
			}
		case A_MOD:
			switch (get(lxchar)) {
			case '=':
				reti(ASOP,ASMOD);
			default:
				saved = lxchar;
				reti(DIVOP,MOD);
			}
		default:
			error('i',"lex act==%d getc()->%d",p,lxchar);

		}

		error('i',"lex, main switch");

	}

}

int lxtitle()
/*
	called after a newline; set linenumber and file name
*/
{
	register c;

	forever
	switch ( get(c) ) {
	default:
		return c;
/*	case EOF:
		return EOF;	*/
	case '\n':
		curloc.line++;
		Nline++;
		break;
	ll:
		break;
	case '#': 	/* # lineno "filename" */
		curloc.line = 0;
		forever 
		switch (get(c)) {
		case '"':
			start_txt();
			forever
			switch (get(c)) {
			case '"':
				pch('\0');
				if (get(c) != '\n') error("unX eol on # line");
				if (*txtstart) {	// stack file name
					if (curr_file == 0) goto push;

					char* fn;

					if ( (fn=file_name[file_stack[curr_file]])
					&& (strcmp(txtstart,fn)==0) ) {
						// same file: ignore
					}
					else if ( (fn=file_name[file_stack[curr_file-1]])
					&& (strcmp(txtstart,fn)==0) ) {
						// previous file: pop
						curr_file--;
					}
					else {	// new file name: push
				push:
						if (MAXFILE*4<Nfile++) error('i',"fileN buffer overflow");
						if (MAXFILE<curr_file++) error('i',"fileN stack overflow");
						file_stack[curr_file] = Nfile;
						
						char* p = new char[txtfree-txtstart+1];
						(void) strcpy(p,txtstart);
						file_name[Nfile] = p;
						Nstr++;
					}
				}
				else {	// back to the original .c file: ""
					curr_file = 0;
				}
				del_txt();
				curloc.file = file_stack[curr_file];
				goto ll;
			case '\n':
				error("unX end of line on '# line'");
			default:
				pch(c);
			}
		case ' ':
			break;
		case '0':
		case '1':
		case '2':
		case '3':
		case '4':
		case '5':
		case '6':
		case '7':
		case '8':
		case '9':
			curloc.line = curloc.line*10+c-'0'; 
			break;
		default:	/* pass #rubbish through */
			pch('#');
			pch('i');
			while (get(c) != '\n') pch(c);
			pch('\0');
			fprintf(out_file,"\n%s\n",txtstart);
			curloc.line++;
			Nline++;
			goto ll;
		case '\n':
			curloc.putline();
			goto ll;
		}
	}
}