toke.c

早期freebsd实现

    case ']':
	tmp = *s++;
	TERM(tmp);
    case '}':
	*s |= 128;
	RETURN(';');
    case '&':
	s++;
	tmp = *s++;
	if (tmp == '&')
	    OPERATOR(ANDAND);
	s--;
	if (expectterm) {
	    d = bufend;
	    while (s < d && isSPACE(*s))
		s++;
	    if (isALPHA(*s) || *s == '_' || *s == '\'')
		*(--s) = '\\';	/* force next ident to WORD */
	    else
		check_uni();
	    OPERATOR(AMPER);
	}
	OPERATOR('&');
    case '|':
	s++;
	tmp = *s++;
	if (tmp == '|')
	    OPERATOR(OROR);
	s--;
	OPERATOR('|');
    case '=':
	s++;
	tmp = *s++;
	if (tmp == '=')
	    EOP(O_EQ);
	if (tmp == '~')
	    OPERATOR(MATCH);
	s--;
	OPERATOR('=');
    case '!':
	s++;
	tmp = *s++;
	if (tmp == '=')
	    EOP(O_NE);
	if (tmp == '~')
	    OPERATOR(NMATCH);
	s--;
	OPERATOR('!');
    case '<':
	if (expectterm) {
	    if (s[1] != '<' && !index(s,'>'))
		check_uni();
	    s = scanstr(s, SCAN_DEF);
	    TERM(RSTRING);
	}
	s++;
	tmp = *s++;
	if (tmp == '<')
	    OPERATOR(LS);
	if (tmp == '=') {
	    tmp = *s++;
	    if (tmp == '>')
		EOP(O_NCMP);
	    s--;
	    ROP(O_LE);
	}
	s--;
	ROP(O_LT);
    case '>':
	s++;
	tmp = *s++;
	if (tmp == '>')
	    OPERATOR(RS);
	if (tmp == '=')
	    ROP(O_GE);
	s--;
	ROP(O_GT);

#define SNARFWORD \
	d = tokenbuf; \
	while (isALNUM(*s) || *s == '\'') \
	    *d++ = *s++; \
	while (d[-1] == '\'') \
	    d--,s--; \
	*d = '\0'; \
	d = tokenbuf;

    case '$':
	if (s[1] == '#' && (isALPHA(s[2]) || s[2] == '_')) {
	    s++;
	    s = scanident(s,bufend,tokenbuf);
	    yylval.stabval = aadd(stabent(tokenbuf,TRUE));
	    TERM(ARYLEN);
	}
	d = s;
	s = scanident(s,bufend,tokenbuf);
	if (reparse) {		/* turn ${foo[bar]} into ($foo[bar]) */
	  do_reparse:
	    s[-1] = ')';
	    s = d;
	    s[1] = s[0];
	    s[0] = '(';
	    goto retry;
	}
	yylval.stabval = stabent(tokenbuf,TRUE);
	expectterm = FALSE;
	if (isSPACE(*s) && oldoldbufptr && oldoldbufptr < bufptr) {
	    s++;
	    while (isSPACE(*oldoldbufptr))
		oldoldbufptr++;
	    if (*oldoldbufptr == 'p' && strnEQ(oldoldbufptr,"print",5)) {
		if (index("&*<%", *s) && isALPHA(s[1]))
		    expectterm = TRUE;		/* e.g. print $fh &sub */
		else if (*s == '.' && isDIGIT(s[1]))
		    expectterm = TRUE;		/* e.g. print $fh .3 */
		else if (index("/?-+", *s) && !isSPACE(s[1]))
		    expectterm = TRUE;		/* e.g. print $fh -1 */
	    }
	}
	RETURN(REG);

    case '@':
	d = s;
	s = scanident(s,bufend,tokenbuf);
	if (reparse)
	    goto do_reparse;
	yylval.stabval = aadd(stabent(tokenbuf,TRUE));
	TERM(ARY);

    case '/':			/* may either be division or pattern */
    case '?':			/* may either be conditional or pattern */
	if (expectterm) {
	    check_uni();
	    s = scanpat(s);
	    TERM(PATTERN);
	}
	tmp = *s++;
	if (tmp == '/')
	    MOP(O_DIVIDE);
	OPERATOR(tmp);

    case '.':
	if (!expectterm || !isDIGIT(s[1])) {
	    tmp = *s++;
	    if (*s == tmp) {
		s++;
		if (*s == tmp) {
		    s++;
		    yylval.ival = 0;
		}
		else
		    yylval.ival = AF_COMMON;
		OPERATOR(DOTDOT);
	    }
	    if (expectterm)
		check_uni();
	    AOP(O_CONCAT);
	}
	/* FALL THROUGH */
    case '0': case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8': case '9':
    case '\'': case '"': case '`':
	s = scanstr(s, SCAN_DEF);
	TERM(RSTRING);

    case '\\':	/* some magic to force next word to be a WORD */
	s++;	/* used by do and sub to force a separate namespace */
	if (!isALPHA(*s) && *s != '_' && *s != '\'') {
	    warn("Spurious backslash ignored");
	    goto retry;
	}
	/* FALL THROUGH */
    case '_':
	SNARFWORD;
	if (d[1] == '_') {
	    if (strEQ(d,"__LINE__") || strEQ(d,"__FILE__")) {
		ARG *arg = op_new(1);

		yylval.arg = arg;
		arg->arg_type = O_ITEM;
		if (d[2] == 'L')
		    (void)sprintf(tokenbuf,"%ld",(long)curcmd->c_line);
		else
		    strcpy(tokenbuf, stab_val(curcmd->c_filestab)->str_ptr);
		arg[1].arg_type = A_SINGLE;
		arg[1].arg_ptr.arg_str = str_make(tokenbuf,strlen(tokenbuf));
		TERM(RSTRING);
	    }
	    else if (strEQ(d,"__END__")) {
		STAB *stab;
		int fd;

		/*SUPPRESS 560*/
		if (!in_eval && (stab = stabent("DATA",FALSE))) {
		    stab->str_pok |= SP_MULTI;
		    if (!stab_io(stab))
			stab_io(stab) = stio_new();
		    stab_io(stab)->ifp = rsfp;
#if defined(HAS_FCNTL) && defined(F_SETFD)
		    fd = fileno(rsfp);
		    fcntl(fd,F_SETFD,fd >= 3);
#endif
		    if (preprocess)
			stab_io(stab)->type = '|';
		    else if ((FILE*)rsfp == stdin)
			stab_io(stab)->type = '-';
		    else
			stab_io(stab)->type = '<';
		    rsfp = Nullfp;
		}
		goto fake_eof;
	    }
	}
	break;
    case 'a': case 'A':
	SNARFWORD;
	if (strEQ(d,"alarm"))
	    UNI(O_ALARM);
	if (strEQ(d,"accept"))
	    FOP22(O_ACCEPT);
	if (strEQ(d,"atan2"))
	    FUN2(O_ATAN2);
	break;
    case 'b': case 'B':
	SNARFWORD;
	if (strEQ(d,"bind"))
	    FOP2(O_BIND);
	if (strEQ(d,"binmode"))
	    FOP(O_BINMODE);
	break;
    case 'c': case 'C':
	SNARFWORD;
	if (strEQ(d,"chop"))
	    LFUN(O_CHOP);
	if (strEQ(d,"continue"))
	    OPERATOR(CONTINUE);
	if (strEQ(d,"chdir")) {
	    (void)stabent("ENV",TRUE);	/* may use HOME */
	    UNI(O_CHDIR);
	}
	if (strEQ(d,"close"))
	    FOP(O_CLOSE);
	if (strEQ(d,"closedir"))
	    FOP(O_CLOSEDIR);
	if (strEQ(d,"cmp"))
	    EOP(O_SCMP);
	if (strEQ(d,"caller"))
	    UNI(O_CALLER);
	if (strEQ(d,"crypt")) {
#ifdef FCRYPT
	    static int cryptseen = 0;

	    if (!cryptseen++)
		init_des();
#endif
	    FUN2(O_CRYPT);
	}
	if (strEQ(d,"chmod"))
	    LOP(O_CHMOD);
	if (strEQ(d,"chown"))
	    LOP(O_CHOWN);
	if (strEQ(d,"connect"))
	    FOP2(O_CONNECT);
	if (strEQ(d,"cos"))
	    UNI(O_COS);
	if (strEQ(d,"chroot"))
	    UNI(O_CHROOT);
	break;
    case 'd': case 'D':
	SNARFWORD;
	if (strEQ(d,"do")) {
	    d = bufend;
	    while (s < d && isSPACE(*s))
		s++;
	    if (isALPHA(*s) || *s == '_')
		*(--s) = '\\';	/* force next ident to WORD */
	    OPERATOR(DO);
	}
	if (strEQ(d,"die"))
	    LOP(O_DIE);
	if (strEQ(d,"defined"))
	    LFUN(O_DEFINED);
	if (strEQ(d,"delete"))
	    OPERATOR(DELETE);
	if (strEQ(d,"dbmopen"))
	    HFUN3(O_DBMOPEN);
	if (strEQ(d,"dbmclose"))
	    HFUN(O_DBMCLOSE);
	if (strEQ(d,"dump"))
	    LOOPX(O_DUMP);
	break;
    case 'e': case 'E':
	SNARFWORD;
	if (strEQ(d,"else"))
	    OPERATOR(ELSE);
	if (strEQ(d,"elsif")) {
	    yylval.ival = curcmd->c_line;
	    OPERATOR(ELSIF);
	}
	if (strEQ(d,"eq") || strEQ(d,"EQ"))
	    EOP(O_SEQ);
	if (strEQ(d,"exit"))
	    UNI(O_EXIT);
	if (strEQ(d,"eval")) {
	    allstabs = TRUE;		/* must initialize everything since */
	    UNI(O_EVAL);		/* we don't know what will be used */
	}
	if (strEQ(d,"eof"))
	    FOP(O_EOF);
	if (strEQ(d,"exp"))
	    UNI(O_EXP);
	if (strEQ(d,"each"))
	    HFUN(O_EACH);
	if (strEQ(d,"exec")) {
	    set_csh();
	    LOP(O_EXEC_OP);
	}
	if (strEQ(d,"endhostent"))
	    FUN0(O_EHOSTENT);
	if (strEQ(d,"endnetent"))
	    FUN0(O_ENETENT);
	if (strEQ(d,"endservent"))
	    FUN0(O_ESERVENT);
	if (strEQ(d,"endprotoent"))
	    FUN0(O_EPROTOENT);
	if (strEQ(d,"endpwent"))
	    FUN0(O_EPWENT);
	if (strEQ(d,"endgrent"))
	    FUN0(O_EGRENT);
	break;
    case 'f': case 'F':
	SNARFWORD;
	if (strEQ(d,"for") || strEQ(d,"foreach")) {
	    yylval.ival = curcmd->c_line;
	    while (s < bufend && isSPACE(*s))
		s++;
	    if (isALPHA(*s))
		fatal("Missing $ on loop variable");
	    OPERATOR(FOR);
	}
	if (strEQ(d,"format")) {
	    d = bufend;
	    while (s < d && isSPACE(*s))
		s++;
	    if (isALPHA(*s) || *s == '_')
		*(--s) = '\\';	/* force next ident to WORD */
	    in_format = TRUE;
	    allstabs = TRUE;		/* must initialize everything since */
	    OPERATOR(FORMAT);		/* we don't know what will be used */
	}
	if (strEQ(d,"fork"))
	    FUN0(O_FORK);
	if (strEQ(d,"fcntl"))
	    FOP3(O_FCNTL);
	if (strEQ(d,"fileno"))
	    FOP(O_FILENO);
	if (strEQ(d,"flock"))
	    FOP2(O_FLOCK);
	break;
    case 'g': case 'G':
	SNARFWORD;
	if (strEQ(d,"gt") || strEQ(d,"GT"))
	    ROP(O_SGT);
	if (strEQ(d,"ge") || strEQ(d,"GE"))
	    ROP(O_SGE);
	if (strEQ(d,"grep"))
	    FL2(O_GREP);
	if (strEQ(d,"goto"))
	    LOOPX(O_GOTO);
	if (strEQ(d,"gmtime"))
	    UNI(O_GMTIME);
	if (strEQ(d,"getc"))
	    FOP(O_GETC);
	if (strnEQ(d,"get",3)) {
	    d += 3;
	    if (*d == 'p') {
		if (strEQ(d,"ppid"))
		    FUN0(O_GETPPID);
		if (strEQ(d,"pgrp"))
		    UNI(O_GETPGRP);
		if (strEQ(d,"priority"))
		    FUN2(O_GETPRIORITY);
		if (strEQ(d,"protobyname"))
		    UNI(O_GPBYNAME);
		if (strEQ(d,"protobynumber"))
		    FUN1(O_GPBYNUMBER);
		if (strEQ(d,"protoent"))
		    FUN0(O_GPROTOENT);
		if (strEQ(d,"pwent"))
		    FUN0(O_GPWENT);
		if (strEQ(d,"pwnam"))
		    FUN1(O_GPWNAM);
		if (strEQ(d,"pwuid"))
		    FUN1(O_GPWUID);
		if (strEQ(d,"peername"))
		    FOP(O_GETPEERNAME);
	    }
	    else if (*d == 'h') {
		if (strEQ(d,"hostbyname"))
		    UNI(O_GHBYNAME);
		if (strEQ(d,"hostbyaddr"))
		    FUN2(O_GHBYADDR);
		if (strEQ(d,"hostent"))
		    FUN0(O_GHOSTENT);
	    }
	    else if (*d == 'n') {
		if (strEQ(d,"netbyname"))
		    UNI(O_GNBYNAME);
		if (strEQ(d,"netbyaddr"))
		    FUN2(O_GNBYADDR);
		if (strEQ(d,"netent"))
		    FUN0(O_GNETENT);
	    }
	    else if (*d == 's') {
		if (strEQ(d,"servbyname"))
		    FUN2(O_GSBYNAME);
		if (strEQ(d,"servbyport"))
		    FUN2(O_GSBYPORT);
		if (strEQ(d,"servent"))
		    FUN0(O_GSERVENT);
		if (strEQ(d,"sockname"))
		    FOP(O_GETSOCKNAME);
		if (strEQ(d,"sockopt"))
		    FOP3(O_GSOCKOPT);
	    }
	    else if (*d == 'g') {
		if (strEQ(d,"grent"))
		    FUN0(O_GGRENT);
		if (strEQ(d,"grnam"))
		    FUN1(O_GGRNAM);
		if (strEQ(d,"grgid"))
		    FUN1(O_GGRGID);
	    }
	    else if (*d == 'l') {
		if (strEQ(d,"login"))
		    FUN0(O_GETLOGIN);
	    }
	    d -= 3;
	}
	break;
    case 'h': case 'H':
	SNARFWORD;
	if (strEQ(d,"hex"))
	    UNI(O_HEX);
	break;
    case 'i': case 'I':
	SNARFWORD;
	if (strEQ(d,"if")) {
	    yylval.ival = curcmd->c_line;
	    OPERATOR(IF);
	}
	if (strEQ(d,"index"))
	    FUN2x(O_INDEX);
	if (strEQ(d,"int"))
	    UNI(O_INT);
	if (strEQ(d,"ioctl"))
	    FOP3(O_IOCTL);
	break;
    case 'j': case 'J':
	SNARFWORD;
	if (strEQ(d,"join"))
	    FL2(O_JOIN);
	break;
    case 'k': case 'K':
	SNARFWORD;
	if (strEQ(d,"keys"))
	    HFUN(O_KEYS);
	if (strEQ(d,"kill"))
	    LOP(O_KILL);
	break;
    case 'l': case 'L':
	SNARFWORD;
	if (strEQ(d,"last"))
	    LOOPX(O_LAST);
	if (strEQ(d,"local"))
	    OPERATOR(LOCAL);
	if (strEQ(d,"length"))
	    UNI(O_LENGTH);
	if (strEQ(d,"lt") || strEQ(d,"LT"))
	    ROP(O_SLT);
	if (strEQ(d,"le") || strEQ(d,"LE"))
	    ROP(O_SLE);
	if (strEQ(d,"localtime"))
	    UNI(O_LOCALTIME);
	if (strEQ(d,"log"))
	    UNI(O_LOG);
	if (strEQ(d,"link"))
	    FUN2(O_LINK);
	if (strEQ(d,"listen"))
	    FOP2(O_LISTEN);
	if (strEQ(d,"lstat"))
	    FOP(O_LSTAT);
	break;
    case 'm': case 'M':
	if (s[1] == '\'') {
	    d = "m";
	    s++;
	}
	else {
	    SNARFWORD;
	}
	if (strEQ(d,"m")) {
	    s = scanpat(s-1);
	    if (yylval.arg)
		TERM(PATTERN);
	    else
		RETURN(1);	/* force error */
	}
	switch (d[1]) {
	case 'k':
	    if (strEQ(d,"mkdir"))
		FUN2(O_MKDIR);
	    break;
	case 's':
	    if (strEQ(d,"msgctl"))
		FUN3(O_MSGCTL);
	    if (strEQ(d,"msgget"))
		FUN2(O_MSGGET);
	    if (strEQ(d,"msgrcv"))
		FUN5(O_MSGRCV);
	    if (strEQ(d,"msgsnd"))
		FUN3(O_MSGSND);
	    break;
	}
	break;
    case 'n': case 'N':
	SNARFWORD;
	if (strEQ(d,"next"))
	    LOOPX(O_NEXT);
	if (strEQ(d,"ne") || strEQ(d,"NE"))
	    EOP(O_SNE);
	break;
    case 'o': case 'O':
	SNARFWORD;
	if (strEQ(d,"open"))
	    OPERATOR(OPEN);
	if (strEQ(d,"ord"))
	    UNI(O_ORD);
	if (strEQ(d,"oct"))
	    UNI(O_OCT);
	if (strEQ(d,"opendir"))
	    FOP2(O_OPEN_DIR);
	break;
    case 'p': case 'P':
	SNARFWORD;
	if (strEQ(d,"print")) {
	    checkcomma(s,d,"filehandle");
	    LOP(O_PRINT);
	}
	if (strEQ(d,"printf")) {
	    checkcomma(s,d,"filehandle");
	    LOP(O_PRTF);
	}