lex.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

				continue;
			if(cp == symb+2)
				yyerror("malformed hex constant");
			goto ncu;
		}
	if(c < '0' || c > '7')
		goto dc;
	for(;;) {
		if(c >= '0' && c <= '7') {
			*cp++ = c;
			c = GETC();
			continue;
		}
		goto ncu;
	}

dc:
	if(c == '.')
		goto casedot;
	if(c == 'e' || c == 'E')
		goto casee;

ncu:
	if((c == 'U' || c == 'u') && !(c1 & Numuns)) {
		c = GETC();
		c1 |= Numuns;
		goto ncu;
	}
	if((c == 'L' || c == 'l') && !(c1 & Numvlong)) {
		c = GETC();
		if(c1 & Numlong)
			c1 |= Numvlong;
		c1 |= Numlong;
		goto ncu;
	}
	*cp = 0;
	peekc = c;
	if(mpatov(symb, &yylval.vval))
		yyerror("overflow in constant");

	vv = yylval.vval;
	if(c1 & Numvlong) {
		if((c1 & Numuns) || convvtox(vv, TVLONG) < 0) {
			c = LUVLCONST;
			t = TUVLONG;
			goto nret;
		}
		c = LVLCONST;
		t = TVLONG;
		goto nret;
	}
	if(c1 & Numlong) {
		if((c1 & Numuns) || convvtox(vv, TLONG) < 0) {
			c = LULCONST;
			t = TULONG;
			goto nret;
		}
		c = LLCONST;
		t = TLONG;
		goto nret;
	}
	if((c1 & Numuns) || convvtox(vv, TINT) < 0) {
		c = LUCONST;
		t = TUINT;
		goto nret;
	}
	c = LCONST;
	t = TINT;
	goto nret;

nret:
	yylval.vval = convvtox(vv, t);
	if(yylval.vval != vv){
		nearln = lineno;
		warn(Z, "truncated constant: %T %s", types[t], symb);
	}
	return c;

casedot:
	for(;;) {
		*cp++ = c;
		c = GETC();
		if(!isdigit(c))
			break;
	}
	if(c != 'e' && c != 'E')
		goto caseout;

casee:
	*cp++ = 'e';
	c = GETC();
	if(c == '+' || c == '-') {
		*cp++ = c;
		c = GETC();
	}
	if(!isdigit(c))
		yyerror("malformed fp constant exponent");
	while(isdigit(c)) {
		*cp++ = c;
		c = GETC();
	}

caseout:
	if(c == 'L' || c == 'l') {
		c = GETC();
		c1 |= Numlong;
	} else
	if(c == 'F' || c == 'f') {
		c = GETC();
		c1 |= Numflt;
	}
	*cp = 0;
	peekc = c;
	yylval.dval = strtod(symb, nil);
	if(isInf(yylval.dval, 1) || isInf(yylval.dval, -1)) {
		yyerror("overflow in float constant");
		yylval.dval = 0;
	}
	if(c1 & Numflt)
		return LFCONST;
	return LDCONST;
}

/*
 * convert a string, s, to vlong in *v
 * return conversion overflow.
 * required syntax is [0[x]]d*
 */
int
mpatov(char *s, vlong *v)
{
	vlong n, nn;
	int c;

	n = 0;
	c = *s;
	if(c == '0')
		goto oct;
	while(c = *s++) {
		if(c >= '0' && c <= '9')
			nn = n*10 + c-'0';
		else
			goto bad;
		if(n < 0 && nn >= 0)
			goto bad;
		n = nn;
	}
	goto out;

oct:
	s++;
	c = *s;
	if(c == 'x' || c == 'X')
		goto hex;
	while(c = *s++) {
		if(c >= '0' || c <= '7')
			nn = n*8 + c-'0';
		else
			goto bad;
		if(n < 0 && nn >= 0)
			goto bad;
		n = nn;
	}
	goto out;

hex:
	s++;
	while(c = *s++) {
		if(c >= '0' && c <= '9')
			c += 0-'0';
		else
		if(c >= 'a' && c <= 'f')
			c += 10-'a';
		else
		if(c >= 'A' && c <= 'F')
			c += 10-'A';
		else
			goto bad;
		nn = n*16 + c;
		if(n < 0 && nn >= 0)
			goto bad;
		n = nn;
	}
out:
	*v = n;
	return 0;

bad:
	*v = ~0;
	return 1;
}

int
getc(void)
{
	int c;

	if(peekc != IGN) {
		c = peekc;
		peekc = IGN;
	} else
		c = GETC();
	if(c == '\n')
		lineno++;
	if(c == EOF) {
		yyerror("End of file");
		errorexit();
	}
	return c;
}

long
getr(void)
{
	int c, i;
	char str[UTFmax+1];
	Rune rune;


	c = getc();
	if(c < Runeself)
		return c;
	i = 0;
	str[i++] = c;

loop:
	c = getc();
	str[i++] = c;
	if(!fullrune(str, i))
		goto loop;
	c = chartorune(&rune, str);
	if(rune == Runeerror && c == 1) {
		nearln = lineno;
		diag(Z, "illegal rune in string");
		for(c=0; c<i; c++)
			print(" %.2x", *(uchar*)(str+c));
		print("\n");
	}
	return rune;
}

int
getnsc(void)
{
	int c;

	if(peekc != IGN) {
		c = peekc;
		peekc = IGN;
	} else
		c = GETC();
	for(;;) {
		if(!isspace(c))
			return c;
		if(c == '\n') {
			lineno++;
			return c;
		}
		c = GETC();
	}
}

void
unget(int c)
{

	peekc = c;
	if(c == '\n')
		lineno--;
}

long
escchar(long e, int longflg, int escflg)
{
	long c, l;
	int i;

loop:
	c = getr();
	if(c == '\n') {
		yyerror("newline in string");
		return EOF;
	}
	if(c != '\\') {
		if(c == e)
			c = EOF;
		return c;
	}
	c = getr();
	if(c == 'x') {
		/*
		 * note this is not ansi,
		 * supposed to only accept 2 hex
		 */
		i = 2;
		if(longflg)
			i = 4;
		l = 0;
		for(; i>0; i--) {
			c = getc();
			if(c >= '0' && c <= '9') {
				l = l*16 + c-'0';
				continue;
			}
			if(c >= 'a' && c <= 'f') {
				l = l*16 + c-'a' + 10;
				continue;
			}
			if(c >= 'A' && c <= 'F') {
				l = l*16 + c-'A' + 10;
				continue;
			}
			unget(c);
			break;
		}
		if(escflg)
			l |= ESC;
		return l;
	}
	if(c >= '0' && c <= '7') {
		/*
		 * note this is not ansi,
		 * supposed to only accept 3 oct
		 */
		i = 2;
		if(longflg)
			i = 5;
		l = c - '0';
		for(; i>0; i--) {
			c = getc();
			if(c >= '0' && c <= '7') {
				l = l*8 + c-'0';
				continue;
			}
			unget(c);
		}
		if(escflg)
			l |= ESC;
		return l;
	}
	switch(c)
	{
	case '\n':	goto loop;
	case 'n':	return '\n';
	case 't':	return '\t';
	case 'b':	return '\b';
	case 'r':	return '\r';
	case 'f':	return '\f';
	case 'a':	return '\a';
	case 'v':	return '\v';
	}
	return c;
}

struct
{
	char	*name;
	ushort	lexical;
	ushort	type;
} itab[] =
{
	"auto",		LAUTO,		0,
	"break",	LBREAK,		0,
	"case",		LCASE,		0,
	"char",		LCHAR,		TCHAR,
	"const",	LCONSTNT,	0,
	"continue",	LCONTINUE,	0,
	"default",	LDEFAULT,	0,
	"do",		LDO,		0,
	"double",	LDOUBLE,	TDOUBLE,
	"else",		LELSE,		0,
	"enum",		LENUM,		0,
	"extern",	LEXTERN,	0,
	"float",	LFLOAT,		TFLOAT,
	"for",		LFOR,		0,
	"goto",		LGOTO,		0,
	"if",		LIF,		0,
	"inline",	LINLINE,	0,
	"int",		LINT,		TINT,
	"long",		LLONG,		TLONG,
	"register",	LREGISTER,	0,
	"restrict",	LRESTRICT,	0,
	"return",	LRETURN,	0,
	"SET",		LSET,		0,
	"short",	LSHORT,		TSHORT,
	"signed",	LSIGNED,	0,
	"signof",	LSIGNOF,	0,
	"sizeof",	LSIZEOF,	0,
	"static",	LSTATIC,	0,
	"struct",	LSTRUCT,	0,
	"switch",	LSWITCH,	0,
	"typedef",	LTYPEDEF,	0,
	"typestr",	LTYPESTR,	0,
	"union",	LUNION,		0,
	"unsigned",	LUNSIGNED,	0,
	"USED",		LUSED,		0,
	"void",		LVOID,		TVOID,
	"volatile",	LVOLATILE,	0,
	"while",	LWHILE,		0,
	0
};

void
cinit(void)
{
	Sym *s;
	int i;
	Type *t;

	nerrors = 0;
	lineno = 1;
	iostack = I;
	iofree = I;
	peekc = IGN;
	nhunk = 0;

	types[TXXX] = T;
	types[TCHAR] = typ(TCHAR, T);
	types[TUCHAR] = typ(TUCHAR, T);
	types[TSHORT] = typ(TSHORT, T);
	types[TUSHORT] = typ(TUSHORT, T);
	types[TINT] = typ(TINT, T);
	types[TUINT] = typ(TUINT, T);
	types[TLONG] = typ(TLONG, T);
	types[TULONG] = typ(TULONG, T);
	types[TVLONG] = typ(TVLONG, T);
	types[TUVLONG] = typ(TUVLONG, T);
	types[TFLOAT] = typ(TFLOAT, T);
	types[TDOUBLE] = typ(TDOUBLE, T);
	types[TVOID] = typ(TVOID, T);
	types[TENUM] = typ(TENUM, T);
	types[TFUNC] = typ(TFUNC, types[TINT]);
	types[TIND] = typ(TIND, types[TVOID]);

	for(i=0; i<NHASH; i++)
		hash[i] = S;
	for(i=0; itab[i].name; i++) {
		s = slookup(itab[i].name);
		s->lexical = itab[i].lexical;
		if(itab[i].type != 0)
			s->type = types[itab[i].type];
	}
	blockno = 0;
	autobn = 0;
	autoffset = 0;

	t = typ(TARRAY, types[TCHAR]);
	t->width = 0;
	symstring = slookup(".string");
	symstring->class = CSTATIC;
	symstring->type = t;

	t = typ(TARRAY, types[TCHAR]);
	t->width = 0;

	nodproto = new(OPROTO, Z, Z);
	dclstack = D;

	pathname = allocn(pathname, 0, 100);
	if(mygetwd(pathname, 99) == 0) {
		pathname = allocn(pathname, 100, 900);
		if(mygetwd(pathname, 999) == 0)
			strcpy(pathname, "/???");
	}

	fmtinstall('O', Oconv);
	fmtinstall('T', Tconv);
	fmtinstall('F', FNconv);
	fmtinstall('L', Lconv);
	fmtinstall('Q', Qconv);
	fmtinstall('|', VBconv);
}

int
filbuf(void)
{
	Io *i;

loop:
	i = iostack;
	if(i == I)
		return EOF;
	if(i->f < 0)
		goto pop;
	fi.c = read(i->f, i->b, BUFSIZ) - 1;
	if(fi.c < 0) {
		close(i->f);
		linehist(0, 0);
		goto pop;
	}
	fi.p = i->b + 1;
	return i->b[0] & 0xff;

pop:
	iostack = i->link;
	i->link = iofree;
	iofree = i;
	i = iostack;
	if(i == I)
		return EOF;
	fi.p = i->p;
	fi.c = i->c;
	if(--fi.c < 0)
		goto loop;
	return *fi.p++ & 0xff;
}

int
Oconv(Fmt *fp)
{
	int a;

	a = va_arg(fp->args, int);
	if(a < OXXX || a > OEND)
		return fmtprint(fp, "***badO %d***", a);

	return fmtstrcpy(fp, onames[a]);
}

int
Lconv(Fmt *fp)
{
	char str[STRINGSZ], s[STRINGSZ];
	Hist *h;
	struct
	{
		Hist*	incl;	/* start of this include file */
		long	idel;	/* delta line number to apply to include */
		Hist*	line;	/* start of this #line directive */
		long	ldel;	/* delta line number to apply to #line */
	} a[HISTSZ];
	long l, d;
	int i, n;

	l = va_arg(fp->args, long);
	n = 0;
	for(h = hist; h != H; h = h->link) {
		if(l < h->line)
			break;
		if(h->name) {
			if(h->offset != 0) {		/* #line directive, not #pragma */
				if(n > 0 && n < HISTSZ && h->offset >= 0) {
					a[n-1].line = h;
					a[n-1].ldel = h->line - h->offset + 1;
				}
			} else {
				if(n < HISTSZ) {	/* beginning of file */
					a[n].incl = h;
					a[n].idel = h->line;
					a[n].line = 0;
				}
				n++;
			}
			continue;
		}
		n--;
		if(n > 0 && n < HISTSZ) {
			d = h->line - a[n].incl->line;
			a[n-1].ldel += d;
			a[n-1].idel += d;
		}
	}
	if(n > HISTSZ)
		n = HISTSZ;
	str[0] = 0;
	for(i=n-1; i>=0; i--) {
		if(i != n-1) {
			if(fp->flags & ~(FmtWidth|FmtPrec))	/* BUG ROB - was f3 */
				break;
			strcat(str, " ");
		}
		if(a[i].line)
			snprint(s, STRINGSZ, "%s:%ld[%s:%ld]",
				a[i].line->name, l-a[i].ldel+1,
				a[i].incl->name, l-a[i].idel+1);
		else
			snprint(s, STRINGSZ, "%s:%ld",
				a[i].incl->name, l-a[i].idel+1);
		if(strlen(s)+strlen(str) >= STRINGSZ-10)
			break;
		strcat(str, s);
		l = a[i].incl->line - 1;	/* now print out start of this file */
	}
	if(n == 0)
		strcat(str, "<eof>");
	return fmtstrcpy(fp, str);
}

int
Tconv(Fmt *fp)
{
	char str[STRINGSZ+20], s[STRINGSZ+20];
	Type *t, *t1;
	int et;
	long n;

	str[0] = 0;
	for(t = va_arg(fp->args, Type*); t != T; t = t->link) {
		et = t->etype;
		if(str[0])
			strcat(str, " ");
		if(t->garb&~GINCOMPLETE) {
			sprint(s, "%s ", gnames[t->garb&~GINCOMPLETE]);
			if(strlen(str) + strlen(s) < STRINGSZ)
				strcat(str, s);
		}
		sprint(s, "%s", tnames[et]);
		if(strlen(str) + strlen(s) < STRINGSZ)
			strcat(str, s);
		if(et == TFUNC && (t1 = t->down)) {
			sprint(s, "(%T", t1);
			if(strlen(str) + strlen(s) < STRINGSZ)
				strcat(str, s);
			while(t1 = t1->down) {
				sprint(s, ", %T", t1);
				if(strlen(str) + strlen(s) < STRINGSZ)
					strcat(str, s);
			}
			if(strlen(str) + strlen(s) < STRINGSZ)
				strcat(str, ")");
		}
		if(et == TARRAY) {
			n = t->width;
			if(t->link && t->link->width)
				n /= t->link->width;
			sprint(s, "[%ld]", n);
			if(strlen(str) + strlen(s) < STRINGSZ)
				strcat(str, s);
		}
		if(t->nbits) {
			sprint(s, " %d:%d", t->shift, t->nbits);
			if(strlen(str) + strlen(s) < STRINGSZ)
				strcat(str, s);
		}
		if(typesu[et]) {
			if(t->tag) {
				strcat(str, " ");
				if(strlen(str) + strlen(t->tag->name) < STRINGSZ)
					strcat(str, t->tag->name);
			} else
				strcat(str, " {}");
			break;
		}
	}
	return fmtstrcpy(fp, str);
}

int
FNconv(Fmt *fp)
{
	char *str;
	Node *n;

	n = va_arg(fp->args, Node*);
	str = "<indirect>";
	if(n != Z && (n->op == ONAME || n->op == ODOT || n->op == OELEM))
		str = n->sym->name;
	return fmtstrcpy(fp, str);
}

int
Qconv(Fmt *fp)
{
	char str[STRINGSZ+20], *s;
	long b;
	int i;

	str[0] = 0;
	for(b = va_arg(fp->args, long); b;) {
		i = bitno(b);
		if(str[0])
			strcat(str, " ");
		s = qnames[i];
		if(strlen(str) + strlen(s) >= STRINGSZ)
			break;
		strcat(str, s);
		b &= ~(1L << i);
	}
	return fmtstrcpy(fp, str);
}

int
VBconv(Fmt *fp)
{
	char str[STRINGSZ];
	int i, n, t, pc;

	n = va_arg(fp->args, int);
	pc = 0;	/* BUG: was printcol */
	i = 0;
	while(pc < n) {
		t = (pc+4) & ~3;
		if(t <= n) {
			str[i++] = '\t';
			pc = t;
			continue;
		}
		str[i++] = ' ';
		pc++;
	}
	str[i] = 0;

	return fmtstrcpy(fp, str);
}

/*
 * real allocs
 */
void*
alloc(long n)
{
	void *p;

	while((uintptr)hunk & MAXALIGN) {
		hunk++;
		nhunk--;
	}
	while(nhunk < n)
		gethunk();
	p = hunk;
	nhunk -= n;
	hunk += n;
	return p;
}

void*
allocn(void *p, long on, long n)
{
	void *q;

	q = (uchar*)p + on;
	if(q != hunk || nhunk < n) {
		while(nhunk < on+n)
			gethunk();
		memmove(hunk, p, on);
		p = hunk;
		hunk += on;
		nhunk -= on;
	}
	hunk += n;
	nhunk -= n;
	return p;
}

void
setinclude(char *p)
{
	int i;
	char *e;

	while(*p != 0) {
		e = strchr(p, ' ');
		if(e != 0)
			*e = '\0';

		for(i=1; i < ninclude; i++)
			if(strcmp(p, include[i]) == 0)
				break;

		if(i >= ninclude)
			include[ninclude++] = p;

		if(ninclude > nelem(include)) {
			diag(Z, "ninclude too small %d", nelem(include));
			exits("ninclude");
		}

		if(e == 0)
			break;
		p = e+1;
	}
}