4.4bsd-patches

早期freebsd实现

From: Vern Paxson <[e-mail]>

I've ported Research Awk to flex and tested it moderately.  Note that I
didn't have time to support the nice error-message context stuff (where
it points out exactly where on a line it thinks an error occurred), as
the original code made a lot of assumptions regarding the internal
buffering of a lex scanner that are no longer valid with flex.  Also, the
sources had a function called "isnumber" which conflicted with a macro by
the same name in <ctype.h>, so I changed its name to is_a_number.

Let me know if you find more problems.

		Vern

diff -wc orig/awk.lx.l awk.lx.l
*** orig/awk.lx.l	Fri Oct 29 10:39:54 1993
--- awk.lx.l	Fri Oct 29 11:07:35 1993
***************
*** 25,37 ****
  THIS SOFTWARE.
  ****************************************************************/
  
- /* some of this depends on behavior of lex that
-    may not be preserved in other implementations of lex.
- */
- 
- #undef	input	/* defeat lex */
- #undef	unput
- 
  #include <stdlib.h>
  #include <string.h>
  #include "awk.h"
--- 25,30 ----
***************
*** 60,65 ****
--- 53,70 ----
  uchar	cbuf[CBUFLEN];
  uchar	*s;
  int	clen, cflag;
+ 
+ /* some of this depends on behavior of lex that
+    may not be preserved in other implementations of lex.
+ */
+ 
+ static	int	my_input( YY_CHAR *buf, int max_size );
+ 
+ #undef YY_INPUT
+ #define YY_INPUT(buf,result,max_size) result = my_input(buf, max_size);
+ 
+ #undef YY_USER_INIT
+ #define YY_USER_INIT init_input_source();
  %}
  
  A	[a-zA-Z_]
***************
*** 70,76 ****
  WS	[ \t]
  
  %%
! 	switch (yybgin-yysvec-1) {	/* witchcraft */
  	case 0:
  		BEGIN A;
  		break;
--- 75,81 ----
  WS	[ \t]
  
  %%
! 	switch ((yy_start - 1) / 2) {	/* witchcraft */
  	case 0:
  		BEGIN A;
  		break;
***************
*** 116,128 ****
  
  <A>"$"{D}+	{ yylval.cp = fieldadr(atoi(yytext+1)); RET(FIELD); }
  <A>"$NF"	{ unputstr("(NF)"); return(INDIRECT); }
! <A>"$"{A}{B}*	{ int c, n;
! 		  c = input(); unput(c);
! 		  if (c == '(' || c == '[' || infunc && (n=isarg(yytext+1)) >= 0) {
! 			unputstr(yytext+1);
  			return(INDIRECT);
  		  } else {
! 			yylval.cp = setsymtab(yytext+1,"",0.0,STR|NUM,symtab);
  			RET(IVAR);
  		  }
  		}
--- 121,139 ----
  
  <A>"$"{D}+	{ yylval.cp = fieldadr(atoi(yytext+1)); RET(FIELD); }
  <A>"$NF"	{ unputstr("(NF)"); return(INDIRECT); }
! <A>"$"{A}{B}*	{
! 		  int c;
! 		  char *yytext_copy = strdup(yytext);
! 		  c = input(); unput(c);	/* look for '(' or '[' */
! 		  if (c == '(' || c == '[' ||
! 		      infunc && isarg(yytext_copy+1) >= 0) {
! 			  unputstr(yytext_copy+1);
! 			  free(yytext_copy);
  			  return(INDIRECT);
  		  } else {
! 			  yylval.cp =
! 				setsymtab(yytext_copy+1,"",0.0,STR|NUM,symtab);
! 			  free(yytext_copy);
  			  RET(IVAR);
  		  }
  		}
***************
*** 172,183 ****
  <A>fflush	{ yylval.i = FFLUSH; RET(BLTIN); }
  
  <A>{A}{B}*	{ int n, c;
  		  c = input(); unput(c);	/* look for '(' */
! 		  if (c != '(' && infunc && (n=isarg(yytext)) >= 0) {
  			yylval.i = n;
  			RET(ARG);
  		  } else {
! 			yylval.cp = setsymtab(yytext,"",0.0,STR|NUM,symtab);
  			if (c == '(') {
  				RET(CALL);
  			} else {
--- 183,198 ----
  <A>fflush	{ yylval.i = FFLUSH; RET(BLTIN); }
  
  <A>{A}{B}*	{ int n, c;
+ 		  char *yytext_copy = strdup(yytext);
  		  c = input(); unput(c);	/* look for '(' */
! 		  if (c != '(' && infunc && (n=isarg(yytext_copy)) >= 0) {
  			yylval.i = n;
+ 			free(yytext_copy);
  			RET(ARG);
  		  } else {
! 			yylval.cp =
! 				setsymtab(yytext_copy,"",0.0,STR|NUM,symtab);
! 			free(yytext_copy);
  			if (c == '(') {
  				RET(CALL);
  			} else {
***************
*** 185,190 ****
--- 200,206 ----
  			}
  		  }
  		}
+ 
  <A>\"		{ BEGIN str; clen = 0; }
  
  <A>"}"		{ if (--bracecnt < 0) ERROR "extra }" SYNTAX; BEGIN sc; RET(';'); }
***************
*** 235,280 ****
  	clen = 0;
  }
  
! /* input() and unput() are transcriptions of the standard lex
!    macros for input and output with additions for error message
!    printing.  God help us all if someone changes how lex works.
! */
! 
! uchar	ebuf[300];
! uchar	*ep = ebuf;
! 
! input(void)	/* get next lexical input character */
  {
- 	register int c;
  	extern uchar *lexprog;
  
! 	if (yysptr > yysbuf)
! 		c = U(*--yysptr);
! 	else if (lexprog != NULL) {	/* awk '...' */
! 		if (c = *lexprog)
! 			lexprog++;
! 	} else				/* awk -f ... */
! 		c = pgetc();
! 	if (c == '\n')
! 		yylineno++;
! 	else if (c == EOF)
! 		c = 0;
! 	if (ep >= ebuf + sizeof ebuf)
! 		ep = ebuf;
! 	return *ep++ = c;
  }
  
! void unput(int c)	/* put lexical character back on input */
! {
! 	yytchar = c;
! 	if (yytchar == '\n')
! 		yylineno--;
! 	*yysptr++ = yytchar;
! 	if (--ep < ebuf)
! 		ep = ebuf + sizeof(ebuf) - 1;
  }
  
- 
  void unputstr(char *s)	/* put a string back on input */
  {
  	int i;
--- 251,281 ----
  	clen = 0;
  }
  
! static int my_input( YY_CHAR *buf, int max_size )
  {
  	extern uchar *lexprog;
  
! 	if ( lexprog ) {		/* awk '...' */
! 		int num_chars = strlen( lexprog );
! 		if ( num_chars > max_size )
! 			{
! 			num_chars = max_size;
! 			strncpy( buf, lexprog, num_chars );
  			}
+ 		else
+ 			strcpy( buf, lexprog );
+ 		lexprog += num_chars;
+ 		return num_chars;
  
! 	} else {			/* awk -f ... */
! 		int c = pgetc();
! 		if (c == EOF)
! 			return 0;
! 		buf[0] = c;
! 		return 1;
  	}
+ }
  
  void unputstr(char *s)	/* put a string back on input */
  {
  	int i;
***************
*** 281,284 ****
--- 282,290 ----
  
  	for (i = strlen(s)-1; i >= 0; i--)
  		unput(s[i]);
+ }
+ 
+ int lex_input()
+ {
+ 	return input();
  }
diff -wc orig/lib.c lib.c
*** orig/lib.c	Fri Oct 29 10:39:55 1993
--- lib.c	Fri Oct 29 11:22:33 1993
***************
*** 129,135 ****
  					xfree(recloc->sval);
  				recloc->sval = record;
  				recloc->tval = REC | STR | DONTFREE;
! 				if (isnumber(recloc->sval)) {
  					recloc->fval = atof(recloc->sval);
  					recloc->tval |= NUM;
  				}
--- 129,135 ----
  					xfree(recloc->sval);
  				recloc->sval = record;
  				recloc->tval = REC | STR | DONTFREE;
! 				if (is_a_number(recloc->sval)) {
  					recloc->fval = atof(recloc->sval);
  					recloc->tval |= NUM;
  				}
***************
*** 202,208 ****
  	p = qstring(p, '\0');
  	q = setsymtab(s, p, 0.0, STR, symtab);
  	setsval(q, p);
! 	if (isnumber(q->sval)) {
  		q->fval = atof(q->sval);
  		q->tval |= NUM;
  	}
--- 202,208 ----
  	p = qstring(p, '\0');
  	q = setsymtab(s, p, 0.0, STR, symtab);
  	setsval(q, p);
! 	if (is_a_number(q->sval)) {
  		q->fval = atof(q->sval);
  		q->tval |= NUM;
  	}
***************
*** 268,274 ****
  	maxfld = i;
  	donefld = 1;
  	for (p = fldtab+1; p <= fldtab+maxfld; p++) {
! 		if(isnumber(p->sval)) {
  			p->fval = atof(p->sval);
  			p->tval |= NUM;
  		}
--- 268,274 ----
  	maxfld = i;
  	donefld = 1;
  	for (p = fldtab+1; p <= fldtab+maxfld; p++) {
! 		if(is_a_number(p->sval)) {
  			p->fval = atof(p->sval);
  			p->tval |= NUM;
  		}
***************
*** 410,416 ****
  
  	if (beenhere++)
  		return;
! 	while ((c = input()) != EOF && c != '\0')
  		bclass(c);
  	bcheck2(bracecnt, '{', '}');
  	bcheck2(brackcnt, '[', ']');
--- 410,416 ----
  
  	if (beenhere++)
  		return;
! 	while ((c = lex_input()) != EOF && c != '\0')
  		bclass(c);
  	bcheck2(bracecnt, '{', '}');
  	bcheck2(brackcnt, '[', ']');
***************
*** 458,463 ****
--- 458,464 ----
  
  void eprint(void)	/* try to print context around error */
  {
+ #if 0
  	uchar *p, *q;
  	int c;
  	static int been_here = 0;
***************
*** 490,495 ****
--- 491,497 ----
  		}
  	putc('\n', stderr);
  	ep = ebuf;
+ #endif
  }
  
  void bclass(int c)
***************
*** 534,540 ****
  
  #define	MAXEXPON	38	/* maximum exponent for fp number. should be IEEE */
  
! isnumber(uchar *s)	/* probably should be done by a library function */
  {
  	register int d1, d2;
  	int point;
--- 536,542 ----
  
  #define	MAXEXPON	38	/* maximum exponent for fp number. should be IEEE */
  
! is_a_number(uchar *s)	/* probably should be done by a library function */
  {
  	register int d1, d2;
  	int point;
diff -wc orig/main.c main.c
*** orig/main.c	Fri Oct 29 10:39:54 1993
--- main.c	Fri Oct 29 10:43:31 1993
***************
*** 175,177 ****
--- 175,189 ----
  		curpfile++;
  	}
  }
+ 
+ void init_input_source(void)
+ {
+ 	if (yyin == NULL) {
+ 		if (pfile[curpfile] == 0)
+ 			return;
+ 		if (strcmp((char *) pfile[curpfile], "-") == 0)
+ 			yyin = stdin;
+ 		else if ((yyin = fopen((char *) pfile[curpfile], "r")) == NULL)
+ 			ERROR "can't open file %s", pfile[curpfile] FATAL;
+ 	}
+ }
diff -wc orig/proto.h proto.h
*** orig/proto.h	Fri Oct 29 10:39:52 1993
--- proto.h	Fri Oct 29 11:22:28 1993
***************
*** 22,28 ****
  THIS SOFTWARE.
  ****************************************************************/
  
- extern	int	yywrap(void);
  extern	void	setfname(Cell *);
  extern	int	constnode(Node *);
  extern	uchar	*strnode(Node *);
--- 22,27 ----
***************
*** 31,42 ****
  
  extern	int	yylex(void);
  extern	void	startreg(void);
! extern	int	input(void);
! extern	void	unput(int);
  extern	void	unputstr(char *);
- extern	int	yylook(void);
- extern	int	yyback(int *, int);
- extern	int	yyinput(void);
  
  extern	fa	*makedfa(uchar *, int);
  extern	fa	*mkdfa(uchar *, int);
--- 30,37 ----
  
  extern	int	yylex(void);
  extern	void	startreg(void);
! extern	int	lex_input(void);
  extern	void	unputstr(char *);
  
  extern	fa	*makedfa(uchar *, int);
  extern	fa	*mkdfa(uchar *, int);
***************
*** 66,71 ****
--- 61,67 ----
  
  extern	int	main(int, uchar **);
  extern	int	pgetc(void);
+ extern	void	init_input_source(void);
  
  extern	Node	*nodealloc(int);
  extern	Node	*exptostat(Node *);
***************
*** 130,136 ****
  extern	void	bclass(int);
  extern	double	errcheck(double, uchar *);
  extern	int	isclvar(uchar *);
! extern	int	isnumber(uchar *);
  
  extern	void	run(Node *);
  extern	Cell	*r_execute(Node *);
--- 126,132 ----
  extern	void	bclass(int);
  extern	double	errcheck(double, uchar *);
  extern	int	isclvar(uchar *);
! extern	int	is_a_number(uchar *);
  
  extern	void	run(Node *);
  extern	Cell	*r_execute(Node *);
diff -wc orig/run.c run.c
*** orig/run.c	Fri Oct 29 10:39:56 1993
--- run.c	Fri Oct 29 11:22:39 1993
***************
*** 360,366 ****
  			strcpy(record, buf);
  			recloc->sval = record;
  			recloc->tval = REC | STR | DONTFREE;
! 			if (isnumber(recloc->sval)) {
  				recloc->fval = atof(recloc->sval);
  				recloc->tval |= NUM;
  			}
--- 360,366 ----
  			strcpy(record, buf);
  			recloc->sval = record;
  			recloc->tval = REC | STR | DONTFREE;
! 			if (is_a_number(recloc->sval)) {
  				recloc->fval = atof(recloc->sval);
  				recloc->tval |= NUM;
  			}
***************
*** 618,624 ****
  
  	x = execute(a[0]);
  	m = getfval(x);
! 	if (m == 0 && !isnumber(s = getsval(x)))	/* suspicion! */
  		ERROR "illegal field $(%s), name \"%s\"", s, x->nval FATAL;
    /* can x->nval ever be null??? */
  		/* ERROR "illegal field $(%s)", s FATAL; */
--- 618,624 ----
  
  	x = execute(a[0]);
  	m = getfval(x);
! 	if (m == 0 && !is_a_number(s = getsval(x)))	/* suspicion! */
  		ERROR "illegal field $(%s), name \"%s\"", s, x->nval FATAL;
    /* can x->nval ever be null??? */
  		/* ERROR "illegal field $(%s)", s FATAL; */
***************
*** 1097,1103 ****
  				sprintf((char *)num, "%d", n);
  				temp = *patbeg;
  				*patbeg = '\0';
! 				if (isnumber(s))
  					setsymtab(num, s, atof((char *)s), STR|NUM, (Array *) ap->sval);
  				else
  					setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
--- 1097,1103 ----
  				sprintf((char *)num, "%d", n);
  				temp = *patbeg;
  				*patbeg = '\0';
! 				if (is_a_number(s))
  					setsymtab(num, s, atof((char *)s), STR|NUM, (Array *) ap->sval);
  				else
  					setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
***************
*** 1114,1120 ****
  		}
  		n++;
  		sprintf((char *)num, "%d", n);
! 		if (isnumber(s))
  			setsymtab(num, s, atof((char *)s), STR|NUM, (Array *) ap->sval);
  		else
  			setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
--- 1114,1120 ----
  		}
  		n++;
  		sprintf((char *)num, "%d", n);
! 		if (is_a_number(s))
  			setsymtab(num, s, atof((char *)s), STR|NUM, (Array *) ap->sval);
  		else
  			setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
***************
*** 1134,1140 ****
  			temp = *s;
  			*s = '\0';
  			sprintf((char *)num, "%d", n);
! 			if (isnumber(t))
  				setsymtab(num, t, atof((char *)t), STR|NUM, (Array *) ap->sval);
  			else
  				setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
--- 1134,1140 ----
  			temp = *s;
  			*s = '\0';
  			sprintf((char *)num, "%d", n);
! 			if (is_a_number(t))
  				setsymtab(num, t, atof((char *)t), STR|NUM, (Array *) ap->sval);
  			else
  				setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
***************
*** 1151,1157 ****
  			temp = *s;
  			*s = '\0';
  			sprintf((char *)num, "%d", n);
! 			if (isnumber(t))
  				setsymtab(num, t, atof((char *)t), STR|NUM, (Array *) ap->sval);
  			else
  				setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
--- 1151,1157 ----
  			temp = *s;
  			*s = '\0';
  			sprintf((char *)num, "%d", n);
! 			if (is_a_number(t))
  				setsymtab(num, t, atof((char *)t), STR|NUM, (Array *) ap->sval);
  			else
  				setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
diff -wc orig/tran.c tran.c
*** orig/tran.c	Fri Oct 29 10:39:57 1993
--- tran.c	Fri Oct 29 11:22:43 1993
***************
*** 110,116 ****
  	cp->sval = (uchar *) ARGVtab;
  	for (i = 0; i < ac; i++) {
  		sprintf((char *)temp, "%d", i);
! 		if (isnumber(*av))
  			setsymtab(temp, *av, atof(*av), STR|NUM, ARGVtab);
  		else
  			setsymtab(temp, *av, 0.0, STR, ARGVtab);
--- 110,116 ----
  	cp->sval = (uchar *) ARGVtab;
  	for (i = 0; i < ac; i++) {
  		sprintf((char *)temp, "%d", i);
! 		if (is_a_number(*av))
  			setsymtab(temp, *av, atof(*av), STR|NUM, ARGVtab);
  		else
  			setsymtab(temp, *av, 0.0, STR, ARGVtab);
***************
*** 130,136 ****
  		if ((p = (uchar *) strchr((char *) *envp, '=')) == NULL)
  			continue;
  		*p++ = 0;	/* split into two strings at = */
! 		if (isnumber(p))
  			setsymtab(*envp, p, atof(p), STR|NUM, ENVtab);
  		else
  			setsymtab(*envp, p, 0.0, STR, ENVtab);
--- 130,136 ----
  		if ((p = (uchar *) strchr((char *) *envp, '=')) == NULL)
  			continue;
  		*p++ = 0;	/* split into two strings at = */
! 		if (is_a_number(p))
  			setsymtab(*envp, p, atof(p), STR|NUM, ENVtab);
  		else
  			setsymtab(*envp, p, 0.0, STR, ENVtab);
***************
*** 332,338 ****
  		recbld();
  	if (!isnum(vp)) {	/* not a number */
  		vp->fval = atof(vp->sval);	/* best guess */
! 		if (isnumber(vp->sval) && !(vp->tval&CON))
  			vp->tval |= NUM;	/* make NUM only sparingly */
  	}
  	dprintf( ("getfval %o: %s = %g, t=%o\n", vp, vp->nval, vp->fval, vp->tval) );
--- 332,338 ----
  		recbld();
  	if (!isnum(vp)) {	/* not a number */
  		vp->fval = atof(vp->sval);	/* best guess */
! 		if (is_a_number(vp->sval) && !(vp->tval&CON))
  			vp->tval |= NUM;	/* make NUM only sparingly */
  	}
  	dprintf( ("getfval %o: %s = %g, t=%o\n", vp, vp->nval, vp->fval, vp->tval) );