Return to OpenBSD-PATCHES CVS log

Up to [local] / src / usr.bin / awk

April 16, 1999 version of the "one true awk"--64-bit fixes and more.

/*	$OpenBSD: OpenBSD-PATCHES,v 1.4 1999/04/20 17:31:28 millert Exp $	*/

I merged the April 16, 1999 version of the "one true awk" from Brian Kernighan's
web page http://cm.bell-labs.com/who/bwk/

Mods to the distribution sources:
1) OpenBSD tags in all files.
2) setlocale(3) call added to main().
3) Use __progname instead of argv[0] to determine name.
4) left out win95 bits, distribution makefile, backup copies of yacc generated
   files.


Start of historical section:

--------------------------------------------------------------
These are the changes made from Research Awk to allow it to
build on OpenBSD with flex.  Below is the inof that came
with the original set of diffs (from 4.4BSD).  ytab.* was
changed to awkgram.* for make's benefit.

 - todd

From: Vern Paxson <vern@daffy.ee.lbl.gov>

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buffering of a lex scanner that are no longer valid with flex.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

--- /home/millert/tmp/awk/awk.1	Sun Jan 19 18:06:25 1997
+++ awk.1	Sun Jan 19 17:51:39 1997
@@ -1,3 +1,4 @@
+.\"	$OpenBSD: OpenBSD-PATCHES,v 1.4 1999/04/20 17:31:28 millert Exp $
 .de EX
 .nf
 .ft CW
@@ -13,7 +14,7 @@
 .SH NAME
 awk \- pattern-directed scanning and processing language
 .SH SYNOPSIS
-.B awk
+.B awk|nawk
 [
 .BI \-F
 .I fs
--- /home/millert/tmp/awk/awklex.l	Sun Jan 19 18:06:24 1997
+++ awklex.l	Sun Jan 19 18:00:52 1997
@@ -29,13 +29,15 @@
    may not be preserved in other implementations of lex.
 */
 
+#ifndef FLEX_SCANNER
 #undef	input	/* defeat lex */
 #undef	unput
+#endif /* !FLEX_SCANNER */
 
 #include <stdlib.h>
 #include <string.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 extern YYSTYPE	yylval;
 extern int	infunc;
@@ -60,6 +62,20 @@
 char	*s;
 Gstring	*gs = 0;	/* initialized in main() */
 int	cflag;
+
+#ifdef FLEX_SCANNER
+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();
+
+#define	FIRST	((yy_start - 1) / 2)
+#else /* FLEX_SCANNER */
+#define	FIRST	(yybgin - yysvec - 1)
+#endif /* FLEX_SCANNER */
 %}
 
 A	[a-zA-Z_]
@@ -70,7 +86,7 @@
 WS	[ \t]
 
 %%
-	switch (yybgin-yysvec-1) {	/* witchcraft */
+	switch (FIRST) {	/* witchcraft */
 	case 0:
 		BEGIN A;
 		break;
@@ -116,14 +132,20 @@
 
 <A>"$"{D}+	{ yylval.cp = fieldadr(atoi(yytext+1)); RET(FIELD); }
 <A>"$NF"	{ unputstr("(NF)"); return(INDIRECT); }
-<A>"$"{A}{B}*	{ int c, n;
-		  c = input(); unput(c);
-		  if (c == '(' || c == '[' || (infunc && (n=isarg(yytext+1)) >= 0)) {
-			unputstr(yytext+1);
-			return(INDIRECT);
+<A>"$"{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+1, "", 0.0, STR|NUM, symtab);
-			RET(IVAR);
+			  yylval.cp =
+				setsymtab(yytext_copy+1,"",0.0,STR|NUM,symtab);
+			  free(yytext_copy);
+			  RET(IVAR);
 		  }
 		}
 <A>"$"		{ RET(INDIRECT); }
@@ -173,12 +195,15 @@
 <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)) >= 0) {
+		  if (c != '(' && infunc && (n=isarg(yytext_copy)) >= 0) {
 			yylval.i = n;
+			free(yytext_copy);
 			RET(ARG);
 		  } else {
-			yylval.cp = setsymtab(yytext, "", 0.0, STR|NUM, symtab);
+			yylval.cp = setsymtab(yytext_copy, "", 0.0, STR|NUM, symtab);
+			free(yytext_copy);
 			if (c == '(') {
 				RET(CALL);
 			} else {
@@ -237,6 +262,32 @@
 	caddreset(gs);
 }
 
+#ifdef FLEX_SCANNER
+static int my_input( YY_CHAR *buf, int max_size )
+{
+	extern uschar *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;
+	}
+}
+#else /* FLEX_SCANNER */
 /* 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.
@@ -275,7 +326,7 @@
 	if (--ep < ebuf)
 		ep = ebuf + sizeof(ebuf) - 1;
 }
-
+#endif /* FLEX_SCANNER */
 
 void unputstr(char *s)	/* put a string back on input */
 {
@@ -285,6 +336,11 @@
 		unput(s[i]);
 }
 
+int lex_input()
+{
+	return input();
+}
+
 /* growing-string code */
 
 const int CBUFLEN = 400;
@@ -330,3 +386,20 @@
 	free((void *) gs->cbuf);
 	free((void *) gs);
 }
+
+#ifdef FLEX_SCANNER
+void init_input_source(void)
+{
+	extern int curpfile;
+	extern char *pfile[];
+
+	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;
+	}
+}
+#endif
--- /home/millert/tmp/awk/b.c	Sun Jan 19 18:06:24 1997
+++ b.c	Sun Jan 19 18:00:55 1997
@@ -31,7 +31,7 @@
 #include <string.h>
 #include <stdlib.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 #define	HAT	(NCHARS-1)	/* matches ^ in regular expr */
 				/* NCHARS is 2**n */
--- /home/millert/tmp/awk/lib.c	Sun Jan 19 18:06:24 1997
+++ lib.c	Sun Jan 19 18:01:45 1997
@@ -29,7 +29,7 @@
 #include <errno.h>
 #include <stdlib.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 FILE	*infile	= NULL;
 char	*file	= "";
@@ -431,7 +431,7 @@
 
 	if (beenhere++)
 		return;
-	while ((c = input()) != EOF && c != '\0')
+	while ((c = lex_input()) != EOF && c != '\0')
 		bclass(c);
 	bcheck2(bracecnt, '{', '}');
 	bcheck2(brackcnt, '[', ']');
@@ -479,6 +479,7 @@
 
 void eprint(void)	/* try to print context around error */
 {
+#if 0
 	char *p, *q;
 	int c;
 	static int been_here = 0;
@@ -511,6 +512,7 @@
 		}
 	putc('\n', stderr);
 	ep = ebuf;
+#endif
 }
 
 void bclass(int c)
--- /home/millert/tmp/awk/main.c	Sun Jan 19 18:06:24 1997
+++ main.c	Sun Jan 19 18:00:57 1997
@@ -27,11 +27,12 @@
 #define DEBUG
 #include <stdio.h>
 #include <ctype.h>
+#include <locale.h>
 #include <stdlib.h>
 #include <string.h>
 #include <signal.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 extern	char	**environ;
 extern	int	nfields;
@@ -53,7 +54,12 @@
 	char *fs = NULL, *marg;
 	int temp;
 
-	cmdname = argv[0];
+	setlocale(LC_ALL, "");
+
+	if ((cmdname = strrchr(argv[0], '/')) != NULL)
+		cmdname++;
+	else
+		cmdname = argv[0];
 	if (argc == 1) {
 		fprintf(stderr, "Usage: %s [-f programfile | 'program'] [-Ffieldsep] [-v var=value] [-mf n] [-mr n] [files]\n", cmdname);
 		exit(1);
--- /home/millert/tmp/awk/maketab.c	Sun Jan 19 18:06:24 1997
+++ maketab.c	Sun Jan 19 18:01:08 1997
@@ -25,14 +25,14 @@
 /*
  * this program makes the table to link function names
  * and type indices that is used by execute() in run.c.
- * it finds the indices in ytab.h, produced by yacc.
+ * it finds the indices in awkgram.h, produced by yacc.
  */
 
 #include <stdio.h>
 #include <string.h>
 #include <stdlib.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 struct xx
 {	int token;
@@ -120,12 +120,12 @@
 
 	printf("#include <stdio.h>\n");
 	printf("#include \"awk.h\"\n");
-	printf("#include \"ytab.h\"\n\n");
+	printf("#include \"awkgram.h\"\n\n");
 	for (i = SIZE; --i >= 0; )
 		names[i] = "";
 
-	if ((fp = fopen("ytab.h", "r")) == NULL) {
-		fprintf(stderr, "maketab can't open ytab.h!\n");
+	if ((fp = fopen("awkgram.h", "r")) == NULL) {
+		fprintf(stderr, "maketab can't open awkgram.h!\n");
 		exit(1);
 	}
 	printf("static char *printname[%d] = {\n", SIZE);
--- /home/millert/tmp/awk/parse.c	Sun Jan 19 18:06:24 1997
+++ parse.c	Sun Jan 19 18:01:11 1997
@@ -27,7 +27,7 @@
 #include <string.h>
 #include <stdlib.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 Node *nodealloc(int n)
 {
--- /home/millert/tmp/awk/proto.h	Sun Jan 19 18:06:24 1997
+++ proto.h	Sun Jan 19 17:51:39 1997
@@ -22,7 +22,6 @@
 USE OR PERFORMANCE OF THIS SOFTWARE.
 ****************************************************************/
 
-extern	int	yywrap(void);
 extern	void	setfname(Cell *);
 extern	int	constnode(Node *);
 extern	char	*strnode(Node *);
@@ -31,12 +30,8 @@
 
 extern	int	yylex(void);
 extern	void	startreg(void);
-extern	int	input(void);
-extern	void	unput(int);
+extern	int	lex_input(void);
 extern	void	unputstr(char *);
-extern	int	yylook(void);
-extern	int	yyback(int *, int);
-extern	int	yyinput(void);
 
 extern	fa	*makedfa(char *, int);
 extern	fa	*mkdfa(char *, int);
@@ -65,6 +60,7 @@
 extern	void	freefa(fa *);
 
 extern	int	pgetc(void);
+extern	void	init_input_source(void);
 
 extern	Node	*nodealloc(int);
 extern	Node	*exptostat(Node *);
--- /home/millert/tmp/awk/run.c	Sun Jan 19 18:06:24 1997
+++ run.c	Sun Jan 19 18:01:15 1997
@@ -31,7 +31,7 @@
 #include <stdlib.h>
 #include <time.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 #define tempfree(x)	if (istemp(x)) tfree(x); else
 
--- /home/millert/tmp/awk/tran.c	Sun Jan 19 18:06:24 1997
+++ tran.c	Sun Jan 19 18:01:18 1997
@@ -29,7 +29,7 @@
 #include <string.h>
 #include <stdlib.h>
 #include "awk.h"
-#include "ytab.h"
+#include "awkgram.h"
 
 #define	FULLTAB	2	/* rehash when table gets this x full */
 #define	GROWTAB 4	/* grow table by this factor */