xref: /freebsd-src/contrib/one-true-awk/tran.c (revision 17853db4b0dc36ed32af039cd803f13b692913da)
12a55deb1SDavid E. O'Brien /****************************************************************
22a55deb1SDavid E. O'Brien Copyright (C) Lucent Technologies 1997
32a55deb1SDavid E. O'Brien All Rights Reserved
42a55deb1SDavid E. O'Brien 
52a55deb1SDavid E. O'Brien Permission to use, copy, modify, and distribute this software and
62a55deb1SDavid E. O'Brien its documentation for any purpose and without fee is hereby
72a55deb1SDavid E. O'Brien granted, provided that the above copyright notice appear in all
82a55deb1SDavid E. O'Brien copies and that both that the copyright notice and this
92a55deb1SDavid E. O'Brien permission notice and warranty disclaimer appear in supporting
102a55deb1SDavid E. O'Brien documentation, and that the name Lucent Technologies or any of
112a55deb1SDavid E. O'Brien its entities not be used in advertising or publicity pertaining
122a55deb1SDavid E. O'Brien to distribution of the software without specific, written prior
132a55deb1SDavid E. O'Brien permission.
142a55deb1SDavid E. O'Brien 
152a55deb1SDavid E. O'Brien LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
162a55deb1SDavid E. O'Brien INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
172a55deb1SDavid E. O'Brien IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
182a55deb1SDavid E. O'Brien SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
192a55deb1SDavid E. O'Brien WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
202a55deb1SDavid E. O'Brien IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
212a55deb1SDavid E. O'Brien ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
222a55deb1SDavid E. O'Brien THIS SOFTWARE.
232a55deb1SDavid E. O'Brien ****************************************************************/
242a55deb1SDavid E. O'Brien 
252a55deb1SDavid E. O'Brien #define	DEBUG
262a55deb1SDavid E. O'Brien #include <stdio.h>
272a55deb1SDavid E. O'Brien #include <math.h>
282a55deb1SDavid E. O'Brien #include <ctype.h>
292a55deb1SDavid E. O'Brien #include <string.h>
302a55deb1SDavid E. O'Brien #include <stdlib.h>
312a55deb1SDavid E. O'Brien #include "awk.h"
322a55deb1SDavid E. O'Brien 
332a55deb1SDavid E. O'Brien #define	FULLTAB	2	/* rehash when table gets this x full */
342a55deb1SDavid E. O'Brien #define	GROWTAB 4	/* grow table by this factor */
352a55deb1SDavid E. O'Brien 
362a55deb1SDavid E. O'Brien Array	*symtab;	/* main symbol table */
372a55deb1SDavid E. O'Brien 
382a55deb1SDavid E. O'Brien char	**FS;		/* initial field sep */
392a55deb1SDavid E. O'Brien char	**RS;		/* initial record sep */
402a55deb1SDavid E. O'Brien char	**OFS;		/* output field sep */
412a55deb1SDavid E. O'Brien char	**ORS;		/* output record sep */
422a55deb1SDavid E. O'Brien char	**OFMT;		/* output format for numbers */
432a55deb1SDavid E. O'Brien char	**CONVFMT;	/* format for conversions in getsval */
442a55deb1SDavid E. O'Brien Awkfloat *NF;		/* number of fields in current record */
452a55deb1SDavid E. O'Brien Awkfloat *NR;		/* number of current record */
462a55deb1SDavid E. O'Brien Awkfloat *FNR;		/* number of current record in current file */
472a55deb1SDavid E. O'Brien char	**FILENAME;	/* current filename argument */
482a55deb1SDavid E. O'Brien Awkfloat *ARGC;		/* number of arguments from command line */
492a55deb1SDavid E. O'Brien char	**SUBSEP;	/* subscript separator for a[i,j,k]; default \034 */
502a55deb1SDavid E. O'Brien Awkfloat *RSTART;	/* start of re matched with ~; origin 1 (!) */
512a55deb1SDavid E. O'Brien Awkfloat *RLENGTH;	/* length of same */
522a55deb1SDavid E. O'Brien 
53c263f9bfSRuslan Ermilov Cell	*fsloc;		/* FS */
542a55deb1SDavid E. O'Brien Cell	*nrloc;		/* NR */
552a55deb1SDavid E. O'Brien Cell	*nfloc;		/* NF */
562a55deb1SDavid E. O'Brien Cell	*fnrloc;	/* FNR */
57b5253557SWarner Losh Cell	*ofsloc;	/* OFS */
58b5253557SWarner Losh Cell	*orsloc;	/* ORS */
59b5253557SWarner Losh Cell	*rsloc;		/* RS */
60*17853db4SWarner Losh Cell	*ARGVcell;	/* cell with symbol table containing ARGV[...] */
612a55deb1SDavid E. O'Brien Cell	*rstartloc;	/* RSTART */
622a55deb1SDavid E. O'Brien Cell	*rlengthloc;	/* RLENGTH */
63b5253557SWarner Losh Cell	*subseploc;	/* SUBSEP */
642a55deb1SDavid E. O'Brien Cell	*symtabloc;	/* SYMTAB */
652a55deb1SDavid E. O'Brien 
662a55deb1SDavid E. O'Brien Cell	*nullloc;	/* a guaranteed empty cell */
672a55deb1SDavid E. O'Brien Node	*nullnode;	/* zero&null, converted into a node for comparisons */
682a55deb1SDavid E. O'Brien Cell	*literal0;
692a55deb1SDavid E. O'Brien 
702a55deb1SDavid E. O'Brien extern Cell **fldtab;
712a55deb1SDavid E. O'Brien 
722a55deb1SDavid E. O'Brien void syminit(void)	/* initialize symbol table with builtin vars */
732a55deb1SDavid E. O'Brien {
742a55deb1SDavid E. O'Brien 	literal0 = setsymtab("0", "0", 0.0, NUM|STR|CON|DONTFREE, symtab);
752a55deb1SDavid E. O'Brien 	/* this is used for if(x)... tests: */
762a55deb1SDavid E. O'Brien 	nullloc = setsymtab("$zero&null", "", 0.0, NUM|STR|CON|DONTFREE, symtab);
772a55deb1SDavid E. O'Brien 	nullnode = celltonode(nullloc, CCON);
782a55deb1SDavid E. O'Brien 
79c263f9bfSRuslan Ermilov 	fsloc = setsymtab("FS", " ", 0.0, STR|DONTFREE, symtab);
80c263f9bfSRuslan Ermilov 	FS = &fsloc->sval;
81b5253557SWarner Losh 	rsloc = setsymtab("RS", "\n", 0.0, STR|DONTFREE, symtab);
82b5253557SWarner Losh 	RS = &rsloc->sval;
83b5253557SWarner Losh 	ofsloc = setsymtab("OFS", " ", 0.0, STR|DONTFREE, symtab);
84b5253557SWarner Losh 	OFS = &ofsloc->sval;
85b5253557SWarner Losh 	orsloc = setsymtab("ORS", "\n", 0.0, STR|DONTFREE, symtab);
86b5253557SWarner Losh 	ORS = &orsloc->sval;
872a55deb1SDavid E. O'Brien 	OFMT = &setsymtab("OFMT", "%.6g", 0.0, STR|DONTFREE, symtab)->sval;
882a55deb1SDavid E. O'Brien 	CONVFMT = &setsymtab("CONVFMT", "%.6g", 0.0, STR|DONTFREE, symtab)->sval;
892a55deb1SDavid E. O'Brien 	FILENAME = &setsymtab("FILENAME", "", 0.0, STR|DONTFREE, symtab)->sval;
902a55deb1SDavid E. O'Brien 	nfloc = setsymtab("NF", "", 0.0, NUM, symtab);
912a55deb1SDavid E. O'Brien 	NF = &nfloc->fval;
922a55deb1SDavid E. O'Brien 	nrloc = setsymtab("NR", "", 0.0, NUM, symtab);
932a55deb1SDavid E. O'Brien 	NR = &nrloc->fval;
942a55deb1SDavid E. O'Brien 	fnrloc = setsymtab("FNR", "", 0.0, NUM, symtab);
952a55deb1SDavid E. O'Brien 	FNR = &fnrloc->fval;
96b5253557SWarner Losh 	subseploc = setsymtab("SUBSEP", "\034", 0.0, STR|DONTFREE, symtab);
97b5253557SWarner Losh 	SUBSEP = &subseploc->sval;
982a55deb1SDavid E. O'Brien 	rstartloc = setsymtab("RSTART", "", 0.0, NUM, symtab);
992a55deb1SDavid E. O'Brien 	RSTART = &rstartloc->fval;
1002a55deb1SDavid E. O'Brien 	rlengthloc = setsymtab("RLENGTH", "", 0.0, NUM, symtab);
1012a55deb1SDavid E. O'Brien 	RLENGTH = &rlengthloc->fval;
1022a55deb1SDavid E. O'Brien 	symtabloc = setsymtab("SYMTAB", "", 0.0, ARR, symtab);
103f39dd6a9SWarner Losh 	free(symtabloc->sval);
1042a55deb1SDavid E. O'Brien 	symtabloc->sval = (char *) symtab;
1052a55deb1SDavid E. O'Brien }
1062a55deb1SDavid E. O'Brien 
1072a55deb1SDavid E. O'Brien void arginit(int ac, char **av)	/* set up ARGV and ARGC */
1082a55deb1SDavid E. O'Brien {
109*17853db4SWarner Losh 	Array *ap;
1102a55deb1SDavid E. O'Brien 	Cell *cp;
1112a55deb1SDavid E. O'Brien 	int i;
1122a55deb1SDavid E. O'Brien 	char temp[50];
1132a55deb1SDavid E. O'Brien 
1142a55deb1SDavid E. O'Brien 	ARGC = &setsymtab("ARGC", "", (Awkfloat) ac, NUM, symtab)->fval;
1152a55deb1SDavid E. O'Brien 	cp = setsymtab("ARGV", "", 0.0, ARR, symtab);
116*17853db4SWarner Losh 	ap = makesymtab(NSYMTAB);	/* could be (int) ARGC as well */
117f39dd6a9SWarner Losh 	free(cp->sval);
118*17853db4SWarner Losh 	cp->sval = (char *) ap;
1192a55deb1SDavid E. O'Brien 	for (i = 0; i < ac; i++) {
120f39dd6a9SWarner Losh 		double result;
121f39dd6a9SWarner Losh 
1222a55deb1SDavid E. O'Brien 		sprintf(temp, "%d", i);
123f39dd6a9SWarner Losh 		if (is_number(*av, & result))
124*17853db4SWarner Losh 			setsymtab(temp, *av, result, STR|NUM, ap);
1252a55deb1SDavid E. O'Brien 		else
126*17853db4SWarner Losh 			setsymtab(temp, *av, 0.0, STR, ap);
1272a55deb1SDavid E. O'Brien 		av++;
1282a55deb1SDavid E. O'Brien 	}
129*17853db4SWarner Losh 	ARGVcell = cp;
1302a55deb1SDavid E. O'Brien }
1312a55deb1SDavid E. O'Brien 
1322a55deb1SDavid E. O'Brien void envinit(char **envp)	/* set up ENVIRON variable */
1332a55deb1SDavid E. O'Brien {
134*17853db4SWarner Losh 	Array *ap;
1352a55deb1SDavid E. O'Brien 	Cell *cp;
1362a55deb1SDavid E. O'Brien 	char *p;
1372a55deb1SDavid E. O'Brien 
1382a55deb1SDavid E. O'Brien 	cp = setsymtab("ENVIRON", "", 0.0, ARR, symtab);
139*17853db4SWarner Losh 	ap = makesymtab(NSYMTAB);
140f39dd6a9SWarner Losh 	free(cp->sval);
141*17853db4SWarner Losh 	cp->sval = (char *) ap;
1422a55deb1SDavid E. O'Brien 	for ( ; *envp; envp++) {
143f39dd6a9SWarner Losh 		double result;
144f39dd6a9SWarner Losh 
1452a55deb1SDavid E. O'Brien 		if ((p = strchr(*envp, '=')) == NULL)
1462a55deb1SDavid E. O'Brien 			continue;
1472a55deb1SDavid E. O'Brien 		if( p == *envp ) /* no left hand side name in env string */
1482a55deb1SDavid E. O'Brien 			continue;
1492a55deb1SDavid E. O'Brien 		*p++ = 0;	/* split into two strings at = */
150f39dd6a9SWarner Losh 		if (is_number(p, & result))
151*17853db4SWarner Losh 			setsymtab(*envp, p, result, STR|NUM, ap);
1522a55deb1SDavid E. O'Brien 		else
153*17853db4SWarner Losh 			setsymtab(*envp, p, 0.0, STR, ap);
1542a55deb1SDavid E. O'Brien 		p[-1] = '=';	/* restore in case env is passed down to a shell */
1552a55deb1SDavid E. O'Brien 	}
1562a55deb1SDavid E. O'Brien }
1572a55deb1SDavid E. O'Brien 
1582a55deb1SDavid E. O'Brien Array *makesymtab(int n)	/* make a new symbol table */
1592a55deb1SDavid E. O'Brien {
1602a55deb1SDavid E. O'Brien 	Array *ap;
1612a55deb1SDavid E. O'Brien 	Cell **tp;
1622a55deb1SDavid E. O'Brien 
163f39dd6a9SWarner Losh 	ap = (Array *) malloc(sizeof(*ap));
164f39dd6a9SWarner Losh 	tp = (Cell **) calloc(n, sizeof(*tp));
1652a55deb1SDavid E. O'Brien 	if (ap == NULL || tp == NULL)
1662a55deb1SDavid E. O'Brien 		FATAL("out of space in makesymtab");
1672a55deb1SDavid E. O'Brien 	ap->nelem = 0;
1682a55deb1SDavid E. O'Brien 	ap->size = n;
1692a55deb1SDavid E. O'Brien 	ap->tab = tp;
1702a55deb1SDavid E. O'Brien 	return(ap);
1712a55deb1SDavid E. O'Brien }
1722a55deb1SDavid E. O'Brien 
1732a55deb1SDavid E. O'Brien void freesymtab(Cell *ap)	/* free a symbol table */
1742a55deb1SDavid E. O'Brien {
1752a55deb1SDavid E. O'Brien 	Cell *cp, *temp;
1762a55deb1SDavid E. O'Brien 	Array *tp;
1772a55deb1SDavid E. O'Brien 	int i;
1782a55deb1SDavid E. O'Brien 
1792a55deb1SDavid E. O'Brien 	if (!isarr(ap))
1802a55deb1SDavid E. O'Brien 		return;
1812a55deb1SDavid E. O'Brien 	tp = (Array *) ap->sval;
1822a55deb1SDavid E. O'Brien 	if (tp == NULL)
1832a55deb1SDavid E. O'Brien 		return;
1842a55deb1SDavid E. O'Brien 	for (i = 0; i < tp->size; i++) {
1852a55deb1SDavid E. O'Brien 		for (cp = tp->tab[i]; cp != NULL; cp = temp) {
1862a55deb1SDavid E. O'Brien 			xfree(cp->nval);
1872a55deb1SDavid E. O'Brien 			if (freeable(cp))
1882a55deb1SDavid E. O'Brien 				xfree(cp->sval);
1892a55deb1SDavid E. O'Brien 			temp = cp->cnext;	/* avoids freeing then using */
1902a55deb1SDavid E. O'Brien 			free(cp);
191007c6572SDag-Erling Smørgrav 			tp->nelem--;
1922a55deb1SDavid E. O'Brien 		}
19310ce5b99SWarner Losh 		tp->tab[i] = NULL;
1942a55deb1SDavid E. O'Brien 	}
195007c6572SDag-Erling Smørgrav 	if (tp->nelem != 0)
196007c6572SDag-Erling Smørgrav 		WARNING("can't happen: inconsistent element count freeing %s", ap->nval);
1972a55deb1SDavid E. O'Brien 	free(tp->tab);
1982a55deb1SDavid E. O'Brien 	free(tp);
1992a55deb1SDavid E. O'Brien }
2002a55deb1SDavid E. O'Brien 
201813da98dSDavid E. O'Brien void freeelem(Cell *ap, const char *s)	/* free elem s from ap (i.e., ap["s"] */
2022a55deb1SDavid E. O'Brien {
2032a55deb1SDavid E. O'Brien 	Array *tp;
2042a55deb1SDavid E. O'Brien 	Cell *p, *prev = NULL;
2052a55deb1SDavid E. O'Brien 	int h;
2062a55deb1SDavid E. O'Brien 
2072a55deb1SDavid E. O'Brien 	tp = (Array *) ap->sval;
2082a55deb1SDavid E. O'Brien 	h = hash(s, tp->size);
2092a55deb1SDavid E. O'Brien 	for (p = tp->tab[h]; p != NULL; prev = p, p = p->cnext)
2102a55deb1SDavid E. O'Brien 		if (strcmp(s, p->nval) == 0) {
2112a55deb1SDavid E. O'Brien 			if (prev == NULL)	/* 1st one */
2122a55deb1SDavid E. O'Brien 				tp->tab[h] = p->cnext;
2132a55deb1SDavid E. O'Brien 			else			/* middle somewhere */
2142a55deb1SDavid E. O'Brien 				prev->cnext = p->cnext;
2152a55deb1SDavid E. O'Brien 			if (freeable(p))
2162a55deb1SDavid E. O'Brien 				xfree(p->sval);
2172a55deb1SDavid E. O'Brien 			free(p->nval);
2182a55deb1SDavid E. O'Brien 			free(p);
2192a55deb1SDavid E. O'Brien 			tp->nelem--;
2202a55deb1SDavid E. O'Brien 			return;
2212a55deb1SDavid E. O'Brien 		}
2222a55deb1SDavid E. O'Brien }
2232a55deb1SDavid E. O'Brien 
224813da98dSDavid E. O'Brien Cell *setsymtab(const char *n, const char *s, Awkfloat f, unsigned t, Array *tp)
2252a55deb1SDavid E. O'Brien {
2262a55deb1SDavid E. O'Brien 	int h;
2272a55deb1SDavid E. O'Brien 	Cell *p;
2282a55deb1SDavid E. O'Brien 
2292a55deb1SDavid E. O'Brien 	if (n != NULL && (p = lookup(n, tp)) != NULL) {
230f39dd6a9SWarner Losh 		DPRINTF("setsymtab found %p: n=%s s=\"%s\" f=%g t=%o\n",
231f39dd6a9SWarner Losh 			(void*)p, NN(p->nval), NN(p->sval), p->fval, p->tval);
2322a55deb1SDavid E. O'Brien 		return(p);
2332a55deb1SDavid E. O'Brien 	}
234f39dd6a9SWarner Losh 	p = (Cell *) malloc(sizeof(*p));
2352a55deb1SDavid E. O'Brien 	if (p == NULL)
2362a55deb1SDavid E. O'Brien 		FATAL("out of space for symbol table at %s", n);
2372a55deb1SDavid E. O'Brien 	p->nval = tostring(n);
2382a55deb1SDavid E. O'Brien 	p->sval = s ? tostring(s) : tostring("");
2392a55deb1SDavid E. O'Brien 	p->fval = f;
2402a55deb1SDavid E. O'Brien 	p->tval = t;
2412a55deb1SDavid E. O'Brien 	p->csub = CUNK;
2422a55deb1SDavid E. O'Brien 	p->ctype = OCELL;
2432a55deb1SDavid E. O'Brien 	tp->nelem++;
2442a55deb1SDavid E. O'Brien 	if (tp->nelem > FULLTAB * tp->size)
2452a55deb1SDavid E. O'Brien 		rehash(tp);
2462a55deb1SDavid E. O'Brien 	h = hash(n, tp->size);
2472a55deb1SDavid E. O'Brien 	p->cnext = tp->tab[h];
2482a55deb1SDavid E. O'Brien 	tp->tab[h] = p;
249f39dd6a9SWarner Losh 	DPRINTF("setsymtab set %p: n=%s s=\"%s\" f=%g t=%o\n",
250f39dd6a9SWarner Losh 		(void*)p, p->nval, p->sval, p->fval, p->tval);
2512a55deb1SDavid E. O'Brien 	return(p);
2522a55deb1SDavid E. O'Brien }
2532a55deb1SDavid E. O'Brien 
254813da98dSDavid E. O'Brien int hash(const char *s, int n)	/* form hash value for string s */
2552a55deb1SDavid E. O'Brien {
2562a55deb1SDavid E. O'Brien 	unsigned hashval;
2572a55deb1SDavid E. O'Brien 
2582a55deb1SDavid E. O'Brien 	for (hashval = 0; *s != '\0'; s++)
2592a55deb1SDavid E. O'Brien 		hashval = (*s + 31 * hashval);
2602a55deb1SDavid E. O'Brien 	return hashval % n;
2612a55deb1SDavid E. O'Brien }
2622a55deb1SDavid E. O'Brien 
2632a55deb1SDavid E. O'Brien void rehash(Array *tp)	/* rehash items in small table into big one */
2642a55deb1SDavid E. O'Brien {
2652a55deb1SDavid E. O'Brien 	int i, nh, nsz;
2662a55deb1SDavid E. O'Brien 	Cell *cp, *op, **np;
2672a55deb1SDavid E. O'Brien 
2682a55deb1SDavid E. O'Brien 	nsz = GROWTAB * tp->size;
269f39dd6a9SWarner Losh 	np = (Cell **) calloc(nsz, sizeof(*np));
2702a55deb1SDavid E. O'Brien 	if (np == NULL)		/* can't do it, but can keep running. */
2712a55deb1SDavid E. O'Brien 		return;		/* someone else will run out later. */
2722a55deb1SDavid E. O'Brien 	for (i = 0; i < tp->size; i++) {
2732a55deb1SDavid E. O'Brien 		for (cp = tp->tab[i]; cp; cp = op) {
2742a55deb1SDavid E. O'Brien 			op = cp->cnext;
2752a55deb1SDavid E. O'Brien 			nh = hash(cp->nval, nsz);
2762a55deb1SDavid E. O'Brien 			cp->cnext = np[nh];
2772a55deb1SDavid E. O'Brien 			np[nh] = cp;
2782a55deb1SDavid E. O'Brien 		}
2792a55deb1SDavid E. O'Brien 	}
2802a55deb1SDavid E. O'Brien 	free(tp->tab);
2812a55deb1SDavid E. O'Brien 	tp->tab = np;
2822a55deb1SDavid E. O'Brien 	tp->size = nsz;
2832a55deb1SDavid E. O'Brien }
2842a55deb1SDavid E. O'Brien 
285813da98dSDavid E. O'Brien Cell *lookup(const char *s, Array *tp)	/* look for s in tp */
2862a55deb1SDavid E. O'Brien {
2872a55deb1SDavid E. O'Brien 	Cell *p;
2882a55deb1SDavid E. O'Brien 	int h;
2892a55deb1SDavid E. O'Brien 
2902a55deb1SDavid E. O'Brien 	h = hash(s, tp->size);
2912a55deb1SDavid E. O'Brien 	for (p = tp->tab[h]; p != NULL; p = p->cnext)
2922a55deb1SDavid E. O'Brien 		if (strcmp(s, p->nval) == 0)
2932a55deb1SDavid E. O'Brien 			return(p);	/* found it */
2942a55deb1SDavid E. O'Brien 	return(NULL);			/* not found */
2952a55deb1SDavid E. O'Brien }
2962a55deb1SDavid E. O'Brien 
2972a55deb1SDavid E. O'Brien Awkfloat setfval(Cell *vp, Awkfloat f)	/* set float val of a Cell */
2982a55deb1SDavid E. O'Brien {
2992a55deb1SDavid E. O'Brien 	int fldno;
3002a55deb1SDavid E. O'Brien 
301b5253557SWarner Losh 	f += 0.0;		/* normalise negative zero to positive zero */
3022a55deb1SDavid E. O'Brien 	if ((vp->tval & (NUM | STR)) == 0)
3032a55deb1SDavid E. O'Brien 		funnyvar(vp, "assign to");
3042a55deb1SDavid E. O'Brien 	if (isfld(vp)) {
305f39dd6a9SWarner Losh 		donerec = false;	/* mark $0 invalid */
3062a55deb1SDavid E. O'Brien 		fldno = atoi(vp->nval);
3072a55deb1SDavid E. O'Brien 		if (fldno > *NF)
3082a55deb1SDavid E. O'Brien 			newfld(fldno);
309f39dd6a9SWarner Losh 		DPRINTF("setting field %d to %g\n", fldno, f);
310b5253557SWarner Losh 	} else if (&vp->fval == NF) {
311f39dd6a9SWarner Losh 		donerec = false;	/* mark $0 invalid */
312b5253557SWarner Losh 		setlastfld(f);
313f32a6403SWarner Losh 		DPRINTF("setfval: setting NF to %g\n", f);
3142a55deb1SDavid E. O'Brien 	} else if (isrec(vp)) {
315f39dd6a9SWarner Losh 		donefld = false;	/* mark $1... invalid */
316f39dd6a9SWarner Losh 		donerec = true;
317f39dd6a9SWarner Losh 		savefs();
318b5253557SWarner Losh 	} else if (vp == ofsloc) {
319f39dd6a9SWarner Losh 		if (!donerec)
320b5253557SWarner Losh 			recbld();
3212a55deb1SDavid E. O'Brien 	}
3222a55deb1SDavid E. O'Brien 	if (freeable(vp))
3232a55deb1SDavid E. O'Brien 		xfree(vp->sval); /* free any previous string */
324b5253557SWarner Losh 	vp->tval &= ~(STR|CONVC|CONVO); /* mark string invalid */
325b5253557SWarner Losh 	vp->fmt = NULL;
3262a55deb1SDavid E. O'Brien 	vp->tval |= NUM;	/* mark number ok */
3270840e960SXin LI 	if (f == -0)  /* who would have thought this possible? */
3280840e960SXin LI 		f = 0;
329f39dd6a9SWarner Losh 	DPRINTF("setfval %p: %s = %g, t=%o\n", (void*)vp, NN(vp->nval), f, vp->tval);
3302a55deb1SDavid E. O'Brien 	return vp->fval = f;
3312a55deb1SDavid E. O'Brien }
3322a55deb1SDavid E. O'Brien 
333813da98dSDavid E. O'Brien void funnyvar(Cell *vp, const char *rw)
3342a55deb1SDavid E. O'Brien {
3352a55deb1SDavid E. O'Brien 	if (isarr(vp))
3362a55deb1SDavid E. O'Brien 		FATAL("can't %s %s; it's an array name.", rw, vp->nval);
3372a55deb1SDavid E. O'Brien 	if (vp->tval & FCN)
3382a55deb1SDavid E. O'Brien 		FATAL("can't %s %s; it's a function.", rw, vp->nval);
3392a55deb1SDavid E. O'Brien 	WARNING("funny variable %p: n=%s s=\"%s\" f=%g t=%o",
340f39dd6a9SWarner Losh 		(void *)vp, vp->nval, vp->sval, vp->fval, vp->tval);
3412a55deb1SDavid E. O'Brien }
3422a55deb1SDavid E. O'Brien 
343813da98dSDavid E. O'Brien char *setsval(Cell *vp, const char *s)	/* set string val of a Cell */
3442a55deb1SDavid E. O'Brien {
3452a55deb1SDavid E. O'Brien 	char *t;
3462a55deb1SDavid E. O'Brien 	int fldno;
347b5253557SWarner Losh 	Awkfloat f;
3482a55deb1SDavid E. O'Brien 
349f39dd6a9SWarner Losh 	DPRINTF("starting setsval %p: %s = \"%s\", t=%o, r,f=%d,%d\n",
350f39dd6a9SWarner Losh 		(void*)vp, NN(vp->nval), s, vp->tval, donerec, donefld);
3512a55deb1SDavid E. O'Brien 	if ((vp->tval & (NUM | STR)) == 0)
3522a55deb1SDavid E. O'Brien 		funnyvar(vp, "assign to");
353f32a6403SWarner Losh 	if (CSV && (vp == rsloc))
354f32a6403SWarner Losh 		WARNING("danger: don't set RS when --csv is in effect");
355f32a6403SWarner Losh 	if (CSV && (vp == fsloc))
356f32a6403SWarner Losh 		WARNING("danger: don't set FS when --csv is in effect");
3572a55deb1SDavid E. O'Brien 	if (isfld(vp)) {
358f39dd6a9SWarner Losh 		donerec = false;	/* mark $0 invalid */
3592a55deb1SDavid E. O'Brien 		fldno = atoi(vp->nval);
3602a55deb1SDavid E. O'Brien 		if (fldno > *NF)
3612a55deb1SDavid E. O'Brien 			newfld(fldno);
362f39dd6a9SWarner Losh 		DPRINTF("setting field %d to %s (%p)\n", fldno, s, (const void*)s);
3632a55deb1SDavid E. O'Brien 	} else if (isrec(vp)) {
364f39dd6a9SWarner Losh 		donefld = false;	/* mark $1... invalid */
365f39dd6a9SWarner Losh 		donerec = true;
366f39dd6a9SWarner Losh 		savefs();
367b5253557SWarner Losh 	} else if (vp == ofsloc) {
368f39dd6a9SWarner Losh 		if (!donerec)
369b5253557SWarner Losh 			recbld();
3702a55deb1SDavid E. O'Brien 	}
371adb46ac4SWarner Losh 	t = s ? tostring(s) : tostring("");	/* in case it's self-assign */
3722a55deb1SDavid E. O'Brien 	if (freeable(vp))
3732a55deb1SDavid E. O'Brien 		xfree(vp->sval);
374f32a6403SWarner Losh 	vp->tval &= ~(NUM|DONTFREE|CONVC|CONVO);
375d2f6e492SDavid E. O'Brien 	vp->tval |= STR;
376b5253557SWarner Losh 	vp->fmt = NULL;
377f39dd6a9SWarner Losh 	DPRINTF("setsval %p: %s = \"%s (%p) \", t=%o r,f=%d,%d\n",
378f39dd6a9SWarner Losh 		(void*)vp, NN(vp->nval), t, (void*)t, vp->tval, donerec, donefld);
379b5253557SWarner Losh 	vp->sval = t;
380b5253557SWarner Losh 	if (&vp->fval == NF) {
381f39dd6a9SWarner Losh 		donerec = false;	/* mark $0 invalid */
382b5253557SWarner Losh 		f = getfval(vp);
383b5253557SWarner Losh 		setlastfld(f);
384f32a6403SWarner Losh 		DPRINTF("setsval: setting NF to %g\n", f);
385b5253557SWarner Losh 	}
386b5253557SWarner Losh 
387b5253557SWarner Losh 	return(vp->sval);
3882a55deb1SDavid E. O'Brien }
3892a55deb1SDavid E. O'Brien 
3902a55deb1SDavid E. O'Brien Awkfloat getfval(Cell *vp)	/* get float val of a Cell */
3912a55deb1SDavid E. O'Brien {
3922a55deb1SDavid E. O'Brien 	if ((vp->tval & (NUM | STR)) == 0)
3932a55deb1SDavid E. O'Brien 		funnyvar(vp, "read value of");
394f39dd6a9SWarner Losh 	if (isfld(vp) && !donefld)
3952a55deb1SDavid E. O'Brien 		fldbld();
396f39dd6a9SWarner Losh 	else if (isrec(vp) && !donerec)
3972a55deb1SDavid E. O'Brien 		recbld();
3982a55deb1SDavid E. O'Brien 	if (!isnum(vp)) {	/* not a number */
399f39dd6a9SWarner Losh 		double fval;
400f39dd6a9SWarner Losh 		bool no_trailing;
401f39dd6a9SWarner Losh 
402f39dd6a9SWarner Losh 		if (is_valid_number(vp->sval, true, & no_trailing, & fval)) {
403f39dd6a9SWarner Losh 			vp->fval = fval;
404f39dd6a9SWarner Losh 			if (no_trailing && !(vp->tval&CON))
4052a55deb1SDavid E. O'Brien 				vp->tval |= NUM;	/* make NUM only sparingly */
406f39dd6a9SWarner Losh 		} else
407f39dd6a9SWarner Losh 			vp->fval = 0.0;
4082a55deb1SDavid E. O'Brien 	}
409f39dd6a9SWarner Losh 	DPRINTF("getfval %p: %s = %g, t=%o\n",
410f39dd6a9SWarner Losh 		(void*)vp, NN(vp->nval), vp->fval, vp->tval);
4112a55deb1SDavid E. O'Brien 	return(vp->fval);
4122a55deb1SDavid E. O'Brien }
4132a55deb1SDavid E. O'Brien 
414f39dd6a9SWarner Losh static const char *get_inf_nan(double d)
415f39dd6a9SWarner Losh {
416f39dd6a9SWarner Losh 	if (isinf(d)) {
417f39dd6a9SWarner Losh 		return (d < 0 ? "-inf" : "+inf");
418f39dd6a9SWarner Losh 	} else if (isnan(d)) {
419f39dd6a9SWarner Losh 		return (signbit(d) != 0 ? "-nan" : "+nan");
420f39dd6a9SWarner Losh 	} else
421f39dd6a9SWarner Losh 		return NULL;
422f39dd6a9SWarner Losh }
423f39dd6a9SWarner Losh 
424813da98dSDavid E. O'Brien static char *get_str_val(Cell *vp, char **fmt)        /* get string val of a Cell */
4252a55deb1SDavid E. O'Brien {
426b5253557SWarner Losh 	char s[256];
4272a55deb1SDavid E. O'Brien 	double dtemp;
428f39dd6a9SWarner Losh 	const char *p;
4292a55deb1SDavid E. O'Brien 
4302a55deb1SDavid E. O'Brien 	if ((vp->tval & (NUM | STR)) == 0)
4312a55deb1SDavid E. O'Brien 		funnyvar(vp, "read value of");
432f39dd6a9SWarner Losh 	if (isfld(vp) && ! donefld)
4332a55deb1SDavid E. O'Brien 		fldbld();
434f39dd6a9SWarner Losh 	else if (isrec(vp) && ! donerec)
4352a55deb1SDavid E. O'Brien 		recbld();
436b5253557SWarner Losh 
437b5253557SWarner Losh 	/*
438b5253557SWarner Losh 	 * ADR: This is complicated and more fragile than is desirable.
439b5253557SWarner Losh 	 * Retrieving a string value for a number associates the string
440b5253557SWarner Losh 	 * value with the scalar.  Previously, the string value was
441b5253557SWarner Losh 	 * sticky, meaning if converted via OFMT that became the value
442b5253557SWarner Losh 	 * (even though POSIX wants it to be via CONVFMT). Or if CONVFMT
443b5253557SWarner Losh 	 * changed after a string value was retrieved, the original value
444b5253557SWarner Losh 	 * was maintained and used.  Also not per POSIX.
445b5253557SWarner Losh 	 *
446b5253557SWarner Losh 	 * We work around this design by adding two additional flags,
447b5253557SWarner Losh 	 * CONVC and CONVO, indicating how the string value was
448b5253557SWarner Losh 	 * obtained (via CONVFMT or OFMT) and _also_ maintaining a copy
449b5253557SWarner Losh 	 * of the pointer to the xFMT format string used for the
450b5253557SWarner Losh 	 * conversion.  This pointer is only read, **never** dereferenced.
451b5253557SWarner Losh 	 * The next time we do a conversion, if it's coming from the same
452b5253557SWarner Losh 	 * xFMT as last time, and the pointer value is different, we
453b5253557SWarner Losh 	 * know that the xFMT format string changed, and we need to
454b5253557SWarner Losh 	 * redo the conversion. If it's the same, we don't have to.
455b5253557SWarner Losh 	 *
456b5253557SWarner Losh 	 * There are also several cases where we don't do a conversion,
457b5253557SWarner Losh 	 * such as for a field (see the checks below).
458b5253557SWarner Losh 	 */
459b5253557SWarner Losh 
460b5253557SWarner Losh 	/* Don't duplicate the code for actually updating the value */
461b5253557SWarner Losh #define update_str_val(vp) \
462b5253557SWarner Losh 	{ \
463b5253557SWarner Losh 		if (freeable(vp)) \
464b5253557SWarner Losh 			xfree(vp->sval); \
465f39dd6a9SWarner Losh 		if ((p = get_inf_nan(vp->fval)) != NULL) \
466f39dd6a9SWarner Losh 			strcpy(s, p); \
467f39dd6a9SWarner Losh 		else if (modf(vp->fval, &dtemp) == 0)	/* it's integral */ \
468b5253557SWarner Losh 			snprintf(s, sizeof (s), "%.30g", vp->fval); \
469b5253557SWarner Losh 		else \
470b5253557SWarner Losh 			snprintf(s, sizeof (s), *fmt, vp->fval); \
471b5253557SWarner Losh 		vp->sval = tostring(s); \
472b5253557SWarner Losh 		vp->tval &= ~DONTFREE; \
473b5253557SWarner Losh 		vp->tval |= STR; \
4742a55deb1SDavid E. O'Brien 	}
475b5253557SWarner Losh 
476b5253557SWarner Losh 	if (isstr(vp) == 0) {
477b5253557SWarner Losh 		update_str_val(vp);
478b5253557SWarner Losh 		if (fmt == OFMT) {
479b5253557SWarner Losh 			vp->tval &= ~CONVC;
480b5253557SWarner Losh 			vp->tval |= CONVO;
481b5253557SWarner Losh 		} else {
482b5253557SWarner Losh 			/* CONVFMT */
483b5253557SWarner Losh 			vp->tval &= ~CONVO;
484b5253557SWarner Losh 			vp->tval |= CONVC;
485b5253557SWarner Losh 		}
486b5253557SWarner Losh 		vp->fmt = *fmt;
487b5253557SWarner Losh 	} else if ((vp->tval & DONTFREE) != 0 || ! isnum(vp) || isfld(vp)) {
488b5253557SWarner Losh 		goto done;
489b5253557SWarner Losh 	} else if (isstr(vp)) {
490b5253557SWarner Losh 		if (fmt == OFMT) {
491b5253557SWarner Losh 			if ((vp->tval & CONVC) != 0
492b5253557SWarner Losh 			    || ((vp->tval & CONVO) != 0 && vp->fmt != *fmt)) {
493b5253557SWarner Losh 				update_str_val(vp);
494b5253557SWarner Losh 				vp->tval &= ~CONVC;
495b5253557SWarner Losh 				vp->tval |= CONVO;
496b5253557SWarner Losh 				vp->fmt = *fmt;
497b5253557SWarner Losh 			}
498b5253557SWarner Losh 		} else {
499b5253557SWarner Losh 			/* CONVFMT */
500b5253557SWarner Losh 			if ((vp->tval & CONVO) != 0
501b5253557SWarner Losh 			    || ((vp->tval & CONVC) != 0 && vp->fmt != *fmt)) {
502b5253557SWarner Losh 				update_str_val(vp);
503b5253557SWarner Losh 				vp->tval &= ~CONVO;
504b5253557SWarner Losh 				vp->tval |= CONVC;
505b5253557SWarner Losh 				vp->fmt = *fmt;
506b5253557SWarner Losh 			}
507b5253557SWarner Losh 		}
508b5253557SWarner Losh 	}
509b5253557SWarner Losh done:
510f39dd6a9SWarner Losh 	DPRINTF("getsval %p: %s = \"%s (%p)\", t=%o\n",
511f39dd6a9SWarner Losh 		(void*)vp, NN(vp->nval), vp->sval, (void*)vp->sval, vp->tval);
5122a55deb1SDavid E. O'Brien 	return(vp->sval);
5132a55deb1SDavid E. O'Brien }
5142a55deb1SDavid E. O'Brien 
515813da98dSDavid E. O'Brien char *getsval(Cell *vp)       /* get string val of a Cell */
516813da98dSDavid E. O'Brien {
517813da98dSDavid E. O'Brien       return get_str_val(vp, CONVFMT);
518813da98dSDavid E. O'Brien }
519813da98dSDavid E. O'Brien 
520813da98dSDavid E. O'Brien char *getpssval(Cell *vp)     /* get string val of a Cell for print */
521813da98dSDavid E. O'Brien {
522813da98dSDavid E. O'Brien       return get_str_val(vp, OFMT);
523813da98dSDavid E. O'Brien }
524813da98dSDavid E. O'Brien 
525813da98dSDavid E. O'Brien 
526813da98dSDavid E. O'Brien char *tostring(const char *s)	/* make a copy of string s */
5272a55deb1SDavid E. O'Brien {
528f39dd6a9SWarner Losh 	char *p = strdup(s);
529f39dd6a9SWarner Losh 	if (p == NULL)
530f39dd6a9SWarner Losh 		FATAL("out of space in tostring on %s", s);
531f39dd6a9SWarner Losh 	return(p);
532f39dd6a9SWarner Losh }
533f39dd6a9SWarner Losh 
534f39dd6a9SWarner Losh char *tostringN(const char *s, size_t n)	/* make a copy of string s */
535f39dd6a9SWarner Losh {
5362a55deb1SDavid E. O'Brien 	char *p;
5372a55deb1SDavid E. O'Brien 
538f39dd6a9SWarner Losh 	p = (char *) malloc(n);
5392a55deb1SDavid E. O'Brien 	if (p == NULL)
5402a55deb1SDavid E. O'Brien 		FATAL("out of space in tostring on %s", s);
5412a55deb1SDavid E. O'Brien 	strcpy(p, s);
5422a55deb1SDavid E. O'Brien 	return(p);
5432a55deb1SDavid E. O'Brien }
5442a55deb1SDavid E. O'Brien 
545f39dd6a9SWarner Losh Cell *catstr(Cell *a, Cell *b) /* concatenate a and b */
546f39dd6a9SWarner Losh {
547f39dd6a9SWarner Losh 	Cell *c;
548f39dd6a9SWarner Losh 	char *p;
549f39dd6a9SWarner Losh 	char *sa = getsval(a);
550f39dd6a9SWarner Losh 	char *sb = getsval(b);
551f39dd6a9SWarner Losh 	size_t l = strlen(sa) + strlen(sb) + 1;
552f39dd6a9SWarner Losh 	p = (char *) malloc(l);
553f39dd6a9SWarner Losh 	if (p == NULL)
554f39dd6a9SWarner Losh 		FATAL("out of space concatenating %s and %s", sa, sb);
555f39dd6a9SWarner Losh 	snprintf(p, l, "%s%s", sa, sb);
556f39dd6a9SWarner Losh 
557f39dd6a9SWarner Losh 	l++;	// add room for ' '
558f39dd6a9SWarner Losh 	char *newbuf = (char *) malloc(l);
559f39dd6a9SWarner Losh 	if (newbuf == NULL)
560f39dd6a9SWarner Losh 		FATAL("out of space concatenating %s and %s", sa, sb);
561f39dd6a9SWarner Losh 	// See string() in lex.c; a string "xx" is stored in the symbol
562f39dd6a9SWarner Losh 	// table as "xx ".
563f39dd6a9SWarner Losh 	snprintf(newbuf, l, "%s ", p);
564f39dd6a9SWarner Losh 	c = setsymtab(newbuf, p, 0.0, CON|STR|DONTFREE, symtab);
565f39dd6a9SWarner Losh 	free(p);
566f39dd6a9SWarner Losh 	free(newbuf);
567f39dd6a9SWarner Losh 	return c;
568f39dd6a9SWarner Losh }
569f39dd6a9SWarner Losh 
570813da98dSDavid E. O'Brien char *qstring(const char *is, int delim)	/* collect string up to next delim */
5712a55deb1SDavid E. O'Brien {
5722a55deb1SDavid E. O'Brien 	int c, n;
573f39dd6a9SWarner Losh 	const uschar *s = (const uschar *) is;
5742a55deb1SDavid E. O'Brien 	uschar *buf, *bp;
5752a55deb1SDavid E. O'Brien 
576007c6572SDag-Erling Smørgrav 	if ((buf = (uschar *) malloc(strlen(is)+3)) == NULL)
5772a55deb1SDavid E. O'Brien 		FATAL( "out of space in qstring(%s)", s);
5782a55deb1SDavid E. O'Brien 	for (bp = buf; (c = *s) != delim; s++) {
5792a55deb1SDavid E. O'Brien 		if (c == '\n')
580f32a6403SWarner Losh 			SYNTAX( "newline in string %.20s...", is );
5812a55deb1SDavid E. O'Brien 		else if (c != '\\')
5822a55deb1SDavid E. O'Brien 			*bp++ = c;
5832a55deb1SDavid E. O'Brien 		else {	/* \something */
5842a55deb1SDavid E. O'Brien 			c = *++s;
5852a55deb1SDavid E. O'Brien 			if (c == 0) {	/* \ at end */
5862a55deb1SDavid E. O'Brien 				*bp++ = '\\';
5872a55deb1SDavid E. O'Brien 				break;	/* for loop */
5882a55deb1SDavid E. O'Brien 			}
5892a55deb1SDavid E. O'Brien 			switch (c) {
5902a55deb1SDavid E. O'Brien 			case '\\':	*bp++ = '\\'; break;
5912a55deb1SDavid E. O'Brien 			case 'n':	*bp++ = '\n'; break;
5922a55deb1SDavid E. O'Brien 			case 't':	*bp++ = '\t'; break;
5932a55deb1SDavid E. O'Brien 			case 'b':	*bp++ = '\b'; break;
5942a55deb1SDavid E. O'Brien 			case 'f':	*bp++ = '\f'; break;
5952a55deb1SDavid E. O'Brien 			case 'r':	*bp++ = '\r'; break;
596f39dd6a9SWarner Losh 			case 'v':	*bp++ = '\v'; break;
597f39dd6a9SWarner Losh 			case 'a':	*bp++ = '\a'; break;
5982a55deb1SDavid E. O'Brien 			default:
5992a55deb1SDavid E. O'Brien 				if (!isdigit(c)) {
6002a55deb1SDavid E. O'Brien 					*bp++ = c;
6012a55deb1SDavid E. O'Brien 					break;
6022a55deb1SDavid E. O'Brien 				}
6032a55deb1SDavid E. O'Brien 				n = c - '0';
6042a55deb1SDavid E. O'Brien 				if (isdigit(s[1])) {
6052a55deb1SDavid E. O'Brien 					n = 8 * n + *++s - '0';
6062a55deb1SDavid E. O'Brien 					if (isdigit(s[1]))
6072a55deb1SDavid E. O'Brien 						n = 8 * n + *++s - '0';
6082a55deb1SDavid E. O'Brien 				}
6092a55deb1SDavid E. O'Brien 				*bp++ = n;
6102a55deb1SDavid E. O'Brien 				break;
6112a55deb1SDavid E. O'Brien 			}
6122a55deb1SDavid E. O'Brien 		}
6132a55deb1SDavid E. O'Brien 	}
6142a55deb1SDavid E. O'Brien 	*bp++ = 0;
6152a55deb1SDavid E. O'Brien 	return (char *) buf;
6162a55deb1SDavid E. O'Brien }
617b5253557SWarner Losh 
618b5253557SWarner Losh const char *flags2str(int flags)
619b5253557SWarner Losh {
620b5253557SWarner Losh 	static const struct ftab {
621b5253557SWarner Losh 		const char *name;
622b5253557SWarner Losh 		int value;
623b5253557SWarner Losh 	} flagtab[] = {
624b5253557SWarner Losh 		{ "NUM", NUM },
625b5253557SWarner Losh 		{ "STR", STR },
626b5253557SWarner Losh 		{ "DONTFREE", DONTFREE },
627b5253557SWarner Losh 		{ "CON", CON },
628b5253557SWarner Losh 		{ "ARR", ARR },
629b5253557SWarner Losh 		{ "FCN", FCN },
630b5253557SWarner Losh 		{ "FLD", FLD },
631b5253557SWarner Losh 		{ "REC", REC },
632b5253557SWarner Losh 		{ "CONVC", CONVC },
633b5253557SWarner Losh 		{ "CONVO", CONVO },
634b5253557SWarner Losh 		{ NULL, 0 }
635b5253557SWarner Losh 	};
636b5253557SWarner Losh 	static char buf[100];
637b5253557SWarner Losh 	int i;
638b5253557SWarner Losh 	char *cp = buf;
639b5253557SWarner Losh 
640b5253557SWarner Losh 	for (i = 0; flagtab[i].name != NULL; i++) {
641b5253557SWarner Losh 		if ((flags & flagtab[i].value) != 0) {
642b5253557SWarner Losh 			if (cp > buf)
643b5253557SWarner Losh 				*cp++ = '|';
644b5253557SWarner Losh 			strcpy(cp, flagtab[i].name);
645b5253557SWarner Losh 			cp += strlen(cp);
646b5253557SWarner Losh 		}
647b5253557SWarner Losh 	}
648b5253557SWarner Losh 
649b5253557SWarner Losh 	return buf;
650b5253557SWarner Losh }
651