Home | History | Annotate | Line # | Download | only in sort
init.c revision 1.3
      1 /*	$NetBSD: init.c,v 1.3 2000/10/16 21:41:05 jdolecek Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * This code is derived from software contributed to Berkeley by
      8  * Peter McIlroy.
      9  *
     10  * Redistribution and use in source and binary forms, with or without
     11  * modification, are permitted provided that the following conditions
     12  * are met:
     13  * 1. Redistributions of source code must retain the above copyright
     14  *    notice, this list of conditions and the following disclaimer.
     15  * 2. Redistributions in binary form must reproduce the above copyright
     16  *    notice, this list of conditions and the following disclaimer in the
     17  *    documentation and/or other materials provided with the distribution.
     18  * 3. All advertising materials mentioning features or use of this software
     19  *    must display the following acknowledgement:
     20  *	This product includes software developed by the University of
     21  *	California, Berkeley and its contributors.
     22  * 4. Neither the name of the University nor the names of its contributors
     23  *    may be used to endorse or promote products derived from this software
     24  *    without specific prior written permission.
     25  *
     26  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     27  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     28  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     29  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     30  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     31  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     32  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     33  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     34  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     35  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     36  * SUCH DAMAGE.
     37  */
     38 
     39 #include "sort.h"
     40 
     41 #ifndef lint
     42 __RCSID("$NetBSD: init.c,v 1.3 2000/10/16 21:41:05 jdolecek Exp $");
     43 __SCCSID("@(#)init.c	8.1 (Berkeley) 6/6/93");
     44 #endif /* not lint */
     45 
     46 #include <ctype.h>
     47 #include <string.h>
     48 
     49 static void insertcol __P((struct field *));
     50 static const char *setcolumn __P((const char *, struct field *, int));
     51 int setfield __P((const char *, struct field *, int));
     52 
     53 extern struct coldesc clist[(ND+1)*2];
     54 extern int ncols;
     55 u_char gweights[NBINS];
     56 
     57 /*
     58  * clist (list of columns which correspond to one or more icol or tcol)
     59  * is in increasing order of columns.
     60  * Fields are kept in increasing order of fields.
     61  */
     62 
     63 /*
     64  * keep clist in order--inserts a column in a sorted array
     65  */
     66 static void
     67 insertcol(field)
     68 	struct field *field;
     69 {
     70 	int i;
     71 	for (i = 0; i < ncols; i++)
     72 		if (field->icol.num <= clist[i].num)
     73 			break;
     74 	if (field->icol.num != clist[i].num) {
     75 		memmove(clist+i+1, clist+i, sizeof(COLDESC)*(ncols-i));
     76 		clist[i].num = field->icol.num;
     77 		ncols++;
     78 	}
     79 	if (field->tcol.num && field->tcol.num != field->icol.num) {
     80 		for (i = 0; i < ncols; i++)
     81 			if (field->tcol.num <= clist[i].num)
     82 				break;
     83 		if (field->tcol.num != clist[i].num) {
     84 			memmove(clist+i+1, clist+i,sizeof(COLDESC)*(ncols-i));
     85 			clist[i].num = field->tcol.num;
     86 			ncols++;
     87 		}
     88 	}
     89 }
     90 
     91 /*
     92  * matches fields with the appropriate columns--n^2 but who cares?
     93  */
     94 void
     95 fldreset(fldtab)
     96 	struct field *fldtab;
     97 {
     98 	int i;
     99 	fldtab[0].tcol.p = clist+ncols-1;
    100 	for (++fldtab; fldtab->icol.num; ++fldtab) {
    101 		for (i = 0; fldtab->icol.num != clist[i].num; i++);
    102 		fldtab->icol.p = clist + i;
    103 		if (!fldtab->tcol.num)
    104 			continue;
    105 		for (i = 0; fldtab->tcol.num != clist[i].num; i++);
    106 		fldtab->tcol.p = clist + i;
    107 	}
    108 }
    109 
    110 /*
    111  * interprets a column in a -k field
    112  */
    113 static const char *
    114 setcolumn(pos, cur_fld, gflag)
    115 	const char *pos;
    116 	struct field *cur_fld;
    117 	int gflag;
    118 {
    119 	struct column *col;
    120 	int tmp;
    121 	col = cur_fld->icol.num ? (&(*cur_fld).tcol) : (&(*cur_fld).icol);
    122 	pos += sscanf(pos, "%d", &(col->num));
    123 	while (isdigit(*pos))
    124 		pos++;
    125 	if (col->num <= 0 && !(col->num == 0 && col == &(cur_fld->tcol)))
    126 		errx(2, "field numbers must be positive");
    127 	if (*pos == '.') {
    128 		if (!col->num)
    129 			errx(2, "cannot indent end of line");
    130 		pos += sscanf(++pos, "%d", &(col->indent));
    131 		while (isdigit(*pos))
    132 			pos++;
    133 		if (&cur_fld->icol == col)
    134 			col->indent--;
    135 		if (col->indent < 0)
    136 			errx(2, "illegal offset");
    137 	}
    138 	if (optval(*pos, cur_fld->tcol.num))
    139 		while ((tmp = optval(*pos, cur_fld->tcol.num))) {
    140 			cur_fld->flags |= tmp;
    141 			pos++;
    142 	}
    143 	if (cur_fld->icol.num == 0)
    144 		cur_fld->icol.num = 1;
    145 	return (pos);
    146 }
    147 
    148 int
    149 setfield(pos, cur_fld, gflag)
    150 	const char *pos;
    151 	struct field *cur_fld;
    152 	int gflag;
    153 {
    154 	static int nfields = 0;
    155 	int tmp;
    156 	if (++nfields == ND)
    157 		errx(2, "too many sort keys. (Limit is %d)", ND-1);
    158 	cur_fld->weights = ascii;
    159 	cur_fld->mask = alltable;
    160 	pos = setcolumn(pos, cur_fld, gflag);
    161 	if (*pos == '\0')			/* key extends to EOL. */
    162 		cur_fld->tcol.num = 0;
    163 	else {
    164 		if (*pos != ',')
    165 			errx(2, "illegal field descriptor");
    166 		setcolumn((++pos), cur_fld, gflag);
    167 	}
    168 	if (!cur_fld->flags)
    169 		cur_fld->flags = gflag;
    170 	tmp = cur_fld->flags;
    171 
    172 	/*
    173 	 * Assign appropriate mask table and weight table.
    174 	 * If the global weights are reversed, the local field
    175 	 * must be "re-reversed".
    176 	 */
    177 	if (((tmp & R) ^ (gflag & R)) && tmp & F)
    178 		cur_fld->weights = RFtable;
    179 	else if (tmp & F)
    180 		cur_fld->weights = Ftable;
    181 	else if ((tmp & R) ^ (gflag & R))
    182 		cur_fld->weights = Rascii;
    183 	if (tmp & I)
    184 		cur_fld->mask = itable;
    185 	else if (tmp & D)
    186 		cur_fld->mask = dtable;
    187 	cur_fld->flags |= (gflag & (BI | BT));
    188 	if (!cur_fld->tcol.indent)	/* BT has no meaning at end of field */
    189 		cur_fld->flags &= (D|F|I|N|R|BI);
    190 	if (cur_fld->tcol.num && !(!(cur_fld->flags & BI)
    191 	    && cur_fld->flags & BT) && (cur_fld->tcol.num <= cur_fld->icol.num
    192 	    && cur_fld->tcol.indent < cur_fld->icol.indent))
    193 		errx(2, "fields out of order");
    194 	insertcol(cur_fld);
    195 	return (cur_fld->tcol.num);
    196 }
    197 
    198 int
    199 optval(desc, tcolflag)
    200 	int desc, tcolflag;
    201 {
    202 	switch(desc) {
    203 		case 'b':
    204 			if (!tcolflag)
    205 				return(BI);
    206 			else
    207 				return(BT);
    208 		case 'd': return(D);
    209 		case 'f': return(F);
    210 		case 'i': return(I);
    211 		case 'n': return(N);
    212 		case 'r': return(R);
    213 		default:  return(0);
    214 	}
    215 }
    216 
    217 void
    218 fixit(argc, argv)
    219 	int *argc;
    220 	char **argv;
    221 {
    222 	int i, j, v, w, x;
    223 	static char vbuf[ND*20], *vpos, *tpos;
    224 	vpos = vbuf;
    225 
    226 	for (i = 1; i < *argc; i++) {
    227 		if (argv[i][0] == '+') {
    228 			tpos = argv[i]+1;
    229 			argv[i] = vpos;
    230 			vpos += sprintf(vpos, "-k");
    231 			tpos += sscanf(tpos, "%d", &v);
    232 			while (isdigit(*tpos))
    233 				tpos++;
    234 			vpos += sprintf(vpos, "%d", v+1);
    235 			if (*tpos == '.') {
    236 				tpos += sscanf(++tpos, "%d", &x);
    237 				vpos += sprintf(vpos, ".%d", x+1);
    238 			}
    239 			while (*tpos)
    240 				*vpos++ = *tpos++;
    241 			vpos += sprintf(vpos, ",");
    242 			if (argv[i+1] &&
    243 			    argv[i+1][0] == '-' && isdigit(argv[i+1][1])) {
    244 				tpos = argv[i+1] + 1;
    245 				tpos += sscanf(tpos, "%d", &w);
    246 				while (isdigit(*tpos))
    247 					tpos++;
    248 				x = 0;
    249 				if (*tpos == '.') {
    250 					tpos += sscanf(++tpos, "%d", &x);
    251 					while (isdigit(*tpos))
    252 						tpos++;
    253 				}
    254 				if (x) {
    255 					vpos += sprintf(vpos, "%d", w+1);
    256 					vpos += sprintf(vpos, ".%d", x);
    257 				} else
    258 					vpos += sprintf(vpos, "%d", w);
    259 				while (*tpos)
    260 					*vpos++ = *tpos++;
    261 				for (j= i+1; j < *argc; j++)
    262 					argv[j] = argv[j+1];
    263 				*argc -= 1;
    264 			}
    265 		}
    266 	}
    267 }
    268 
    269 /*
    270  * ascii, Rascii, Ftable, and RFtable map
    271  * REC_D -> REC_D;  {not REC_D} -> {not REC_D}.
    272  * gweights maps REC_D -> (0 or 255); {not REC_D} -> {not gweights[REC_D]}.
    273  * Note: when sorting in forward order, to encode character zero in a key,
    274  * use \001\001; character 1 becomes \001\002.  In this case, character 0
    275  * is reserved for the field delimiter.  Analagously for -r (fld_d = 255).
    276  * Note: this is only good for ASCII sorting.  For different LC 's,
    277  * all bets are off.  See also num_init in number.c
    278  */
    279 void
    280 settables(gflags)
    281 	int gflags;
    282 {
    283 	u_char *wts;
    284 	int i, incr;
    285 	for (i=0; i < 256; i++) {
    286 		ascii[i] = i;
    287 		if (i > REC_D && i < 255 - REC_D+1)
    288 			Rascii[i] = 255 - i + 1;
    289 		else
    290 			Rascii[i] = 255 - i;
    291 		if (islower(i)) {
    292 			Ftable[i] = Ftable[i- ('a' -'A')];
    293 			RFtable[i] = RFtable[i - ('a' - 'A')];
    294 		} else if (REC_D>= 'A' && REC_D < 'Z' && i < 'a' && i > REC_D) {
    295 			Ftable[i] = i + 1;
    296 			RFtable[i] = Rascii[i] - 1;
    297 		} else {
    298 			Ftable[i] = i;
    299 			RFtable[i] = Rascii[i];
    300 		}
    301 		alltable[i] = 1;
    302 		if (i == '\n' || isprint(i))
    303 			itable[i] = 1;
    304 		else itable[i] = 0;
    305 		if (i == '\n' || i == '\t' || i == ' ' || isalnum(i))
    306 			dtable[i] = 1;
    307 		else dtable[i] = 0;
    308 	}
    309 	Rascii[REC_D] = RFtable[REC_D] = REC_D;
    310 	if (REC_D >= 'A' && REC_D < 'Z')
    311 		++Ftable[REC_D + ('a' - 'A')];
    312 	if (gflags & R && (!(gflags & F) || !SINGL_FLD))
    313 		wts = Rascii;
    314 	else if (!(gflags & F) || !SINGL_FLD)
    315 		wts = ascii;
    316 	else if (gflags & R)
    317 		wts = RFtable;
    318 	else
    319 		wts = Ftable;
    320 	memmove(gweights, wts, sizeof(gweights));
    321 	incr = (gflags & R) ? -1 : 1;
    322 	for (i = 0; i < REC_D; i++)
    323 		gweights[i] += incr;
    324 	gweights[REC_D] = ((gflags & R) ? 255 : 0);
    325 	if (SINGL_FLD && gflags & F) {
    326 		for (i = 0; i < REC_D; i++) {
    327 			ascii[i] += incr;
    328 			Rascii[i] += incr;
    329 		}
    330 		ascii[REC_D] = Rascii[REC_D] = gweights[REC_D];
    331 	}
    332 }
    333