1 /* $NetBSD: fields.c,v 1.4 2000/10/17 15:13:40 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 /* Subroutines to generate sort keys. */ 40 41 #include "sort.h" 42 43 #ifndef lint 44 __RCSID("$NetBSD: fields.c,v 1.4 2000/10/17 15:13:40 jdolecek Exp $"); 45 __SCCSID("@(#)fields.c 8.1 (Berkeley) 6/6/93"); 46 #endif /* not lint */ 47 48 #define blancmange(ptr) { \ 49 if (BLANK & d_mask[*(ptr)]) \ 50 while (BLANK & d_mask[*(++(ptr))]); \ 51 } 52 53 #define NEXTCOL(pos) { \ 54 if (!SEP_FLAG) \ 55 while (BLANK & l_d_mask[*(++pos)]); \ 56 while (!((FLD_D | REC_D_F) & l_d_mask[*++pos])); \ 57 } 58 59 extern u_char *enterfield __P((u_char *, u_char *, struct field *, int)); 60 61 extern u_char *number __P((u_char *, u_char *, u_char *, u_char *, int)); 62 63 extern struct coldesc clist[(ND+1)*2]; 64 extern int ncols; 65 66 #define DECIMAL '.' 67 #define OFFSET 128 68 69 u_char TENS[10]; /* TENS[0] = REC_D <= 128 ? 130 - '0' : 2 -'0'... */ 70 u_char NEGTENS[10]; /* NEGTENS[0] = REC_D <= 128 ? 126 + '0' : 252 +'0' */ 71 u_char *OFF_TENS, *OFF_NTENS; /* TENS - '0', NEGTENS - '0' */ 72 u_char fnum[NBINS], rnum[NBINS]; 73 74 /* 75 * constructs sort key with leading recheader, followed by the key, 76 * followed by the original line. 77 */ 78 length_t 79 enterkey(keybuf, line, size, fieldtable) 80 struct recheader *keybuf; /* pointer to start of key */ 81 DBT *line; 82 int size; 83 struct field fieldtable[]; 84 { 85 int i; 86 u_char *l_d_mask; 87 u_char *lineend, *pos; 88 u_char *endkey, *keypos; 89 struct coldesc *clpos; 90 int col = 1; 91 struct field *ftpos; 92 l_d_mask = d_mask; 93 pos = (u_char *) line->data - 1; 94 lineend = (u_char *) line->data + line->size-1; 95 /* don't include rec_delimiter */ 96 keypos = keybuf->data; 97 98 for (i = 0, clpos = clist; i < ncols; i++, clpos++) { 99 for (; (col < clpos->num) && (pos < lineend); col++) 100 { NEXTCOL(pos); } 101 if (pos >= lineend) 102 break; 103 clpos->start = SEP_FLAG ? pos + 1 : pos; 104 NEXTCOL(pos); 105 clpos->end = pos; 106 col++; 107 if (pos >= lineend) { 108 clpos->end = lineend; 109 ++i; 110 break; 111 } 112 } 113 for (; i <= ncols; i++) 114 clist[i].start = clist[i].end = lineend; 115 if (clist[0].start < (u_char *) line->data) 116 ++clist[0].start; 117 endkey = (u_char *) keybuf + size - line->size; 118 for (ftpos = fieldtable + 1; ftpos->icol.num; ftpos++) 119 if ((keypos = enterfield(keypos, endkey, ftpos, 120 fieldtable->flags)) == NULL) 121 return (1); 122 123 if (UNIQUE) 124 *(keypos-1) = REC_D; 125 keybuf->offset = keypos - keybuf->data; 126 keybuf->length = keybuf->offset + line->size; 127 if (keybuf->length + sizeof(TRECHEADER) > size) 128 return (1); /* line too long for buffer */ 129 memcpy(keybuf->data + keybuf->offset, line->data, line->size); 130 return (0); 131 } 132 133 /* 134 * constructs a field (as defined by -k) within a key 135 */ 136 u_char * 137 enterfield(tablepos, endkey, cur_fld, gflags) 138 struct field *cur_fld; 139 u_char *tablepos, *endkey; 140 int gflags; 141 { 142 u_char *start, *end, *lineend, *mask, *lweight; 143 struct column icol, tcol; 144 u_int flags; 145 u_int Rflag; 146 icol = cur_fld->icol; 147 tcol = cur_fld->tcol; 148 flags = cur_fld->flags; 149 start = icol.p->start; 150 lineend = clist[ncols].end; 151 if (flags & BI) 152 blancmange(start); 153 start += icol.indent; 154 start = min(start, lineend); 155 if (!tcol.num) 156 end = lineend; 157 else { 158 if (tcol.indent) { 159 end = tcol.p->start; 160 if (flags & BT) blancmange(end); 161 end += tcol.indent; 162 end = min(end, lineend); 163 } else 164 end = tcol.p->end; 165 } 166 if (flags & N) { 167 Rflag = (gflags & R ) ^ (flags & R) ? 1 : 0; 168 tablepos = number(tablepos, endkey, start, end, Rflag); 169 return (tablepos); 170 } 171 mask = alltable; 172 mask = cur_fld->mask; 173 lweight = cur_fld->weights; 174 for (; start < end; start++) 175 if (mask[*start]) { 176 if (*start <= 1) { 177 if (tablepos+2 >= endkey) 178 return (NULL); 179 *tablepos++ = lweight[1]; 180 *tablepos++ = lweight[*start ? 2 : 1]; 181 } else { 182 *tablepos++ = lweight[*start]; 183 if (tablepos == endkey) 184 return (NULL); 185 } 186 } 187 *tablepos++ = lweight[0]; 188 return (tablepos == endkey ? NULL : tablepos); 189 } 190 191 /* Uses the first bin to assign sign, expsign, 0, and the first 192 * 61 out of the exponent ( (254 - 3 origins - 4 over/underflows)/4 = 61 ). 193 * When sorting in forward order: 194 * use (0-99) -> (130->240) for sorting the mantissa if REC_D <=128; 195 * else use (0-99)->(2-102). 196 * If the exponent is >=61, use another byte for each additional 253 197 * in the exponent. Cutoff is at 567. 198 * To avoid confusing the exponent and the mantissa, use a field delimiter 199 * if the exponent is exactly 61, 61+252, etc--this is ok, since it's the 200 * only time a field delimiter can come in that position. 201 * Reverse order is done analagously. 202 */ 203 204 u_char * 205 number(pos, bufend, line, lineend, Rflag) 206 u_char *line, *pos, *bufend, *lineend; 207 int Rflag; 208 { 209 int or_sign, parity = 0; 210 int expincr = 1, exponent = -1; 211 int bite, expsign = 1, sign = 1; 212 u_char lastvalue, *nonzero, *tline, *C_TENS; 213 u_char *nweights; 214 215 if (Rflag) 216 nweights = rnum; 217 else 218 nweights = fnum; 219 if (pos > bufend - 8) 220 return (NULL); 221 /* or_sign sets the sort direction: 222 * (-r: +/-)(sign: +/-)(expsign: +/-) */ 223 or_sign = sign ^ expsign ^ Rflag; 224 blancmange(line); 225 if (*line == '-') { /* set the sign */ 226 or_sign ^= 1; 227 sign = 0; 228 line++; 229 } 230 /* eat initial zeroes */ 231 for (; *line == '0' && line < lineend; line++); 232 /* calculate exponents < 0 */ 233 if (*line == DECIMAL) { 234 exponent = 1; 235 while (*++line == '0' && line < lineend) 236 exponent++; 237 expincr = 0; 238 expsign = 0; 239 } 240 /* next character better be a digit */ 241 if (*line < '1' || *line > '9' || line >= lineend) { 242 *pos++ = nweights[127]; 243 return (pos); 244 } 245 if (expincr) { 246 for (tline = line-1; *++tline >= '0' && 247 *tline <= '9' && tline < lineend;) 248 exponent++; 249 } 250 if (exponent > 567) { 251 *pos++ = nweights[sign ? (expsign ? 254 : 128) 252 : (expsign ? 0 : 126)]; 253 warnx("exponent out of bounds"); 254 return (pos); 255 } 256 bite = min(exponent, 61); 257 *pos++ = nweights[(sign) ? (expsign ? 189+bite : 189-bite) 258 : (expsign ? 64-bite : 64+bite)]; 259 if (bite >= 61) { 260 do { 261 exponent -= bite; 262 bite = min(exponent, 254); 263 *pos++ = nweights[or_sign ? 254-bite : bite]; 264 } while (bite == 254); 265 } 266 C_TENS = or_sign ? OFF_NTENS : OFF_TENS; 267 for (; line < lineend; line++) { 268 if (*line >= '0' && *line <= '9') { 269 if (parity) { 270 *pos++ = C_TENS[lastvalue] + (or_sign ? - *line 271 : *line); 272 if (pos == bufend) 273 return (NULL); 274 if (*line != '0' || lastvalue != '0') 275 nonzero = pos; 276 } else 277 lastvalue = *line; 278 parity ^= 1; 279 } else if(*line == DECIMAL) { 280 if(!expincr) /* a decimal already occurred once */ 281 break; 282 expincr = 0; 283 } else 284 break; 285 } 286 if (parity && lastvalue != '0') { 287 *pos++ = or_sign ? OFF_NTENS[lastvalue] - '0' : 288 OFF_TENS[lastvalue] + '0'; 289 } else 290 pos = nonzero; 291 if (pos > bufend-1) 292 return (NULL); 293 *pos++ = or_sign ? nweights[254] : nweights[0]; 294 return (pos); 295 } 296 297 /* This forces a gap around the record delimiter 298 * Thus fnum has vaues over (0,254) -> ((0,REC_D-1),(REC_D+1,255)); 299 * rnum over (0,254) -> (255,REC_D+1),(REC_D-1,0)) 300 */ 301 void 302 num_init() 303 { 304 int i; 305 TENS[0] = REC_D <=128 ? 130 - '0' : 2 - '0'; 306 NEGTENS[0] = REC_D <=128 ? 126 + '0' : 254 + '0'; 307 OFF_TENS = TENS - '0'; 308 OFF_NTENS = NEGTENS - '0'; 309 for (i = 1; i < 10; i++) { 310 TENS[i] = TENS[i-1] + 10; 311 NEGTENS[i] = NEGTENS[i-1] - 10; 312 } 313 for (i = 0; i < REC_D; i++) { 314 fnum[i] = i; 315 rnum[255-i] = i; 316 } 317 for (i = REC_D; i <255; i++) { 318 fnum[i] = i+1; 319 rnum[255-i] = i-1; 320 } 321 } 322