1 /* 2 * Copyright (c) 1989, 1993, 1994 3 * The Regents of the University of California. All rights reserved. 4 * 5 * This code is derived from software contributed to Berkeley by 6 * Michael Rendell of Memorial University of Newfoundland. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 3. All advertising materials mentioning features or use of this software 17 * must display the following acknowledgement: 18 * This product includes software developed by the University of 19 * California, Berkeley and its contributors. 20 * 4. Neither the name of the University nor the names of its contributors 21 * may be used to endorse or promote products derived from this software 22 * without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 34 * SUCH DAMAGE. 35 * 36 * $FreeBSD: src/usr.bin/tsort/tsort.c,v 1.10.2.1 2001/03/04 09:18:23 kris Exp $ 37 * $DragonFly: src/usr.bin/tsort/tsort.c,v 1.3 2003/10/04 20:36:53 hmp Exp $ 38 * 39 * @(#) Copyright (c) 1989, 1993, 1994 The Regents of the University of California. All rights reserved. 40 * @(#)tsort.c 8.3 (Berkeley) 5/4/95 41 */ 42 43 #include <sys/types.h> 44 45 #include <ctype.h> 46 #include <db.h> 47 #include <err.h> 48 #include <errno.h> 49 #include <fcntl.h> 50 #include <stdio.h> 51 #include <stdlib.h> 52 #include <string.h> 53 #include <unistd.h> 54 55 /* 56 * Topological sort. Input is a list of pairs of strings separated by 57 * white space (spaces, tabs, and/or newlines); strings are written to 58 * standard output in sorted order, one per line. 59 * 60 * usage: 61 * tsort [-dlq] [inputfile] 62 * If no input file is specified, standard input is read. 63 * 64 * Should be compatible with AT&T tsort HOWEVER the output is not identical 65 * (i.e. for most graphs there is more than one sorted order, and this tsort 66 * usually generates a different one then the AT&T tsort). Also, cycle 67 * reporting seems to be more accurate in this version (the AT&T tsort 68 * sometimes says a node is in a cycle when it isn't). 69 * 70 * Michael Rendell, michael@stretch.cs.mun.ca - Feb 26, '90 71 */ 72 #define HASHSIZE 53 /* doesn't need to be big */ 73 #define NF_MARK 0x1 /* marker for cycle detection */ 74 #define NF_ACYCLIC 0x2 /* this node is cycle free */ 75 #define NF_NODEST 0x4 /* Unreachable */ 76 77 78 typedef struct node_str NODE; 79 80 struct node_str { 81 NODE **n_prevp; /* pointer to previous node's n_next */ 82 NODE *n_next; /* next node in graph */ 83 NODE **n_arcs; /* array of arcs to other nodes */ 84 int n_narcs; /* number of arcs in n_arcs[] */ 85 int n_arcsize; /* size of n_arcs[] array */ 86 int n_refcnt; /* # of arcs pointing to this node */ 87 int n_flags; /* NF_* */ 88 char n_name[1]; /* name of this node */ 89 }; 90 91 typedef struct _buf { 92 char *b_buf; 93 int b_bsize; 94 } BUF; 95 96 DB *db; 97 NODE *graph, **cycle_buf, **longest_cycle; 98 int debug, longest, quiet; 99 100 void add_arc(char *, char *); 101 int find_cycle(NODE *, NODE *, int, int); 102 NODE *get_node(char *); 103 void *grow_buf(void *, int); 104 void remove_node(NODE *); 105 void tsort(void); 106 void usage(void); 107 108 int 109 main(int argc, char **argv) 110 { 111 register BUF *b; 112 register int c, n; 113 FILE *fp; 114 int bsize, ch, nused; 115 BUF bufs[2]; 116 117 while ((ch = getopt(argc, argv, "dlq")) != -1) 118 switch (ch) { 119 case 'd': 120 debug = 1; 121 break; 122 case 'l': 123 longest = 1; 124 break; 125 case 'q': 126 quiet = 1; 127 break; 128 case '?': 129 default: 130 usage(); 131 } 132 argc -= optind; 133 argv += optind; 134 135 switch (argc) { 136 case 0: 137 fp = stdin; 138 break; 139 case 1: 140 if ((fp = fopen(*argv, "r")) == NULL) 141 err(1, "%s", *argv); 142 break; 143 default: 144 usage(); 145 } 146 147 for (b = bufs, n = 2; --n >= 0; b++) 148 b->b_buf = grow_buf(NULL, b->b_bsize = 1024); 149 150 /* parse input and build the graph */ 151 for (n = 0, c = getc(fp);;) { 152 while (c != EOF && isspace(c)) 153 c = getc(fp); 154 if (c == EOF) 155 break; 156 157 nused = 0; 158 b = &bufs[n]; 159 bsize = b->b_bsize; 160 do { 161 b->b_buf[nused++] = c; 162 if (nused == bsize) 163 b->b_buf = grow_buf(b->b_buf, bsize *= 2); 164 c = getc(fp); 165 } while (c != EOF && !isspace(c)); 166 167 b->b_buf[nused] = '\0'; 168 b->b_bsize = bsize; 169 if (n) 170 add_arc(bufs[0].b_buf, bufs[1].b_buf); 171 n = !n; 172 } 173 (void)fclose(fp); 174 if (n) 175 errx(1, "odd data count"); 176 177 /* do the sort */ 178 tsort(); 179 exit(0); 180 } 181 182 /* double the size of oldbuf and return a pointer to the new buffer. */ 183 void * 184 grow_buf(void *bp, int size) 185 { 186 if ((bp = realloc(bp, (u_int)size)) == NULL) 187 err(1, NULL); 188 return (bp); 189 } 190 191 /* 192 * add an arc from node s1 to node s2 in the graph. If s1 or s2 are not in 193 * the graph, then add them. 194 */ 195 void 196 add_arc(char *s1, char *s2) 197 { 198 register NODE *n1; 199 NODE *n2; 200 int bsize, i; 201 202 n1 = get_node(s1); 203 204 if (!strcmp(s1, s2)) 205 return; 206 207 n2 = get_node(s2); 208 209 /* 210 * Check if this arc is already here. 211 */ 212 for (i = 0; i < n1->n_narcs; i++) 213 if (n1->n_arcs[i] == n2) 214 return; 215 /* 216 * Add it. 217 */ 218 if (n1->n_narcs == n1->n_arcsize) { 219 if (!n1->n_arcsize) 220 n1->n_arcsize = 10; 221 bsize = n1->n_arcsize * sizeof(*n1->n_arcs) * 2; 222 n1->n_arcs = grow_buf(n1->n_arcs, bsize); 223 n1->n_arcsize = bsize / sizeof(*n1->n_arcs); 224 } 225 n1->n_arcs[n1->n_narcs++] = n2; 226 ++n2->n_refcnt; 227 } 228 229 /* Find a node in the graph (insert if not found) and return a pointer to it. */ 230 NODE * 231 get_node(char *name) 232 { 233 DBT data, key; 234 NODE *n; 235 236 if (db == NULL && 237 (db = dbopen(NULL, O_RDWR, 0, DB_HASH, NULL)) == NULL) 238 err(1, "db: %s", name); 239 240 key.data = name; 241 key.size = strlen(name) + 1; 242 243 switch ((*db->get)(db, &key, &data, 0)) { 244 case 0: 245 bcopy(data.data, &n, sizeof(n)); 246 return (n); 247 case 1: 248 break; 249 default: 250 case -1: 251 err(1, "db: %s", name); 252 } 253 254 if ((n = malloc(sizeof(NODE) + key.size)) == NULL) 255 err(1, NULL); 256 257 n->n_narcs = 0; 258 n->n_arcsize = 0; 259 n->n_arcs = NULL; 260 n->n_refcnt = 0; 261 n->n_flags = 0; 262 bcopy(name, n->n_name, key.size); 263 264 /* Add to linked list. */ 265 if ((n->n_next = graph) != NULL) 266 graph->n_prevp = &n->n_next; 267 n->n_prevp = &graph; 268 graph = n; 269 270 /* Add to hash table. */ 271 data.data = &n; 272 data.size = sizeof(n); 273 if ((*db->put)(db, &key, &data, 0)) 274 err(1, "db: %s", name); 275 return (n); 276 } 277 278 279 /* 280 * Clear the NODEST flag from all nodes. 281 */ 282 void 283 clear_cycle(void) 284 { 285 NODE *n; 286 287 for (n = graph; n != NULL; n = n->n_next) 288 n->n_flags &= ~NF_NODEST; 289 } 290 291 /* do topological sort on graph */ 292 void 293 tsort(void) 294 { 295 register NODE *n, *next; 296 register int cnt, i; 297 298 while (graph != NULL) { 299 /* 300 * Keep getting rid of simple cases until there are none left, 301 * if there are any nodes still in the graph, then there is 302 * a cycle in it. 303 */ 304 do { 305 for (cnt = 0, n = graph; n != NULL; n = next) { 306 next = n->n_next; 307 if (n->n_refcnt == 0) { 308 remove_node(n); 309 ++cnt; 310 } 311 } 312 } while (graph != NULL && cnt); 313 314 if (graph == NULL) 315 break; 316 317 if (!cycle_buf) { 318 /* 319 * Allocate space for two cycle logs - one to be used 320 * as scratch space, the other to save the longest 321 * cycle. 322 */ 323 for (cnt = 0, n = graph; n != NULL; n = n->n_next) 324 ++cnt; 325 cycle_buf = malloc((u_int)sizeof(NODE *) * cnt); 326 longest_cycle = malloc((u_int)sizeof(NODE *) * cnt); 327 if (cycle_buf == NULL || longest_cycle == NULL) 328 err(1, NULL); 329 } 330 for (n = graph; n != NULL; n = n->n_next) 331 if (!(n->n_flags & NF_ACYCLIC)) { 332 if ((cnt = find_cycle(n, n, 0, 0))) { 333 if (!quiet) { 334 warnx("cycle in data"); 335 for (i = 0; i < cnt; i++) 336 warnx("%s", 337 longest_cycle[i]->n_name); 338 } 339 remove_node(n); 340 clear_cycle(); 341 break; 342 } else { 343 /* to avoid further checks */ 344 n->n_flags |= NF_ACYCLIC; 345 clear_cycle(); 346 } 347 } 348 349 if (n == NULL) 350 errx(1, "internal error -- could not find cycle"); 351 } 352 } 353 354 /* print node and remove from graph (does not actually free node) */ 355 void 356 remove_node(register NODE *n) 357 { 358 register NODE **np; 359 register int i; 360 361 (void)printf("%s\n", n->n_name); 362 for (np = n->n_arcs, i = n->n_narcs; --i >= 0; np++) 363 --(*np)->n_refcnt; 364 n->n_narcs = 0; 365 *n->n_prevp = n->n_next; 366 if (n->n_next) 367 n->n_next->n_prevp = n->n_prevp; 368 } 369 370 371 /* look for the longest? cycle from node from to node to. */ 372 int 373 find_cycle(NODE *from, NODE *to, int longest_len, int depth) 374 { 375 register NODE **np; 376 register int i, len; 377 378 /* 379 * avoid infinite loops and ignore portions of the graph known 380 * to be acyclic 381 */ 382 if (from->n_flags & (NF_NODEST|NF_MARK|NF_ACYCLIC)) 383 return (0); 384 from->n_flags |= NF_MARK; 385 386 for (np = from->n_arcs, i = from->n_narcs; --i >= 0; np++) { 387 cycle_buf[depth] = *np; 388 if (*np == to) { 389 if (depth + 1 > longest_len) { 390 longest_len = depth + 1; 391 (void)memcpy((char *)longest_cycle, 392 (char *)cycle_buf, 393 longest_len * sizeof(NODE *)); 394 } 395 } else { 396 if ((*np)->n_flags & (NF_MARK|NF_ACYCLIC|NF_NODEST)) 397 continue; 398 len = find_cycle(*np, to, longest_len, depth + 1); 399 400 if (debug) 401 (void)printf("%*s %s->%s %d\n", depth, "", 402 from->n_name, to->n_name, len); 403 404 if (len == 0) 405 (*np)->n_flags |= NF_NODEST; 406 407 if (len > longest_len) 408 longest_len = len; 409 410 if (len > 0 && !longest) 411 break; 412 } 413 } 414 from->n_flags &= ~NF_MARK; 415 return (longest_len); 416 } 417 418 void 419 usage(void) 420 { 421 (void)fprintf(stderr, "usage: tsort [-dlq] [file]\n"); 422 exit(1); 423 } 424