2 * Copyright (c) 1989, 1993, 1994
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed to Berkeley by
6 * Michael Rendell of Memorial University of Newfoundland.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
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.
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
38 static const char copyright[] =
39 "@(#) Copyright (c) 1989, 1993, 1994\n\
40 The Regents of the University of California. All rights reserved.\n";
44 static const char sccsid[] = "@(#)tsort.c 8.3 (Berkeley) 5/4/95";
47 #include <sys/cdefs.h>
48 __FBSDID("$FreeBSD$");
50 #include <sys/types.h>
63 * Topological sort. Input is a list of pairs of strings separated by
64 * white space (spaces, tabs, and/or newlines); strings are written to
65 * standard output in sorted order, one per line.
68 * tsort [-dlq] [inputfile]
69 * If no input file is specified, standard input is read.
71 * Should be compatible with AT&T tsort HOWEVER the output is not identical
72 * (i.e. for most graphs there is more than one sorted order, and this tsort
73 * usually generates a different one then the AT&T tsort). Also, cycle
74 * reporting seems to be more accurate in this version (the AT&T tsort
75 * sometimes says a node is in a cycle when it isn't).
77 * Michael Rendell, michael@stretch.cs.mun.ca - Feb 26, '90
80 #define NF_MARK 0x1 /* marker for cycle detection */
81 #define NF_ACYCLIC 0x2 /* this node is cycle free */
82 #define NF_NODEST 0x4 /* Unreachable */
85 typedef struct node_str NODE;
88 NODE **n_prevp; /* pointer to previous node's n_next */
89 NODE *n_next; /* next node in graph */
90 NODE **n_arcs; /* array of arcs to other nodes */
91 int n_narcs; /* number of arcs in n_arcs[] */
92 int n_arcsize; /* size of n_arcs[] array */
93 int n_refcnt; /* # of arcs pointing to this node */
94 int n_flags; /* NF_* */
95 char n_name[1]; /* name of this node */
104 NODE *graph, **cycle_buf, **longest_cycle;
105 int debug, longest, quiet;
107 void add_arc(char *, char *);
108 int find_cycle(NODE *, NODE *, int, int);
109 NODE *get_node(char *);
110 void *grow_buf(void *, size_t);
111 void remove_node(NODE *);
112 void clear_cycle(void);
117 main(int argc, char *argv[])
122 int bsize, ch, nused;
126 while ((ch = getopt(argc, argv, "dlq")) != -1)
149 if ((fp = fopen(*argv, "r")) == NULL)
156 for (b = bufs, n = 2; --n >= 0; b++)
157 b->b_buf = grow_buf(NULL, b->b_bsize = 1024);
159 /* parse input and build the graph */
160 for (n = 0, c = getc(fp);;) {
161 while (c != EOF && isspace(c))
170 b->b_buf[nused++] = c;
172 b->b_buf = grow_buf(b->b_buf, bsize *= 2);
174 } while (c != EOF && !isspace(c));
176 b->b_buf[nused] = '\0';
179 add_arc(bufs[0].b_buf, bufs[1].b_buf);
184 errx(1, "odd data count");
191 /* double the size of oldbuf and return a pointer to the new buffer. */
193 grow_buf(void *bp, size_t size)
195 if ((bp = realloc(bp, size)) == NULL)
201 * add an arc from node s1 to node s2 in the graph. If s1 or s2 are not in
202 * the graph, then add them.
205 add_arc(char *s1, char *s2)
219 * Check if this arc is already here.
221 for (i = 0; i < n1->n_narcs; i++)
222 if (n1->n_arcs[i] == n2)
227 if (n1->n_narcs == n1->n_arcsize) {
230 bsize = n1->n_arcsize * sizeof(*n1->n_arcs) * 2;
231 n1->n_arcs = grow_buf(n1->n_arcs, bsize);
232 n1->n_arcsize = bsize / sizeof(*n1->n_arcs);
234 n1->n_arcs[n1->n_narcs++] = n2;
238 /* Find a node in the graph (insert if not found) and return a pointer to it. */
246 (db = dbopen(NULL, O_RDWR, 0, DB_HASH, NULL)) == NULL)
247 err(1, "db: %s", name);
250 key.size = strlen(name) + 1;
252 switch ((*db->get)(db, &key, &data, 0)) {
254 bcopy(data.data, &n, sizeof(n));
260 err(1, "db: %s", name);
263 if ((n = malloc(sizeof(NODE) + key.size)) == NULL)
271 bcopy(name, n->n_name, key.size);
273 /* Add to linked list. */
274 if ((n->n_next = graph) != NULL)
275 graph->n_prevp = &n->n_next;
279 /* Add to hash table. */
281 data.size = sizeof(n);
282 if ((*db->put)(db, &key, &data, 0))
283 err(1, "db: %s", name);
289 * Clear the NODEST flag from all nodes.
296 for (n = graph; n != NULL; n = n->n_next)
297 n->n_flags &= ~NF_NODEST;
300 /* do topological sort on graph */
307 while (graph != NULL) {
309 * Keep getting rid of simple cases until there are none left,
310 * if there are any nodes still in the graph, then there is
314 for (cnt = 0, n = graph; n != NULL; n = next) {
316 if (n->n_refcnt == 0) {
321 } while (graph != NULL && cnt);
328 * Allocate space for two cycle logs - one to be used
329 * as scratch space, the other to save the longest
332 for (cnt = 0, n = graph; n != NULL; n = n->n_next)
334 cycle_buf = malloc(sizeof(NODE *) * cnt);
335 longest_cycle = malloc(sizeof(NODE *) * cnt);
336 if (cycle_buf == NULL || longest_cycle == NULL)
339 for (n = graph; n != NULL; n = n->n_next)
340 if (!(n->n_flags & NF_ACYCLIC)) {
341 if ((cnt = find_cycle(n, n, 0, 0))) {
343 warnx("cycle in data");
344 for (i = 0; i < cnt; i++)
346 longest_cycle[i]->n_name);
352 /* to avoid further checks */
353 n->n_flags |= NF_ACYCLIC;
359 errx(1, "internal error -- could not find cycle");
363 /* print node and remove from graph (does not actually free node) */
370 (void)printf("%s\n", n->n_name);
371 for (np = n->n_arcs, i = n->n_narcs; --i >= 0; np++)
374 *n->n_prevp = n->n_next;
376 n->n_next->n_prevp = n->n_prevp;
380 /* look for the longest? cycle from node from to node to. */
382 find_cycle(NODE *from, NODE *to, int longest_len, int depth)
388 * avoid infinite loops and ignore portions of the graph known
391 if (from->n_flags & (NF_NODEST|NF_MARK|NF_ACYCLIC))
393 from->n_flags |= NF_MARK;
395 for (np = from->n_arcs, i = from->n_narcs; --i >= 0; np++) {
396 cycle_buf[depth] = *np;
398 if (depth + 1 > longest_len) {
399 longest_len = depth + 1;
400 (void)memcpy((char *)longest_cycle,
402 longest_len * sizeof(NODE *));
405 if ((*np)->n_flags & (NF_MARK|NF_ACYCLIC|NF_NODEST))
407 len = find_cycle(*np, to, longest_len, depth + 1);
410 (void)printf("%*s %s->%s %d\n", depth, "",
411 from->n_name, to->n_name, len);
414 (*np)->n_flags |= NF_NODEST;
416 if (len > longest_len)
419 if (len > 0 && !longest)
423 from->n_flags &= ~NF_MARK;
424 return (longest_len);
430 (void)fprintf(stderr, "usage: tsort [-dlq] [file]\n");