]> git.cameronkatri.com Git - apple_cmds.git/blob - misc_cmds/tsort/tsort.c
diskdev_cmds: Fix BINDIRs
[apple_cmds.git] / misc_cmds / tsort / tsort.c
1 /* $NetBSD: tsort.c,v 1.13 1998/08/25 20:59:42 ross Exp $ */
2
3 /*
4 * Copyright (c) 1989, 1993, 1994
5 * The Regents of the University of California. All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Michael Rendell of Memorial University of Newfoundland.
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 <sys/cdefs.h>
40 #ifndef lint
41 __COPYRIGHT("@(#) Copyright (c) 1989, 1993, 1994\n\
42 The Regents of the University of California. All rights reserved.\n");
43 #endif /* not lint */
44
45 #ifndef lint
46 #if 0
47 static char sccsid[] = "@(#)tsort.c 8.3 (Berkeley) 5/4/95";
48 #endif
49 __RCSID("$NetBSD: tsort.c,v 1.13 1998/08/25 20:59:42 ross Exp $");
50 #endif /* not lint */
51
52 #include <sys/types.h>
53
54 #include <ctype.h>
55 #include <db.h>
56 #include <err.h>
57 #include <errno.h>
58 #include <fcntl.h>
59 #include <stdio.h>
60 #include <stdlib.h>
61 #include <string.h>
62 #include <unistd.h>
63
64 /*
65 * Topological sort. Input is a list of pairs of strings separated by
66 * white space (spaces, tabs, and/or newlines); strings are written to
67 * standard output in sorted order, one per line.
68 *
69 * usage:
70 * tsort [-l] [inputfile]
71 * If no input file is specified, standard input is read.
72 *
73 * Should be compatable with AT&T tsort HOWEVER the output is not identical
74 * (i.e. for most graphs there is more than one sorted order, and this tsort
75 * usually generates a different one then the AT&T tsort). Also, cycle
76 * reporting seems to be more accurate in this version (the AT&T tsort
77 * sometimes says a node is in a cycle when it isn't).
78 *
79 * Michael Rendell, michael@stretch.cs.mun.ca - Feb 26, '90
80 */
81 #define HASHSIZE 53 /* doesn't need to be big */
82 #define NF_MARK 0x1 /* marker for cycle detection */
83 #define NF_ACYCLIC 0x2 /* this node is cycle free */
84 #define NF_NODEST 0x4 /* Unreachable */
85
86 typedef struct node_str NODE;
87
88 struct node_str {
89 NODE **n_prevp; /* pointer to previous node's n_next */
90 NODE *n_next; /* next node in graph */
91 NODE **n_arcs; /* array of arcs to other nodes */
92 int n_narcs; /* number of arcs in n_arcs[] */
93 int n_arcsize; /* size of n_arcs[] array */
94 int n_refcnt; /* # of arcs pointing to this node */
95 int n_flags; /* NF_* */
96 char n_name[1]; /* name of this node */
97 };
98
99 typedef struct _buf {
100 char *b_buf;
101 int b_bsize;
102 } BUF;
103
104 DB *db;
105 NODE *graph, **cycle_buf, **longest_cycle;
106 int debug, longest, quiet;
107
108 void add_arc __P((char *, char *));
109 void clear_cycle __P((void));
110 int find_cycle __P((NODE *, NODE *, int, int));
111 NODE *get_node __P((char *));
112 void *grow_buf __P((void *, int));
113 int main __P((int, char **));
114 void remove_node __P((NODE *));
115 void tsort __P((void));
116 void usage __P((void));
117
118 int
119 main(argc, argv)
120 int argc;
121 char *argv[];
122 {
123 BUF *b;
124 int c, n;
125 FILE *fp;
126 int bsize, ch, nused;
127 BUF bufs[2];
128
129 fp = NULL;
130 while ((ch = getopt(argc, argv, "dlq")) != -1)
131 switch (ch) {
132 case 'd':
133 debug = 1;
134 break;
135 case 'l':
136 longest = 1;
137 break;
138 case 'q':
139 quiet = 1;
140 break;
141 case '?':
142 default:
143 usage();
144 }
145 argc -= optind;
146 argv += optind;
147
148 switch (argc) {
149 case 0:
150 fp = stdin;
151 break;
152 case 1:
153 if ((fp = fopen(*argv, "r")) == NULL)
154 err(1, "%s", *argv);
155 break;
156 default:
157 usage();
158 }
159
160 for (b = bufs, n = 2; --n >= 0; b++)
161 b->b_buf = grow_buf(NULL, b->b_bsize = 1024);
162
163 /* parse input and build the graph */
164 for (n = 0, c = getc(fp);;) {
165 while (c != EOF && isspace(c))
166 c = getc(fp);
167 if (c == EOF)
168 break;
169
170 nused = 0;
171 b = &bufs[n];
172 bsize = b->b_bsize;
173 do {
174 b->b_buf[nused++] = c;
175 if (nused == bsize)
176 b->b_buf = grow_buf(b->b_buf, bsize *= 2);
177 c = getc(fp);
178 } while (c != EOF && !isspace(c));
179
180 b->b_buf[nused] = '\0';
181 b->b_bsize = bsize;
182 if (n)
183 add_arc(bufs[0].b_buf, bufs[1].b_buf);
184 n = !n;
185 }
186 (void)fclose(fp);
187 if (n)
188 errx(1, "odd data count");
189
190 /* do the sort */
191 tsort();
192 exit(0);
193 }
194
195 /* double the size of oldbuf and return a pointer to the new buffer. */
196 void *
197 grow_buf(bp, size)
198 void *bp;
199 int size;
200 {
201 if ((bp = realloc(bp, (u_int)size)) == NULL)
202 err(1, "realloc");
203 return (bp);
204 }
205
206 /*
207 * add an arc from node s1 to node s2 in the graph. If s1 or s2 are not in
208 * the graph, then add them.
209 */
210 void
211 add_arc(s1, s2)
212 char *s1, *s2;
213 {
214 NODE *n1;
215 NODE *n2;
216 int bsize, i;
217
218 n1 = get_node(s1);
219
220 if (!strcmp(s1, s2))
221 return;
222
223 n2 = get_node(s2);
224
225 /*
226 * Check if this arc is already here.
227 */
228 for (i = 0; i < n1->n_narcs; i++)
229 if (n1->n_arcs[i] == n2)
230 return;
231 /*
232 * Add it.
233 */
234 if (n1->n_narcs == n1->n_arcsize) {
235 if (!n1->n_arcsize)
236 n1->n_arcsize = 10;
237 bsize = n1->n_arcsize * sizeof(*n1->n_arcs) * 2;
238 n1->n_arcs = grow_buf(n1->n_arcs, bsize);
239 n1->n_arcsize = bsize / sizeof(*n1->n_arcs);
240 }
241 n1->n_arcs[n1->n_narcs++] = n2;
242 ++n2->n_refcnt;
243 }
244
245 /* Find a node in the graph (insert if not found) and return a pointer to it. */
246 NODE *
247 get_node(name)
248 char *name;
249 {
250 DBT data, key;
251 NODE *n;
252
253 if (db == NULL &&
254 (db = dbopen(NULL, O_RDWR, 0, DB_HASH, NULL)) == NULL)
255 err(1, "db: %s", name);
256
257 key.data = name;
258 key.size = strlen(name) + 1;
259
260 switch ((*db->get)(db, &key, &data, 0)) {
261 case 0:
262 memmove(&n, data.data, sizeof(n));
263 return (n);
264 case 1:
265 break;
266 default:
267 case -1:
268 err(1, "db: %s", name);
269 }
270
271 if ((n = malloc(sizeof(NODE) + key.size)) == NULL)
272 err(1, "malloc");
273
274 n->n_narcs = 0;
275 n->n_arcsize = 0;
276 n->n_arcs = NULL;
277 n->n_refcnt = 0;
278 n->n_flags = 0;
279 memmove(n->n_name, name, key.size);
280
281 /* Add to linked list. */
282 if ((n->n_next = graph) != NULL)
283 graph->n_prevp = &n->n_next;
284 n->n_prevp = &graph;
285 graph = n;
286
287 /* Add to hash table. */
288 data.data = &n;
289 data.size = sizeof(n);
290 if ((*db->put)(db, &key, &data, 0))
291 err(1, "db: %s", name);
292 return (n);
293 }
294
295
296 /*
297 * Clear the NODEST flag from all nodes.
298 */
299 void
300 clear_cycle()
301 {
302 NODE *n;
303
304 for (n = graph; n != NULL; n = n->n_next)
305 n->n_flags &= ~NF_NODEST;
306 }
307
308 /* do topological sort on graph */
309 void
310 tsort()
311 {
312 NODE *n, *next;
313 int cnt, i;
314
315 while (graph != NULL) {
316 /*
317 * Keep getting rid of simple cases until there are none left,
318 * if there are any nodes still in the graph, then there is
319 * a cycle in it.
320 */
321 do {
322 for (cnt = 0, n = graph; n != NULL; n = next) {
323 next = n->n_next;
324 if (n->n_refcnt == 0) {
325 remove_node(n);
326 ++cnt;
327 }
328 }
329 } while (graph != NULL && cnt);
330
331 if (graph == NULL)
332 break;
333
334 if (!cycle_buf) {
335 /*
336 * Allocate space for two cycle logs - one to be used
337 * as scratch space, the other to save the longest
338 * cycle.
339 */
340 for (cnt = 0, n = graph; n != NULL; n = n->n_next)
341 ++cnt;
342 cycle_buf = malloc((u_int)sizeof(NODE *) * cnt);
343 longest_cycle = malloc((u_int)sizeof(NODE *) * cnt);
344 if (cycle_buf == NULL || longest_cycle == NULL)
345 err(1, "malloc");
346 }
347 for (n = graph; n != NULL; n = n->n_next) {
348 if (!(n->n_flags & NF_ACYCLIC)) {
349 if ((cnt = find_cycle(n, n, 0, 0)) != 0) {
350 if (!quiet) {
351 warnx("cycle in data");
352 for (i = 0; i < cnt; i++)
353 warnx("%s",
354 longest_cycle[i]->n_name);
355 }
356 remove_node(n);
357 clear_cycle();
358 break;
359 } else {
360 /* to avoid further checks */
361 n->n_flags |= NF_ACYCLIC;
362 clear_cycle();
363 }
364 }
365 }
366 if (n == NULL)
367 errx(1, "internal error -- could not find cycle");
368 }
369 }
370
371 /* print node and remove from graph (does not actually free node) */
372 void
373 remove_node(n)
374 NODE *n;
375 {
376 NODE **np;
377 int i;
378
379 (void)printf("%s\n", n->n_name);
380 for (np = n->n_arcs, i = n->n_narcs; --i >= 0; np++)
381 --(*np)->n_refcnt;
382 n->n_narcs = 0;
383 *n->n_prevp = n->n_next;
384 if (n->n_next)
385 n->n_next->n_prevp = n->n_prevp;
386 }
387
388
389 /* look for the longest? cycle from node from to node to. */
390 int
391 find_cycle(from, to, longest_len, depth)
392 NODE *from, *to;
393 int depth, longest_len;
394 {
395 NODE **np;
396 int i, len;
397
398 /*
399 * avoid infinite loops and ignore portions of the graph known
400 * to be acyclic
401 */
402 if (from->n_flags & (NF_NODEST|NF_MARK|NF_ACYCLIC))
403 return (0);
404 from->n_flags |= NF_MARK;
405
406 for (np = from->n_arcs, i = from->n_narcs; --i >= 0; np++) {
407 cycle_buf[depth] = *np;
408 if (*np == to) {
409 if (depth + 1 > longest_len) {
410 longest_len = depth + 1;
411 (void)memcpy((char *)longest_cycle,
412 (char *)cycle_buf,
413 longest_len * sizeof(NODE *));
414 }
415 } else {
416 if ((*np)->n_flags & (NF_MARK|NF_ACYCLIC|NF_NODEST))
417 continue;
418 len = find_cycle(*np, to, longest_len, depth + 1);
419
420 if (debug)
421 (void)printf("%*s %s->%s %d\n", depth, "",
422 from->n_name, to->n_name, len);
423
424 if (len == 0)
425 (*np)->n_flags |= NF_NODEST;
426
427 if (len > longest_len)
428 longest_len = len;
429
430 if (len > 0 && !longest)
431 break;
432 }
433 }
434 from->n_flags &= ~NF_MARK;
435 return (longest_len);
436 }
437
438 void
439 usage()
440 {
441 (void)fprintf(stderr, "usage: tsort [-lq] [file]\n");
442 exit(1);
443 }