Re: db mess
On Sat, Jun 05, 1999 at 07:08:29PM +0200, Mark Kettenis wrote:
> I suspect that db2 from our glibc package is broken, but I have no
> experience with db. Could someone run a few tests? I can provide the tsort
> source file if anyonw wants to try.
>
> Please do so.
Attached. Failure in line 256.
> mandb should use or uses libdb2 from the db2 packages. This makes me wonder
> what all the db's are useful for. Isn't db2 also in glibc? Can someone give
> a quick overview?
>
> I will update the glibc package soon, maybe this fixes our problems, maybe
> not.
>
> I doubt it. It might be related to the fact that the Hurd uses
> seperate bits for O_READ and O_WRITE. This means that:
>
> the Hurd Linux
> ======================================
> O_RDONLY 0x0001 0x0000
> O_WRONLY 0x0002 0x0001
> O_RDWR 0x0003 0x0002
>
> This means that the magic to convert from open modes as given to
> open() to open modes that have to be given to db_open() is likely to
> be wrong. Some programs even hardcode the values (don't use the
> symbolic constants), which is even worse!
Hopefully you are right, because this will be easy to fix. I am not sure if
this would explain the man-db problem, which uses db2.
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' Debian http://www.debian.org finger brinkmd@
Marcus Brinkmann GNU http://www.gnu.org master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de for public PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
/*
* Copyright (c) 1989, 1993, 1994
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Michael Rendell of Memorial University of Newfoundland.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#ifndef lint
static char copyright[] =
"@(#) Copyright (c) 1989, 1993, 1994\n\
The Regents of the University of California. All rights reserved.\n";
#endif /* not lint */
#ifndef lint
static char sccsid[] = "@(#)tsort.c 8.2 (Berkeley) 3/30/94";
#endif /* not lint */
#include <sys/types.h>
#include <ctype.h>
#ifdef libc5
#include <db/db.h>
#else
#if defined __GLIBC__ && __GLIBC__ == 2 && __GLIBC_MINOR__ >= 1
#include <db_185.h>
#else
#include <db.h>
#endif
#endif
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif /* HAVE_UNISTD_H */
/*
* Topological sort. Input is a list of pairs of strings separated by
* white space (spaces, tabs, and/or newlines); strings are written to
* standard output in sorted order, one per line.
*
* usage:
* tsort [-l] [inputfile]
* If no input file is specified, standard input is read.
*
* Should be compatable with AT&T tsort HOWEVER the output is not identical
* (i.e. for most graphs there is more than one sorted order, and this tsort
* usually generates a different one then the AT&T tsort). Also, cycle
* reporting seems to be more accurate in this version (the AT&T tsort
* sometimes says a node is in a cycle when it isn't).
*
* Michael Rendell, michael@stretch.cs.mun.ca - Feb 26, '90
*/
#define HASHSIZE 53 /* doesn't need to be big */
#define NF_MARK 0x1 /* marker for cycle detection */
#define NF_ACYCLIC 0x2 /* this node is cycle free */
#define NF_NODEST 0x4 /* Unreachable */
typedef struct node_str NODE;
struct node_str {
NODE **n_prevp; /* pointer to previous node's n_next */
NODE *n_next; /* next node in graph */
NODE **n_arcs; /* array of arcs to other nodes */
int n_narcs; /* number of arcs in n_arcs[] */
int n_arcsize; /* size of n_arcs[] array */
int n_refcnt; /* # of arcs pointing to this node */
int n_flags; /* NF_* */
char n_name[1]; /* name of this node */
};
typedef struct _buf {
char *b_buf;
int b_bsize;
} BUF;
DB *db;
NODE *graph, **cycle_buf, **longest_cycle;
int debug, longest;
void add_arc __P((char *, char *));
int find_cycle __P((NODE *, NODE *, int, int));
NODE *get_node __P((char *));
void *grow_buf __P((void *, int));
void remove_node __P((NODE *));
void tsort __P((void));
void usage __P((void));
int
main(argc, argv)
int argc;
char *argv[];
{
register BUF *b;
register int c, n;
FILE *fp;
int bsize, ch, nused;
BUF bufs[2];
err_setprogname(argv[0]);
while ((ch = getopt(argc, argv, "dl")) != EOF)
switch (ch) {
case 'd':
debug = 1;
break;
case 'l':
longest = 1;
break;
case '?':
default:
usage();
}
argc -= optind;
argv += optind;
switch (argc) {
case 0:
fp = stdin;
break;
case 1:
if ((fp = fopen(*argv, "r")) == NULL)
err(1, "%s", *argv);
break;
default:
usage();
}
for (b = bufs, n = 2; --n >= 0; b++)
b->b_buf = grow_buf(NULL, b->b_bsize = 1024);
/* parse input and build the graph */
for (n = 0, c = getc(fp);;) {
while (c != EOF && isspace(c))
c = getc(fp);
if (c == EOF)
break;
nused = 0;
b = &bufs[n];
bsize = b->b_bsize;
do {
b->b_buf[nused++] = c;
if (nused == bsize)
b->b_buf = grow_buf(b->b_buf, bsize *= 2);
c = getc(fp);
} while (c != EOF && !isspace(c));
b->b_buf[nused] = '\0';
b->b_bsize = bsize;
if (n)
add_arc(bufs[0].b_buf, bufs[1].b_buf);
n = !n;
}
(void)fclose(fp);
if (n)
errx(1, "odd data count");
/* do the sort */
tsort();
exit(0);
}
/* double the size of oldbuf and return a pointer to the new buffer. */
void *
grow_buf(bp, size)
void *bp;
int size;
{
if ((bp = realloc(bp, (u_int)size)) == NULL)
err(1, NULL);
return (bp);
}
/*
* add an arc from node s1 to node s2 in the graph. If s1 or s2 are not in
* the graph, then add them.
*/
void
add_arc(s1, s2)
char *s1, *s2;
{
register NODE *n1;
NODE *n2;
int bsize, i;
n1 = get_node(s1);
if (!strcmp(s1, s2))
return;
n2 = get_node(s2);
/*
* Check if this arc is already here.
*/
for (i = 0; i < n1->n_narcs; i++)
if (n1->n_arcs[i] == n2)
return;
/*
* Add it.
*/
if (n1->n_narcs == n1->n_arcsize) {
if (!n1->n_arcsize)
n1->n_arcsize = 10;
bsize = n1->n_arcsize * sizeof(*n1->n_arcs) * 2;
n1->n_arcs = grow_buf(n1->n_arcs, bsize);
n1->n_arcsize = bsize / sizeof(*n1->n_arcs);
}
n1->n_arcs[n1->n_narcs++] = n2;
++n2->n_refcnt;
}
/* Find a node in the graph (insert if not found) and return a pointer to it. */
NODE *
get_node(name)
char *name;
{
DBT data, key;
NODE *n;
if (db == NULL &&
(db = dbopen(NULL, O_RDWR, 0, DB_HASH, NULL)) == NULL)
err(1, "db: %s", name);
key.data = name;
key.size = strlen(name) + 1;
switch ((*db->get)(db, &key, &data, 0)) {
case 0:
bcopy(data.data, &n, sizeof(n));
return (n);
case 1:
break;
default:
case -1:
err(1, "db: %s", name);
}
if ((n = malloc(sizeof(NODE) + key.size)) == NULL)
err(1, NULL);
n->n_narcs = 0;
n->n_arcsize = 0;
n->n_arcs = NULL;
n->n_refcnt = 0;
n->n_flags = 0;
bcopy(name, n->n_name, key.size);
/* Add to linked list. */
if ((n->n_next = graph) != NULL)
graph->n_prevp = &n->n_next;
n->n_prevp = &graph;
graph = n;
/* Add to hash table. */
data.data = &n;
data.size = sizeof(n);
if ((*db->put)(db, &key, &data, 0))
err(1, "db: %s", name);
return (n);
}
/*
* Clear the NODEST flag from all nodes.
*/
void
clear_cycle()
{
NODE *n;
for (n = graph; n != NULL; n = n->n_next)
n->n_flags &= ~NF_NODEST;
}
/* do topological sort on graph */
void
tsort()
{
register NODE *n, *next;
register int cnt, i;
while (graph != NULL) {
/*
* Keep getting rid of simple cases until there are none left,
* if there are any nodes still in the graph, then there is
* a cycle in it.
*/
do {
for (cnt = 0, n = graph; n != NULL; n = next) {
next = n->n_next;
if (n->n_refcnt == 0) {
remove_node(n);
++cnt;
}
}
} while (graph != NULL && cnt);
if (graph == NULL)
break;
if (!cycle_buf) {
/*
* Allocate space for two cycle logs - one to be used
* as scratch space, the other to save the longest
* cycle.
*/
for (cnt = 0, n = graph; n != NULL; n = n->n_next)
++cnt;
cycle_buf = malloc((u_int)sizeof(NODE *) * cnt);
longest_cycle = malloc((u_int)sizeof(NODE *) * cnt);
if (cycle_buf == NULL || longest_cycle == NULL)
err(1, NULL);
}
for (n = graph; n != NULL; n = n->n_next)
if (!(n->n_flags & NF_ACYCLIC))
if (cnt = find_cycle(n, n, 0, 0)) {
warnx("cycle in data");
for (i = 0; i < cnt; i++)
warnx("%s",
longest_cycle[i]->n_name);
remove_node(n);
clear_cycle();
break;
} else {
/* to avoid further checks */
n->n_flags |= NF_ACYCLIC;
clear_cycle();
}
if (n == NULL)
errx(1, "internal error -- could not find cycle");
}
}
/* print node and remove from graph (does not actually free node) */
void
remove_node(n)
register NODE *n;
{
register NODE **np;
register int i;
(void)printf("%s\n", n->n_name);
for (np = n->n_arcs, i = n->n_narcs; --i >= 0; np++)
--(*np)->n_refcnt;
n->n_narcs = 0;
*n->n_prevp = n->n_next;
if (n->n_next)
n->n_next->n_prevp = n->n_prevp;
}
/* look for the longest? cycle from node from to node to. */
int
find_cycle(from, to, longest_len, depth)
NODE *from, *to;
int depth, longest_len;
{
register NODE **np;
register int i, len;
/*
* avoid infinite loops and ignore portions of the graph known
* to be acyclic
*/
if (from->n_flags & (NF_NODEST|NF_MARK|NF_ACYCLIC))
return (0);
from->n_flags |= NF_MARK;
for (np = from->n_arcs, i = from->n_narcs; --i >= 0; np++) {
cycle_buf[depth] = *np;
if (*np == to) {
if (depth + 1 > longest_len) {
longest_len = depth + 1;
(void)memcpy((char *)longest_cycle,
(char *)cycle_buf,
longest_len * sizeof(NODE *));
}
} else {
if ((*np)->n_flags & (NF_MARK|NF_ACYCLIC|NF_NODEST))
continue;
len = find_cycle(*np, to, longest_len, depth + 1);
if (debug)
(void)printf("%*s %s->%s %d\n", depth, "",
from->n_name, to->n_name, len);
if (len == 0)
(*np)->n_flags |= NF_NODEST;
if (len > longest_len)
longest_len = len;
if (len > 0 && !longest)
break;
}
}
from->n_flags &= ~NF_MARK;
return (longest_len);
}
void
usage()
{
(void)fprintf(stderr, "usage: tsort [-l] [file]\n");
exit(1);
}
Reply to:
- References:
- db mess
- From: Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de>
- Re: db mess
- From: Mark Kettenis <kettenis@wins.uva.nl>