annotate lib/strstr.c @ 9601:fbf94c8a908a

Share two-way algorithm. * lib/str-two-way.h: New file, merged from... * lib/memmem.c: ...here... * lib/strstr.c: ...and here. * modules/memmem (Files): Use it. * modules/strstr (Files): Likewise. Signed-off-by: Eric Blake <ebb9@byu.net>
author Eric Blake <ebb9@byu.net>
date Thu, 10 Jan 2008 12:06:35 -0700
parents 99268e709d31
children e8d2c6fc33ad
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
9600
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
1 /* Copyright (C) 1991,92,93,94,96,97,98,2000,2004,2007,2008 Free Software
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
2 Foundation, Inc.
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
3 This file is part of the GNU C Library.
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
4
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
5 This program is free software; you can redistribute it and/or modify
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
6 it under the terms of the GNU General Public License as published by
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
7 the Free Software Foundation; either version 2, or (at your option)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
8 any later version.
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
9
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
10 This program is distributed in the hope that it will be useful,
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
13 GNU General Public License for more details.
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
14
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
15 You should have received a copy of the GNU General Public License along
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
16 with this program; if not, write to the Free Software Foundation,
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
17 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
18
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
19 /* This particular implementation was written by Eric Blake, 2008. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
20
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
21 #ifndef _LIBC
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
22 # include <config.h>
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
23 #endif
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
24
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
25 /* Specification of strstr. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
26 #include <string.h>
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
27
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
28 #include <stdbool.h>
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
29
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
30 #ifndef _LIBC
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
31 # define __builtin_expect(expr, val) (expr)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
32 #endif
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
33
9601
fbf94c8a908a Share two-way algorithm.
Eric Blake <ebb9@byu.net>
parents: 9600
diff changeset
34 #define RETURN_TYPE char *
fbf94c8a908a Share two-way algorithm.
Eric Blake <ebb9@byu.net>
parents: 9600
diff changeset
35 #define AVAILABLE(h, h_l, j, n_l) \
fbf94c8a908a Share two-way algorithm.
Eric Blake <ebb9@byu.net>
parents: 9600
diff changeset
36 (!memchr ((h) + (h_l), '\0', (j) + (n_l) - (h_l)) \
fbf94c8a908a Share two-way algorithm.
Eric Blake <ebb9@byu.net>
parents: 9600
diff changeset
37 && ((h_l) = (j) + (n_l)))
fbf94c8a908a Share two-way algorithm.
Eric Blake <ebb9@byu.net>
parents: 9600
diff changeset
38 #include "str-two-way.h"
9600
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
39
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
40 /* Return the first occurrence of NEEDLE in HAYSTACK. Return HAYSTACK
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
41 if NEEDLE is empty, otherwise NULL if NEEDLE is not found in
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
42 HAYSTACK. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
43 char *
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
44 strstr (const char *haystack_start, const char *needle_start)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
45 {
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
46 const char *haystack = haystack_start;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
47 const char *needle = needle_start;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
48 size_t needle_len; /* Length of NEEDLE. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
49 size_t haystack_len; /* Known minimum length of HAYSTACK. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
50 bool ok = true; /* True if NEEDLE is prefix of HAYSTACK. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
51
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
52 /* Determine length of NEEDLE, and in the process, make sure
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
53 HAYSTACK is at least as long (no point processing all of a long
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
54 NEEDLE if HAYSTACK is too short). */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
55 while (*haystack && *needle)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
56 ok &= *haystack++ == *needle++;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
57 if (*needle)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
58 return NULL;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
59 if (ok)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
60 return (char *) haystack_start;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
61
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
62 /* Reduce the size of haystack using strchr, since it has a smaller
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
63 linear coefficient than the Two-Way algorithm. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
64 needle_len = needle - needle_start;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
65 haystack = strchr (haystack_start + 1, *needle_start);
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
66 if (!haystack || __builtin_expect (needle_len == 1, 0))
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
67 return (char *) haystack;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
68 needle -= needle_len;
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
69 haystack_len = (haystack > haystack_start + needle_len ? 1
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
70 : needle_len + haystack_start - haystack);
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
71
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
72 /* Perform the search. Abstract memory is considered to be an array
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
73 of 'unsigned char' values, not an array of 'char' values. See
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
74 ISO C 99 section 6.2.6.1. */
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
75 if (needle_len < LONG_NEEDLE_THRESHOLD)
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
76 return two_way_short_needle ((const unsigned char *) haystack,
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
77 haystack_len,
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
78 (const unsigned char *) needle, needle_len);
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
79 return two_way_long_needle ((const unsigned char *) haystack, haystack_len,
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
80 (const unsigned char *) needle, needle_len);
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
81 }
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
82
99268e709d31 Avoid quadratic strstr implementations.
Eric Blake <ebb9@byu.net>
parents:
diff changeset
83 #undef LONG_NEEDLE_THRESHOLD