1 /* stringlib: fastsearch implementation */
3 #ifndef STRINGLIB_FASTSEARCH_H
4 #define STRINGLIB_FASTSEARCH_H
6 /* fast search/count implementation, based on a mix between boyer-
7 moore and horspool, with a few more bells and whistles on the top.
8 for some more background, see: http://effbot.org/zone/stringlib.htm */
10 /* note: fastsearch may access s[n], which isn't a problem when using
11 Python's ordinary string types, but may cause problems if you're
12 using this code in other contexts. also, the count mode returns -1
13 if there cannot possible be a match in the target string, and 0 if
14 it has actually checked for matches, but didn't find any. callers
19 #define FAST_RSEARCH 2
22 #define STRINGLIB_BLOOM_WIDTH 128
24 #define STRINGLIB_BLOOM_WIDTH 64
26 #define STRINGLIB_BLOOM_WIDTH 32
28 #error "LONG_BIT is smaller than 32"
31 #define STRINGLIB_BLOOM_ADD(mask, ch) \
32 ((mask |= (1UL << ((ch) & (STRINGLIB_BLOOM_WIDTH -1)))))
33 #define STRINGLIB_BLOOM(mask, ch) \
34 ((mask & (1UL << ((ch) & (STRINGLIB_BLOOM_WIDTH -1)))))
36 Py_LOCAL_INLINE(Py_ssize_t)
37 fastsearch(const STRINGLIB_CHAR* s, Py_ssize_t n,
38 const STRINGLIB_CHAR* p, Py_ssize_t m,
39 Py_ssize_t maxcount, int mode)
42 Py_ssize_t skip, count = 0;
43 Py_ssize_t i, j, mlast, w;
47 if (w < 0 || (mode == FAST_COUNT && maxcount == 0))
50 /* look for special cases */
54 /* use special case for 1-character strings */
55 if (mode == FAST_COUNT) {
56 for (i = 0; i < n; i++)
59 if (count == maxcount)
63 } else if (mode == FAST_SEARCH) {
64 for (i = 0; i < n; i++)
67 } else { /* FAST_RSEARCH */
68 for (i = n - 1; i > -1; i--)
79 if (mode != FAST_RSEARCH) {
81 /* create compressed boyer-moore delta 1 table */
83 /* process pattern[:-1] */
84 for (i = 0; i < mlast; i++) {
85 STRINGLIB_BLOOM_ADD(mask, p[i]);
89 /* process pattern[-1] outside the loop */
90 STRINGLIB_BLOOM_ADD(mask, p[mlast]);
92 for (i = 0; i <= w; i++) {
93 /* note: using mlast in the skip path slows things down on x86 */
94 if (s[i+m-1] == p[m-1]) {
96 for (j = 0; j < mlast; j++)
101 if (mode != FAST_COUNT)
104 if (count == maxcount)
109 /* miss: check if next character is part of pattern */
110 if (!STRINGLIB_BLOOM(mask, s[i+m]))
115 /* skip: check if next character is part of pattern */
116 if (!STRINGLIB_BLOOM(mask, s[i+m]))
120 } else { /* FAST_RSEARCH */
122 /* create compressed boyer-moore delta 1 table */
124 /* process pattern[0] outside the loop */
125 STRINGLIB_BLOOM_ADD(mask, p[0]);
126 /* process pattern[:0:-1] */
127 for (i = mlast; i > 0; i--) {
128 STRINGLIB_BLOOM_ADD(mask, p[i]);
133 for (i = w; i >= 0; i--) {
135 /* candidate match */
136 for (j = mlast; j > 0; j--)
142 /* miss: check if previous character is part of pattern */
143 if (i > 0 && !STRINGLIB_BLOOM(mask, s[i-1]))
148 /* skip: check if previous character is part of pattern */
149 if (i > 0 && !STRINGLIB_BLOOM(mask, s[i-1]))
155 if (mode != FAST_COUNT)