src/string/strstr.c

   1 #include <string.h>
   2 #include <stdint.h>
   3
   4 static char *twobyte_strstr(const unsigned char *h, const unsigned char *n)
   5 {
   6         uint16_t nw = n[0]<<8 | n[1], hw = h[0]<<8 | h[1];
   7         for (h++; *h && hw != nw; hw = hw<<8 | *++h);
   8         return *h ? (char *)h-1 : 0;
   9 }
  10
  11 static char *threebyte_strstr(const unsigned char *h, const unsigned char *n)
  12 {
  13         uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8;
  14         uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8;
  15         for (h+=2; *h && hw != nw; hw = (hw|*++h)<<8);
  16         return *h ? (char *)h-2 : 0;
  17 }
  18
  19 static char *fourbyte_strstr(const unsigned char *h, const unsigned char *n)
  20 {
  21         uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8 | n[3];
  22         uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8 | h[3];
  23         for (h+=3; *h && hw != nw; hw = hw<<8 | *++h);
  24         return *h ? (char *)h-3 : 0;
  25 }
  26
  27 #define MAX(a,b) ((a)>(b)?(a):(b))
  28 #define MIN(a,b) ((a)<(b)?(a):(b))
  29
  30 #define BITOP(a,b,op) \
  31  ((a)[(size_t)(b)/(8*sizeof *(a))] op (size_t)1<<((size_t)(b)%(8*sizeof *(a))))
  32
  33 static char *twoway_strstr(const unsigned char *h, const unsigned char *n)
  34 {
  35         const unsigned char *z;
  36         size_t l, ip, jp, k, p, ms, p0, mem, mem0;
  37         size_t byteset[32 / sizeof(size_t)] = { 0 };
  38         size_t shift[256];
  39
  40         /* Computing length of needle and fill shift table */
  41         for (l=0; n[l] && h[l]; l++)
  42                 BITOP(byteset, n[l], |=), shift[n[l]] = l+1;
  43         if (n[l]) return 0; /* hit the end of h */
  44
  45         /* Compute maximal suffix */
  46         ip = -1; jp = 0; k = p = 1;
  47         while (jp+k<l) {
  48                 if (n[ip+k] == n[jp+k]) {
  49                         if (k == p) {
  50                                 jp += p;
  51                                 k = 1;
  52                         } else k++;
  53                 } else if (n[ip+k] > n[jp+k]) {
  54                         jp += k;
  55                         k = 1;
  56                         p = jp - ip;
  57                 } else {
  58                         ip = jp++;
  59                         k = p = 1;
  60                 }
  61         }
  62         ms = ip;
  63         p0 = p;
  64
  65         /* And with the opposite comparison */
  66         ip = -1; jp = 0; k = p = 1;
  67         while (jp+k<l) {
  68                 if (n[ip+k] == n[jp+k]) {
  69                         if (k == p) {
  70                                 jp += p;
  71                                 k = 1;
  72                         } else k++;
  73                 } else if (n[ip+k] < n[jp+k]) {
  74                         jp += k;
  75                         k = 1;
  76                         p = jp - ip;
  77                 } else {
  78                         ip = jp++;
  79                         k = p = 1;
  80                 }
  81         }
  82         if (ip+1 > ms+1) ms = ip;
  83         else p = p0;
  84
  85         /* Periodic needle? */
  86         if (memcmp(n, n+p, ms+1)) {
  87                 mem0 = 0;
  88                 p = MAX(ms, l-ms-1) + 1;
  89         } else mem0 = l-p;
  90         mem = 0;
  91
  92         /* Initialize incremental end-of-haystack pointer */
  93         z = h;
  94
  95         /* Search loop */
  96         for (;;) {
  97                 /* Update incremental end-of-haystack pointer */
  98                 if (z-h < l) {
  99                         /* Fast estimate for MIN(l,63) */
 100                         size_t grow = l | 63;
 101                         const unsigned char *z2 = memchr(z, 0, grow);
 102                         if (z2) {
 103                                 z = z2;
 104                                 if (z-h < l) return 0;
 105                         } else z += grow;
 106                 }
 107
 108                 /* Check last byte first; advance by shift on mismatch */
 109                 if (BITOP(byteset, h[l-1], &)) {
 110                         k = l-shift[h[l-1]];
 111                         //printf("adv by %zu (on %c) at [%s] (%zu;l=%zu)\n", k, h[l-1], h, shift[h[l-1]], l);
 112                         if (k) {
 113                                 if (mem0 && mem && k < p) k = l-p;
 114                                 h += k;
 115                                 mem = 0;
 116                                 continue;
 117                         }
 118                 } else {
 119                         h += l;
 120                         mem = 0;
 121                         continue;
 122                 }
 123
 124                 /* Compare right half */
 125                 for (k=MAX(ms+1,mem); n[k] && n[k] == h[k]; k++);
 126                 if (n[k]) {
 127                         h += k-ms;
 128                         mem = 0;
 129                         continue;
 130                 }
 131                 /* Compare left half */
 132                 for (k=ms+1; k>mem && n[k-1] == h[k-1]; k--);
 133                 if (k <= mem) return (char *)h;
 134                 h += p;
 135                 mem = mem0;
 136         }
 137 }
 138
 139 char *strstr(const char *h, const char *n)
 140 {
 141         /* Return immediately on empty needle */
 142         if (!n[0]) return (char *)h;
 143
 144         /* Use faster algorithms for short needles */
 145         h = strchr(h, *n);
 146         if (!h || !n[1]) return (char *)h;
 147         if (!h[1]) return 0;
 148         if (!n[2]) return twobyte_strstr((void *)h, (void *)n);
 149         if (!h[2]) return 0;
 150         if (!n[3]) return threebyte_strstr((void *)h, (void *)n);
 151         if (!h[3]) return 0;
 152         if (!n[4]) return fourbyte_strstr((void *)h, (void *)n);
 153
 154         return twoway_strstr((void *)h, (void *)n);
 155 }