lib/index.h

   1 /*
   2  *      Sherlock: Data structures used in indices
   3  *
   4  *      (c) 2001--2002 Martin Mares <mj@ucw.cz>
   5  */
   6
   7 /* Words */
   8
   9 #define MAX_WORD_LEN            64
  10
  11 /* Word and string types are defined in lib/custom.h */
  12
  13 /* Index card attributes */
  14
  15 struct card_attr {
  16   u32 card;                             /* Reference to card description (either oid or filepos) */
  17   u32 site_id;
  18 #define INT_ATTR(t,i,o,k,g,p) t i;
  19   CUSTOM_ATTRS                          /* Include all custom attributes */
  20 #undef INT_ATTR
  21   byte weight;
  22   byte flags;
  23   // byte rfu[2];                       /* If no custom attributes are defined */
  24 };
  25
  26 enum card_flag {
  27   CARD_FLAG_EMPTY = 1,                  /* Empty document (redirect, robot file etc.) [scanner] */
  28   CARD_FLAG_ACCENTED = 2,               /* Document contains accented characters [scanner] */
  29   CARD_FLAG_DUP = 4,                    /* Removed as a duplicate [merger] */
  30   CARD_FLAG_MERGED = 8,                 /* Destination of a merge [merger] */
  31 };
  32
  33 #define CARD_POS_SHIFT 5                /* Card positions are shifted this # of bytes to the right */
  34
  35 /* String fingerprints */
  36
  37 struct fingerprint {
  38   byte hash[12];
  39 };
  40
  41 void fingerprint(byte *string, struct fingerprint *fp);
  42
  43 static inline u32
  44 fp_hash(struct fingerprint *fp)
  45 {
  46   return (fp->hash[0] << 24) | (fp->hash[1] << 16) | (fp->hash[2] << 8) | fp->hash[3];
  47 }
  48
  49 /* Reading of tagged text (Unicode values, tags mapped to 0x80000000 and higher) */
  50
  51 #define GET_TAGGED_CHAR(p,u) do {                               \
  52   u = *p;                                                       \
  53   if (u >= 0xc0)                                                \
  54     GET_UTF8_CHAR(p,u);                                         \
  55   else if (u >= 0x80)                                           \
  56     {                                                           \
  57       p++;                                                      \
  58       if (u >= 0xb0)                                            \
  59         {                                                       \
  60           if (u != 0xb0)                                        \
  61             ASSERT(0);                                          \
  62           u += 0x80020000;                                      \
  63         }                                                       \
  64       else if (u >= 0xa0)                                       \
  65         {                                                       \
  66           ASSERT(*p >= 0x80 && *p <= 0xbf);                     \
  67           u = 0x80010000 + ((u & 0x0f) << 6) + (*p++ & 0x3f);   \
  68         }                                                       \
  69       else                                                      \
  70         u += 0x80000000;                                        \
  71     }                                                           \
  72   else                                                          \
  73     p++;                                                        \
  74 } while (0)