| 1 | #ifndef NU_STRCOLL_INTERNAL_H |
| 2 | #define NU_STRCOLL_INTERNAL_H |
| 3 | |
| 4 | /** @defgroup collation_internal Internal collation functions |
| 5 | * |
| 6 | * Functions in this group are mostly for the internal use. PLease use them |
| 7 | * with care. |
| 8 | */ |
| 9 | |
| 10 | #include <libnu/config.h> |
| 11 | #include <libnu/casemap.h> |
| 12 | #include <libnu/defines.h> |
| 13 | #include <libnu/strings.h> |
| 14 | |
| 15 | #if defined (__cplusplus) || defined (c_plusplus) |
| 16 | extern "C" { |
| 17 | #endif |
| 18 | |
| 19 | /** Read (decode) iterator with transformation applied inside of it |
| 20 | * |
| 21 | * @ingroup collation_internal |
| 22 | * @see nu_default_compound_read |
| 23 | * @see nu_nocase_compound_read |
| 24 | */ |
| 25 | typedef const char* (*nu_compound_read_t)( |
| 26 | const char *encoded, const char *encoded_limit, nu_read_iterator_t encoded_read, |
| 27 | uint32_t *unicode, const char **tail); |
| 28 | |
| 29 | /** Weight unicode codepoint (or several codepoints) |
| 30 | * |
| 31 | * 0 should always be weighted to 0. If your weight function need more |
| 32 | * than one codepoint - return negative value, which will be passed back to |
| 33 | * this function along with next codepoint. |
| 34 | * |
| 35 | * When function decided on weight and returned positive result, it has to |
| 36 | * fill weight with how many (Unicode) codepoints nunicode should rollback. |
| 37 | * E.g. function consumed "ZZS" and decided weight (in Hungarian collation), |
| 38 | * it fills 0 to \*weight because no rollback is needed. Then function |
| 39 | * consumed "ZZZ" and no weight available for such contraction - it |
| 40 | * returns weight for "Z" and fills \*weight with 2, to rollback |
| 41 | * redundant "ZZ". |
| 42 | * |
| 43 | * If string suddenly ends before weight function can decide (string limit |
| 44 | * reached), 0 will be passed additionally to the previous string to signal |
| 45 | * end of the string. |
| 46 | * |
| 47 | * @ingroup collation_internal |
| 48 | * @param u unicode codepoint to weight |
| 49 | * @param weight 0 at first call or (on sequential calls) pointer to negative |
| 50 | * weight previously returned by this function |
| 51 | * @param context pointer passed to _nu_strcoll() or _nu_strstr() |
| 52 | * @return positive codepoint weight or negative value if function need more |
| 53 | * codepoints |
| 54 | */ |
| 55 | typedef int32_t (*nu_codepoint_weight_t)(uint32_t u, int32_t *weight, void *context); |
| 56 | |
| 57 | #if (defined NU_WITH_Z_COLLATION) || (defined NU_WITH_N_COLLATION) |
| 58 | |
| 59 | /** Default compound read, equal to simply calling encoded_read(encoded, &unicode) |
| 60 | * |
| 61 | * @ingroup collation_internal |
| 62 | * @param encoded encoded string |
| 63 | * @param encoded_limit upper limit for encoded. NU_UNLIMITED for 0-terminated |
| 64 | * strings |
| 65 | * @param encoded_read read (decode) function |
| 66 | * @param unicode output unicode codepoint |
| 67 | * @param tail output pointer to compound tail, should never be 0 |
| 68 | * @return pointer to next encoded codepoint |
| 69 | */ |
| 70 | static inline |
| 71 | const char* nu_default_compound_read(const char *encoded, const char *encoded_limit, |
| 72 | nu_read_iterator_t encoded_read, uint32_t *unicode, |
| 73 | const char **tail) { |
| 74 | (void)(encoded_limit); |
| 75 | (void)(tail); |
| 76 | |
| 77 | return encoded_read(encoded, unicode); |
| 78 | } |
| 79 | |
| 80 | /** Case-ignoring compound read, equal to calling |
| 81 | * encoded_read(encoded, &unicode) with nu_toupper() applied internally |
| 82 | * |
| 83 | * @ingroup collation_internal |
| 84 | * @param encoded encoded string |
| 85 | * @param encoded_limit upper limit for encoded. NU_UNLIMITED for 0-terminated |
| 86 | * strings |
| 87 | * @param encoded_read read (decode) function |
| 88 | * @param unicode output unicode codepoint |
| 89 | * @param tail output pointer to compound tail, should never be 0 |
| 90 | * @return pointer to next encoded codepoint |
| 91 | */ |
| 92 | static inline |
| 93 | const char* nu_nocase_compound_read(const char *encoded, const char *encoded_limit, |
| 94 | nu_read_iterator_t encoded_read, uint32_t *unicode, |
| 95 | const char **tail) { |
| 96 | |
| 97 | /* re-entry with tail != 0 */ |
| 98 | if (*tail != 0) { |
| 99 | *tail = nu_casemap_read(utf8: *tail, unicode); |
| 100 | |
| 101 | if (*unicode != 0) { |
| 102 | return encoded; |
| 103 | } |
| 104 | |
| 105 | *tail = 0; // fall thru |
| 106 | } |
| 107 | |
| 108 | if (encoded >= encoded_limit) { |
| 109 | *unicode = 0; |
| 110 | return encoded; |
| 111 | } |
| 112 | |
| 113 | const char *p = encoded_read(encoded, unicode); |
| 114 | |
| 115 | if (*unicode == 0) { |
| 116 | return p; |
| 117 | } |
| 118 | |
| 119 | const char *map = NU_FOLDING_FUNCTION(codepoint: *unicode); |
| 120 | if (map != 0) { |
| 121 | *tail = nu_casemap_read(utf8: map, unicode); |
| 122 | } |
| 123 | |
| 124 | return p; |
| 125 | } |
| 126 | |
| 127 | /** Internal interface for nu_strcoll |
| 128 | * |
| 129 | * @ingroup collation_internal |
| 130 | * @param lhs left-hand side encoded string |
| 131 | * @param lhs_limit upper limit for lhs, use NU_UNLIMITED for 0-terminated |
| 132 | * strings |
| 133 | * @param rhs right-hand side encoded string |
| 134 | * @param rhs_limit upper limit for rhs, use NU_UNLIMITED for 0-terminated |
| 135 | * strings |
| 136 | * @param it1 lhs read (decoding) function |
| 137 | * @param it2 rhs read (decoding) function |
| 138 | * @param com1 lhs compound read function |
| 139 | * @param com2 rhs compound read function |
| 140 | * @param weight codepoint weighting function |
| 141 | * @param context pointer which will be passed to weight |
| 142 | * @param collated_left (optional) number of codepoints collated in lhs |
| 143 | * @param collated_right (optional) number of codepoints collated in rhs |
| 144 | * |
| 145 | * @see nu_strcoll |
| 146 | * @see nu_default_compound_read |
| 147 | * @see nu_nocase_compound_read |
| 148 | * @see nu_ducet_weight |
| 149 | */ |
| 150 | NU_EXPORT |
| 151 | int _nu_strcoll(const char *lhs, const char *lhs_limit, |
| 152 | const char *rhs, const char *rhs_limit, |
| 153 | nu_read_iterator_t it1, nu_read_iterator_t it2, |
| 154 | nu_compound_read_t com1, nu_compound_read_t com2, |
| 155 | nu_codepoint_weight_t weight, void *context, |
| 156 | ssize_t *collated_left, ssize_t *collated_right); |
| 157 | |
| 158 | /** Internal interface for nu_strchr |
| 159 | * |
| 160 | * @ingroup collation_internal |
| 161 | * @param lhs left-hand side encoded string |
| 162 | * @param lhs_limit upper limit for lhs, use NU_UNLIMITED for 0-terminated |
| 163 | * strings |
| 164 | * @param c unicode codepoint to look for |
| 165 | * @param read lhs read (decoding) function |
| 166 | * @param com lhs compound read function |
| 167 | * @param casemap casemapping function |
| 168 | * @param casemap_read casemapping result decoding function |
| 169 | * |
| 170 | * @see nu_strchr |
| 171 | * @see nu_default_compound_read |
| 172 | * @see nu_nocase_compound_read |
| 173 | * @see nu_toupper |
| 174 | * @see nu_tolower |
| 175 | */ |
| 176 | NU_EXPORT |
| 177 | const char* _nu_strchr(const char *lhs, const char *lhs_limit, |
| 178 | uint32_t c, nu_read_iterator_t read, |
| 179 | nu_compound_read_t com, |
| 180 | nu_casemapping_t casemap, nu_read_iterator_t casemap_read); |
| 181 | |
| 182 | /** Internal interface for nu_strchr |
| 183 | * |
| 184 | * @ingroup collation_internal |
| 185 | * @see _nu_strchr |
| 186 | */ |
| 187 | NU_EXPORT |
| 188 | const char* _nu_strrchr(const char *encoded, const char *limit, |
| 189 | uint32_t c, nu_read_iterator_t read, |
| 190 | nu_compound_read_t com, |
| 191 | nu_casemapping_t casemap, nu_read_iterator_t casemap_read); |
| 192 | |
| 193 | /** Internal interface for nu_strcoll |
| 194 | * |
| 195 | * @ingroup collation_internal |
| 196 | * @param haystack encoded haystack |
| 197 | * @param haystack_limit upper limit for haystack, use NU_UNLIMITED for |
| 198 | * 0-terminated strings |
| 199 | * @param needle encoded needle string |
| 200 | * @param needle_limit upper limit for needle, use NU_UNLIMITED for |
| 201 | * 0-terminated strings |
| 202 | * @param it1 haystack read (decoding) function |
| 203 | * @param it2 needle read (decoding) function |
| 204 | * @param com1 haystack compound read function |
| 205 | * @param com2 needle compound read function |
| 206 | * @param casemap casemapping function |
| 207 | * @param casemap_read casemapping result decoding function |
| 208 | * @param weight codepoint weighting function |
| 209 | * @param context pointer which will be passed to weight |
| 210 | * |
| 211 | * @see nu_strstr |
| 212 | * @see nu_default_compound_read |
| 213 | * @see nu_nocase_compound_read |
| 214 | * @see nu_toupper |
| 215 | * @see nu_tolower |
| 216 | * @see nu_ducet_weight |
| 217 | */ |
| 218 | NU_EXPORT |
| 219 | const char* _nu_strstr(const char *haystack, const char *haystack_limit, |
| 220 | const char *needle, const char *needle_limit, |
| 221 | nu_read_iterator_t it1, nu_read_iterator_t it2, |
| 222 | nu_compound_read_t com1, nu_compound_read_t com2, |
| 223 | nu_casemapping_t casemap, nu_read_iterator_t casemap_read, |
| 224 | nu_codepoint_weight_t weight, void *context); |
| 225 | |
| 226 | #endif /* (defined NU_WITH_Z_COLLATION) || (defined NU_WITH_N_COLLATION) */ |
| 227 | |
| 228 | #if defined (__cplusplus) || defined (c_plusplus) |
| 229 | } |
| 230 | #endif |
| 231 | |
| 232 | #endif /* NU_STRCOLL_INTERNAL_H */ |
| 233 | |