Class: String

Inherits:
Object
  • Object
show all
Defined in:
(unknown)

Instance Method Summary collapse

Instance Method Details

#transliterate(*args) ⇒ Object

string.transliterate(transform_string) -> string

Transliterates string using transform.



90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
 
# File 'ext/icunicode.c', line 90

static VALUE unicode_transliterate(int argc, VALUE *argv, VALUE string) {
  UChar str[BUF_SIZE];
  int32_t slen = 0;
  UErrorCode status = U_ZERO_ERROR;
  UTransliterator *trans;
  VALUE transform;

  rb_scan_args(argc, argv, "01", &transform);
  if (NIL_P(transform)) transform = rb_str_new2("Latin; Lower; NFD; [^[:letter:] [:space:] [0-9] [:punctuation:]] Remove; NFC");

  to_utf16(string, str, &slen);

  trans = get_trans(transform);
  utrans_transUChars(trans, str, &slen, BUF_SIZE, 0, &slen, &status);

  to_utf8(str, slen);
}

#unicode_sort_key ⇒ Object

string.unicode_sort_key -> string

Returns a string that will sort according to the Unicode collation algorithm.



37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
 
# File 'ext/icunicode.c', line 37

static VALUE unicode_sort_key(VALUE string) {
  char str[BUF_SIZE];
  UChar ustr[BUF_SIZE];
  int32_t len  = 0;
  int32_t ulen = 0;
  UErrorCode status = U_ZERO_ERROR;
  UCollator *col;

  to_utf16(string, ustr, &ulen);

  col = ucol_open("en_US", &status);
  if (U_SUCCESS(status)) {
    len = ucol_getSortKey(col, ustr, ulen, (uint8_t*)str, BUF_SIZE);
    ucol_close(col);
  }

  return rb_str_new(str, len - 1);
}