Commit 396bbdbe authored by bar@mysql.com's avatar bar@mysql.com

Merge abarkov@bk-internal.mysql.com:/home/bk/mysql-4.1

into mysql.com:/usr/home/bar/mysql-4.1
parents 28a02c9b 2d1e9862
......@@ -62,6 +62,9 @@ extern CHARSET_INFO my_charset_utf8_slovak_uca_ci;
extern CHARSET_INFO my_charset_utf8_spanish2_uca_ci;
extern CHARSET_INFO my_charset_utf8_roman_uca_ci;
extern CHARSET_INFO my_charset_utf8_persian_uca_ci;
#ifdef HAVE_CYBOZU_COLLATION
extern CHARSET_INFO my_charset_utf8_general_cs;
#endif
#endif
#endif /* HAVE_UCA_COLLATIONS */
......@@ -146,6 +149,9 @@ my_bool init_compiled_charsets(myf flags __attribute__((unused)))
#ifdef HAVE_CHARSET_utf8
add_compiled_collation(&my_charset_utf8_general_ci);
add_compiled_collation(&my_charset_utf8_bin);
#ifdef HAVE_CYBOZU_COLLATION
add_compiled_collation(&my_charset_utf8_general_cs);
#endif
#ifdef HAVE_UCA_COLLATIONS
add_compiled_collation(&my_charset_utf8_general_uca_ci);
add_compiled_collation(&my_charset_utf8_icelandic_uca_ci);
......
......@@ -2380,6 +2380,172 @@ CHARSET_INFO my_charset_utf8_bin=
&my_collation_mb_bin_handler
};
#ifdef HAVE_CYBOZU_COLLATION
/*
* These functions bacically do the same as their original, except
* that they return 0 only when two comparing unicode strings are
* strictly the same in case-sensitive way. See "save_diff" local
* variable to what they actually do.
*/
static int my_strnncoll_utf8_cs(CHARSET_INFO *cs,
const uchar *s, uint slen,
const uchar *t, uint tlen,
my_bool t_is_prefix)
{
int s_res,t_res;
my_wc_t s_wc,t_wc;
const uchar *se=s+slen;
const uchar *te=t+tlen;
int save_diff = 0;
int diff;
while ( s < se && t < te )
{
int plane;
s_res=my_utf8_uni(cs,&s_wc, s, se);
t_res=my_utf8_uni(cs,&t_wc, t, te);
if ( s_res <= 0 || t_res <= 0 )
{
/* Incorrect string, compare by char value */
return ((int)s[0]-(int)t[0]);
}
if ( save_diff == 0 )
{
save_diff = ((int)s_wc) - ((int)t_wc);
}
plane=(s_wc>>8) & 0xFF;
s_wc = uni_plane[plane] ? uni_plane[plane][s_wc & 0xFF].sort : s_wc;
plane=(t_wc>>8) & 0xFF;
t_wc = uni_plane[plane] ? uni_plane[plane][t_wc & 0xFF].sort : t_wc;
if ( s_wc != t_wc )
{
return ((int) s_wc) - ((int) t_wc);
}
s+=s_res;
t+=t_res;
}
diff = ( (se-s) - (te-t) );
return t_is_prefix ? t-te : ((diff == 0) ? save_diff : diff);
}
static int my_strnncollsp_utf8_cs(CHARSET_INFO *cs,
const uchar *s, uint slen,
const uchar *t, uint tlen)
{
int s_res,t_res;
my_wc_t s_wc,t_wc;
const uchar *se= s+slen;
const uchar *te= t+tlen;
int save_diff = 0;
while ( s < se && t < te )
{
int plane;
s_res=my_utf8_uni(cs,&s_wc, s, se);
t_res=my_utf8_uni(cs,&t_wc, t, te);
if ( s_res <= 0 || t_res <= 0 )
{
/* Incorrect string, compare by char value */
return ((int)s[0]-(int)t[0]);
}
if ( save_diff == 0 )
{
save_diff = ((int)s_wc) - ((int)t_wc);
}
plane=(s_wc>>8) & 0xFF;
s_wc = uni_plane[plane] ? uni_plane[plane][s_wc & 0xFF].sort : s_wc;
plane=(t_wc>>8) & 0xFF;
t_wc = uni_plane[plane] ? uni_plane[plane][t_wc & 0xFF].sort : t_wc;
if ( s_wc != t_wc )
{
return ((int) s_wc) - ((int) t_wc);
}
s+=s_res;
t+=t_res;
}
slen= se-s;
tlen= te-t;
if (slen != tlen)
{
int swap= 0;
if (slen < tlen)
{
slen= tlen;
s= t;
se= te;
swap= -1;
}
/*
This following loop uses the fact that in UTF-8
all multibyte characters are greater than space,
and all multibyte head characters are greater than
space. It means if we meet a character greater
than space, it always means that the longer string
is greater. So we can reuse the same loop from the
8bit version, without having to process full multibute
sequences.
*/
for ( ; s < se; s++)
{
if (*s != ' ')
return ((int)*s - (int) ' ') ^ swap;
}
}
return save_diff;
}
static MY_COLLATION_HANDLER my_collation_cs_handler =
{
NULL, /* init */
my_strnncoll_utf8_cs,
my_strnncollsp_utf8_cs,
my_strnxfrm_utf8,
my_like_range_simple,
my_wildcmp_mb,
my_strcasecmp_utf8,
my_instr_mb,
my_hash_sort_utf8
};
CHARSET_INFO my_charset_utf8_general_cs=
{
254,0,0, /* number */
MY_CS_COMPILED|MY_CS_UNICODE, /* state */
"utf8", /* cs name */
"utf8_general_cs", /* name */
"", /* comment */
NULL, /* tailoring */
ctype_utf8, /* ctype */
to_lower_utf8, /* to_lower */
to_upper_utf8, /* to_upper */
to_upper_utf8, /* sort_order */
NULL, /* contractions */
NULL, /* sort_order_big*/
NULL, /* tab_to_uni */
NULL, /* tab_from_uni */
NULL, /* state_map */
NULL, /* ident_map */
1, /* strxfrm_multiply */
1, /* mbminlen */
3, /* mbmaxlen */
0, /* min_sort_char */
255, /* max_sort_char */
&my_charset_utf8_handler,
&my_collation_cs_handler
};
#endif /* Cybozu Hack */
#ifdef MY_TEST_UTF8
#include <stdio.h>
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment