diff options
author | unknown <serg@serg.mylan> | 2004-05-10 12:39:01 +0200 |
---|---|---|
committer | unknown <serg@serg.mylan> | 2004-05-10 12:39:01 +0200 |
commit | 0bfea087af007309ce06d346ecfb0683a3d012e7 (patch) | |
tree | aa63f206deb5fdf6e373750f990cbe8768ce3c53 /myisam | |
parent | 2d776e36d8dd205efccc18a75292308fd4c5d385 (diff) | |
download | mariadb-git-0bfea087af007309ce06d346ecfb0683a3d012e7.tar.gz |
backport from 4.1:
"phrase search" should not match partial words (it should not match 'paraphrase searches')
Diffstat (limited to 'myisam')
-rw-r--r-- | myisam/ft_boolean_search.c | 29 | ||||
-rw-r--r-- | myisam/ft_parser.c | 9 | ||||
-rw-r--r-- | myisam/ftdefs.h | 5 |
3 files changed, 22 insertions, 21 deletions
diff --git a/myisam/ft_boolean_search.c b/myisam/ft_boolean_search.c index d728c379ea5..61381f80783 100644 --- a/myisam/ft_boolean_search.c +++ b/myisam/ft_boolean_search.c @@ -360,25 +360,34 @@ err: } -/* returns 1 if str0 contain str1 */ +/* returns 1 if str0 ~= /\<str1\>/ */ static int _ftb_strstr(const byte *s0, const byte *e0, const byte *s1, const byte *e1, CHARSET_INFO *cs) { - const byte *p; + const byte *p0, *p1; + my_bool s_after, e_before; - while (s0 < e0) + s_after=true_word_char(s1[0]); + e_before=true_word_char(e1[-1]); + p0=s0; + + while (p0 < e0) { - while (s0 < e0 && cs->to_upper[(uint) (uchar) *s0++] != + while (p0 < e0 && cs->to_upper[(uint) (uchar) *p0++] != cs->to_upper[(uint) (uchar) *s1]) /* no-op */; - if (s0 >= e0) + if (p0 >= e0) return 0; - p=s1+1; - while (s0 < e0 && p < e1 && cs->to_upper[(uint) (uchar) *s0] == - cs->to_upper[(uint) (uchar) *p]) - s0++, p++; - if (p >= e1) + + if (s_after && p0-1 > s0 && true_word_char(p0[-2])) + continue; + + p1=s1+1; + while (p0 < e0 && p1 < e1 && cs->to_upper[(uint) (uchar) *p0] == + cs->to_upper[(uint) (uchar) *p1]) + p0++, p1++; + if (p1 == e1 && (!e_before || p0 == e0 || !true_word_char(p0[0]))) return 1; } return 0; diff --git a/myisam/ft_parser.c b/myisam/ft_parser.c index c25ed6022a0..f397660af6b 100644 --- a/myisam/ft_parser.c +++ b/myisam/ft_parser.c @@ -105,15 +105,6 @@ FT_WORD * ft_linearize(TREE *wtree) DBUG_RETURN(wlist); } -#define true_word_char(X) (isalnum(X) || (X)=='_') -#ifdef HYPHEN_IS_DELIM -#define misc_word_char(X) ((X)=='\'') -#else -#define misc_word_char(X) ((X)=='\'' || (X)=='-') -#endif -#define word_char(X) (true_word_char(X) || misc_word_char(X)) - - /* returns: * 0 - eof * 1 - word found diff --git a/myisam/ftdefs.h b/myisam/ftdefs.h index 62fa4362e19..46acf60d796 100644 --- a/myisam/ftdefs.h +++ b/myisam/ftdefs.h @@ -22,8 +22,9 @@ #include <m_ctype.h> #include <my_tree.h> -#define HYPHEN_IS_DELIM -#define HYPHEN_IS_CONCAT /* not used for now */ +#define true_word_char(X) (isalnum(X) || (X)=='_') +#define misc_word_char(X) ((X)=='\'') +#define word_char(X) (true_word_char(X) || misc_word_char(X)) #define COMPILE_STOPWORDS_IN |