UTF8 now process space as PAD character correctly.

5 files changed, 170 insertions, 8 deletions

diff --git a/mysql-test/r/compare.result b/mysql-test/r/compare.result
index 3ad3465fac0..bf8a5106044 100644
--- a/mysql-test/r/compare.result
+++ b/mysql-test/r/compare.result
@@ -12,3 +12,27 @@ select * from t1;
 id
 000000000001
 drop table t1;
+SELECT 'a' = 'a ';
+'a' = 'a '
+1
+SELECT 'a\0' < 'a';
+'a\0' < 'a'
+1
+SELECT 'a\0' < 'a ';
+'a\0' < 'a '
+1
+SELECT 'a\t' < 'a';
+'a\t' < 'a'
+1
+SELECT 'a\t' < 'a ';
+'a\t' < 'a '
+1
+CREATE TABLE t1 (a char(10) not null);
+INSERT INTO t1 VALUES ('a'),('a\0'),('a\t'),('a ');
+SELECT hex(a),STRCMP(a,'a'), STRCMP(a,'a ') FROM t1;
+hex(a)	STRCMP(a,'a')	STRCMP(a,'a ')
+61	0	0
+6100	-1	-1
+6109	-1	-1
+61	0	0
+DROP TABLE t1;
diff --git a/mysql-test/r/ctype_utf8.result b/mysql-test/r/ctype_utf8.result
index 33a7e91fbeb..28af71b7681 100644
--- a/mysql-test/r/ctype_utf8.result
+++ b/mysql-test/r/ctype_utf8.result
@@ -63,6 +63,30 @@ select 'A' like 'a' collate utf8_bin;
 select _utf8 0xD0B0D0B1D0B2 like concat(_utf8'%',_utf8 0xD0B1,_utf8 '%');
 _utf8 0xD0B0D0B1D0B2 like concat(_utf8'%',_utf8 0xD0B1,_utf8 '%')
 1
+SELECT 'a' = 'a ';
+'a' = 'a '
+1
+SELECT 'a\0' < 'a';
+'a\0' < 'a'
+1
+SELECT 'a\0' < 'a ';
+'a\0' < 'a '
+1
+SELECT 'a\t' < 'a';
+'a\t' < 'a'
+1
+SELECT 'a\t' < 'a ';
+'a\t' < 'a '
+1
+CREATE TABLE t1 (a char(10) character set utf8 not null);
+INSERT INTO t1 VALUES ('a'),('a\0'),('a\t'),('a ');
+SELECT hex(a),STRCMP(a,'a'), STRCMP(a,'a ') FROM t1;
+hex(a)	STRCMP(a,'a')	STRCMP(a,'a ')
+61	0	0
+6100	-1	-1
+6109	-1	-1
+61	0	0
+DROP TABLE t1;
 select insert('txs',2,1,'hi'),insert('is ',4,0,'a'),insert('txxxxt',2,4,'es');
 insert('txs',2,1,'hi')	insert('is ',4,0,'a')	insert('txxxxt',2,4,'es')
 this	is a	test
diff --git a/mysql-test/t/compare.test b/mysql-test/t/compare.test
index e5a2e310866..b0cef48dd3f 100644
--- a/mysql-test/t/compare.test
+++ b/mysql-test/t/compare.test
@@ -13,3 +13,20 @@ select * from t1 where id=000000000001;
 delete from t1 where id=000000000002;
 select * from t1;
 drop table t1;
+
+#
+# Check the following:
+# "a"  == "a "
+# "a\0" < "a"
+# "a\0" < "a "
+
+SELECT 'a' = 'a ';
+SELECT 'a\0' < 'a';
+SELECT 'a\0' < 'a ';
+SELECT 'a\t' < 'a';
+SELECT 'a\t' < 'a ';
+
+CREATE TABLE t1 (a char(10) not null);
+INSERT INTO t1 VALUES ('a'),('a\0'),('a\t'),('a ');
+SELECT hex(a),STRCMP(a,'a'), STRCMP(a,'a ') FROM t1;
+DROP TABLE t1;
diff --git a/mysql-test/t/ctype_utf8.test b/mysql-test/t/ctype_utf8.test
index 964ac84fbbe..4e68efeffc3 100644
--- a/mysql-test/t/ctype_utf8.test
+++ b/mysql-test/t/ctype_utf8.test
@@ -34,6 +34,23 @@ select 'A' like 'a' collate utf8_bin;
 select _utf8 0xD0B0D0B1D0B2 like concat(_utf8'%',_utf8 0xD0B1,_utf8 '%');
 
 #
+# Check the following:
+# "a"  == "a "
+# "a\0" < "a"
+# "a\0" < "a "
+
+SELECT 'a' = 'a ';
+SELECT 'a\0' < 'a';
+SELECT 'a\0' < 'a ';
+SELECT 'a\t' < 'a';
+SELECT 'a\t' < 'a ';
+
+CREATE TABLE t1 (a char(10) character set utf8 not null);
+INSERT INTO t1 VALUES ('a'),('a\0'),('a\t'),('a ');
+SELECT hex(a),STRCMP(a,'a'), STRCMP(a,'a ') FROM t1;
+DROP TABLE t1;
+
+#
 # Fix this, it should return 1:
 #
 #select _utf8 0xD0B0D0B1D0B2 like concat(_utf8'%',_utf8 0xD091,_utf8 '%');
diff --git a/strings/ctype-utf8.c b/strings/ctype-utf8.c
index 886ecfbd0c9..82787f2b65f 100644
--- a/strings/ctype-utf8.c
+++ b/strings/ctype-utf8.c
@@ -1837,18 +1837,98 @@ static int my_strnncoll_utf8(CHARSET_INFO *cs,
 }
 
 
+
 /*
-  TODO: Has to be fixed as strnncollsp in ctype-simple
+  Compare strings, discarding end space
+
+  SYNOPSIS
+    my_strnncollsp_utf8()
+    cs			character set handler
+    a			First string to compare
+    a_length		Length of 'a'
+    b			Second string to compare
+    b_length		Length of 'b'
+
+  IMPLEMENTATION
+    If one string is shorter as the other, then we space extend the other
+    so that the strings have equal length.
+
+    This will ensure that the following things hold:
+
+    "a"  == "a "
+    "a\0" < "a"
+    "a\0" < "a "
+
+  RETURN
+    < 0	 a <  b
+    = 0	 a == b
+    > 0	 a > b
 */
 
-static
-int my_strnncollsp_utf8(CHARSET_INFO * cs, 
-			const uchar *s, uint slen, 
-			const uchar *t, uint tlen)
+static int my_strnncollsp_utf8(CHARSET_INFO *cs, 
+			     const uchar *s, uint slen,
+			     const uchar *t, uint tlen)
 {
-  for ( ; slen && s[slen-1] == ' ' ; slen--);
-  for ( ; tlen && t[tlen-1] == ' ' ; tlen--);
-  return my_strnncoll_utf8(cs,s,slen,t,tlen);
+  int s_res,t_res;
+  my_wc_t s_wc,t_wc;
+  const uchar *se= s+slen;
+  const uchar *te= t+tlen;
+  
+  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]); 
+    }
+    
+    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 0;
 }