BUG#23080148 - BACKPORT BUG 14653594 AND BUG 20683959 TO

               MYSQL-5.5

The bug asks for a backport of bug#1463594 and bug#20682959. This
is required because of the fact that if replication is enabled, master
transaction can commit whereas slave can't commit due to not exact
'enviroment'. This manifestation is seen in bug#22024200.

mysql-test/r/loaddata.result

@@ -507,7 +507,7 @@ DROP TABLE t1;
 # Bug#11765139  58069: LOAD DATA INFILE: VALGRIND REPORTS INVALID MEMORY READS AND WRITES WITH U
 #
 CREATE TABLE t1(f1 INT);
-SELECT 0xE1BB30 INTO OUTFILE 't1.dat';
+SELECT 0xE1C330 INTO OUTFILE 't1.dat';
 LOAD DATA INFILE 't1.dat' IGNORE INTO TABLE t1 CHARACTER SET utf8;
 DROP TABLE t1;
 #
@@ -532,3 +532,27 @@ FIELDS TERMINATED BY 't' LINES TERMINATED BY '';
 Got one of the listed errors
 SET @@sql_mode= @old_mode;
 DROP TABLE t1;
+
+#
+#  Bug#23080148 - Backport of Bug#20683959.
+#  Bug#20683959 LOAD DATA INFILE IGNORES A SPECIFIC ROW SILENTLY
+#               UNDER DB CHARSET IS UTF8.
+#
+CREATE DATABASE d1 CHARSET latin1;
+USE d1;
+CREATE TABLE t1 (val TEXT);
+LOAD DATA INFILE '../../std_data/bug20683959loaddata.txt' INTO TABLE t1;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+1
+SELECT HEX(val) FROM t1;
+HEX(val)
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
+CREATE DATABASE d2 CHARSET utf8;
+USE d2;
+CREATE TABLE t1 (val TEXT);
+LOAD DATA INFILE '../../std_data/bug20683959loaddata.txt' INTO TABLE t1;
+ERROR HY000: Invalid utf8 character string: '"RT @niouzechun: \9058\221A'
+DROP TABLE d1.t1, d2.t1;
+DROP DATABASE d1;
+DROP DATABASE d2;

mysql-test/std_data/bug20683959loaddata.txt

+Ã"RT @niouzechun: 遘√繝上ャ繝斐繧ｨ繝ｳ繝牙耳縺ｪ繧薙□縺代ｌ縺ｩ縲√い繝ｳ繝上ャ繝斐繧ｨ繝ｳ繝峨ｄ諠ｨ蜉噪縺ｪ縺願ｩｱ繧偵≠縺ｾ繧顔ｩ肴･ｵ逧↓鞫ょ叙縺励↑縺炊逕ｱ縺ｯ縲∫樟螳溘莠ｺ逕溘蝓ｺ譛ｬ逧↓縺∪縺上＞縺九↑縺＠荳榊ｹｳ遲峨□縺礼炊荳榊ｰｽ縺縺苓ｾ帙＞

mysql-test/t/loaddata.test

@@ -610,7 +610,7 @@ disconnect con1;
 --echo #
 
 CREATE TABLE t1(f1 INT);
-EVAL SELECT 0xE1BB30 INTO OUTFILE 't1.dat';
+EVAL SELECT 0xE1C330 INTO OUTFILE 't1.dat';
 --disable_warnings
 LOAD DATA INFILE 't1.dat' IGNORE INTO TABLE t1 CHARACTER SET utf8;
 --enable_warnings
@@ -656,3 +656,26 @@ SET @@sql_mode= @old_mode;
 --remove_file $MYSQLTEST_VARDIR/mysql
 DROP TABLE t1;
 
+--echo
+--echo #
+--echo #  Bug#23080148 - Backport of Bug#20683959.
+--echo #  Bug#20683959 LOAD DATA INFILE IGNORES A SPECIFIC ROW SILENTLY
+--echo #               UNDER DB CHARSET IS UTF8.
+--echo #
+
+CREATE DATABASE d1 CHARSET latin1;
+USE d1;
+CREATE TABLE t1 (val TEXT);
+LOAD DATA INFILE '../../std_data/bug20683959loaddata.txt' INTO TABLE t1;
+SELECT COUNT(*) FROM t1;
+SELECT HEX(val) FROM t1;
+
+CREATE DATABASE d2 CHARSET utf8;
+USE d2;
+CREATE TABLE t1 (val TEXT);
+--error ER_INVALID_CHARACTER_STRING
+LOAD DATA INFILE '../../std_data/bug20683959loaddata.txt' INTO TABLE t1;
+
+DROP TABLE d1.t1, d2.t1;
+DROP DATABASE d1;
+DROP DATABASE d2;

sql/sql_load.cc

@@ -1363,8 +1363,8 @@ READ_INFO::READ_INFO(File file_par, uint tot_length, CHARSET_INFO *cs,
   set_if_bigger(length,line_start.length());
   stack=stack_pos=(int*) sql_alloc(sizeof(int)*length);
 
-  if (!(buffer=(uchar*) my_malloc(buff_length+1,MYF(0))))
-    error=1; /* purecov: inspected */
+  if (!(buffer=(uchar*) my_malloc(buff_length+1,MYF(MY_WME))))
+    error= true; /* purecov: inspected */
   else
   {
     end_of_buff=buffer+buff_length;
@@ -1556,37 +1556,50 @@ int READ_INFO::read_field()
 	}
       }
 #ifdef USE_MB
-      if (my_mbcharlen(read_charset, chr) > 1 &&
-          to + my_mbcharlen(read_charset, chr) <= end_of_buff)
-      {
-        uchar* p= to;
-        int ml, i;
-        *to++ = chr;
-
-        ml= my_mbcharlen(read_charset, chr);
+        uint ml= my_mbcharlen(read_charset, chr);
+        if (ml == 0)
+        {
+          *to= '\0';
+          my_error(ER_INVALID_CHARACTER_STRING, MYF(0),
+                   read_charset->csname, buffer);
+          error= true;
+          return 1;
+        }
 
-        for (i= 1; i < ml; i++) 
+        if (ml > 1 &&
+            to + ml <= end_of_buff)
         {
-          chr= GET;
-          if (chr == my_b_EOF)
+          uchar* p= to;
+          *to++ = chr;
+
+          for (uint i= 1; i < ml; i++)
           {
-            /*
-             Need to back up the bytes already ready from illformed
-             multi-byte char 
-            */
-            to-= i;
-            goto found_eof;
+            chr= GET;
+            if (chr == my_b_EOF)
+            {
+              /*
+                Need to back up the bytes already ready from illformed
+                multi-byte char 
+              */
+              to-= i;
+              goto found_eof;
+            }
+            *to++ = chr;
           }
-          *to++ = chr;
-        }
-        if (my_ismbchar(read_charset,
+          if (my_ismbchar(read_charset,
                         (const char *)p,
                         (const char *)to))
-          continue;
-        for (i= 0; i < ml; i++)
-          PUSH(*--to);
-        chr= GET;
-      }
+            continue;
+          for (uint i= 0; i < ml; i++)
+            PUSH(*--to);
+          chr= GET;
+        }
+        else if (ml > 1)
+        {
+          // Buffer is too small, exit while loop, and reallocate.
+          PUSH(chr);
+          break;
+        }
 #endif
       *to++ = (uchar) chr;
     }
@@ -1830,7 +1843,15 @@ int READ_INFO::read_value(int delim, String *val)
   for (chr= GET; my_tospace(chr) != delim && chr != my_b_EOF;)
   {
 #ifdef USE_MB
-    if (my_mbcharlen(read_charset, chr) > 1)
+    uint ml= my_mbcharlen(read_charset, chr);
+    if (ml == 0)
+    {
+      chr= my_b_EOF;
+      val->length(0);
+      return chr;
+    }
+
+    if (ml > 1)
     {
       DBUG_PRINT("read_xml",("multi byte"));
       int i, ml= my_mbcharlen(read_charset, chr);