Bug#33067

  "Update of CSV row incorrect for some BLOBs"
  when reading in rows, move blob columns into temporary storage not
  allocated by Field_blob class or else row update operation will
  alter original row and make mysql think that nothing has been changed.
fix incrementing wrong statistic values.

mysql-test/r/csv.result

@@ -5379,4 +5379,13 @@ SHOW WARNINGS;
 Level	Code	Message
 Error	1178	The storage engine for the table doesn't support nullable columns
 Error	1005	Can't create table 'test.t1' (errno: 138)
+create table t1 (c1 tinyblob not null) engine=csv;
+insert into t1 values("This");
+update t1 set c1="That" where c1="This";
+affected rows: 1
+info: Rows matched: 1  Changed: 1  Warnings: 0
+select * from t1;
+c1
+That
+drop table t1;
 End of 5.1 tests

mysql-test/t/csv.test

@@ -1773,4 +1773,15 @@ drop table t1;
 CREATE TABLE t1(a INT) ENGINE=CSV;
 SHOW WARNINGS;
 
+#
+# BUG#33067 - .
+#
+create table t1 (c1 tinyblob not null) engine=csv;
+insert into t1 values("This");
+--enable_info            
+update t1 set c1="That" where c1="This";
+--disable_info
+select * from t1;
+drop table t1;
+
 --echo End of 5.1 tests

storage/csv/ha_tina.cc

@@ -56,6 +56,7 @@ TODO:
 #define META_BUFFER_SIZE sizeof(uchar) + sizeof(uchar) + sizeof(ulonglong) \
   + sizeof(ulonglong) + sizeof(ulonglong) + sizeof(ulonglong) + sizeof(uchar)
 #define TINA_CHECK_HEADER 254 // The number we use to determine corruption
+#define BLOB_MEMROOT_ALLOC_SIZE 8192
 
 /* The file extension */
 #define CSV_EXT ".CSV"               // The data file
@@ -597,6 +598,8 @@ int ha_tina::find_current_row(uchar *buf)
   bool read_all;
   DBUG_ENTER("ha_tina::find_current_row");
 
+  free_root(&blobroot, MYF(MY_MARK_BLOCKS_FREE));
+
   /*
     We do not read further then local_saved_data_file_length in order
     not to conflict with undergoing concurrent insert.
@@ -684,6 +687,22 @@ int ha_tina::find_current_row(uchar *buf)
       if ((*field)->store(buffer.ptr(), buffer.length(), buffer.charset(),
                           CHECK_FIELD_WARN))
         goto err;
+      if ((*field)->flags & BLOB_FLAG)
+      {
+        Field_blob *blob= *(Field_blob**) field;
+        uchar *src, *tgt;
+        uint length, packlength;
+        
+        packlength= blob->pack_length_no_ptr();
+        length= blob->get_length(blob->ptr);
+        memcpy_fixed(&src, blob->ptr + packlength, sizeof(char*));
+        if (src)
+        {
+          tgt= (uchar*) alloc_root(&blobroot, length);
+          bmove(tgt, src, length);
+          memcpy_fixed(blob->ptr + packlength, &tgt, sizeof(char*));
+        }
+      }
     }
   }
   next_position= end_offset + eoln_len;
@@ -914,9 +933,10 @@ int ha_tina::open_update_temp_file_if_needed()
 int ha_tina::update_row(const uchar * old_data, uchar * new_data)
 {
   int size;
+  int rc= -1;
   DBUG_ENTER("ha_tina::update_row");
 
-  ha_statistic_increment(&SSV::ha_read_rnd_next_count);
+  ha_statistic_increment(&SSV::ha_update_count);
 
   if (table->timestamp_field_type & TIMESTAMP_AUTO_SET_ON_UPDATE)
     table->timestamp_field->set_time();
@@ -931,20 +951,23 @@ int ha_tina::update_row(const uchar * old_data, uchar * new_data)
     The temp_file_length is used to calculate new data file length.
   */
   if (chain_append())
-    DBUG_RETURN(-1);
+    goto err;
 
   if (open_update_temp_file_if_needed())
-    DBUG_RETURN(-1);
+    goto err;
 
   if (my_write(update_temp_file, (uchar*)buffer.ptr(), size,
                MYF(MY_WME | MY_NABP)))
-    DBUG_RETURN(-1);
+    goto err;
   temp_file_length+= size;
+  rc= 0;
 
   /* UPDATE should never happen on the log tables */
   DBUG_ASSERT(!share->is_log_table);
 
-  DBUG_RETURN(0);
+err:
+  DBUG_PRINT("info",("rc = %d", rc));
+  DBUG_RETURN(rc);
 }
 
 
@@ -1050,6 +1073,8 @@ int ha_tina::rnd_init(bool scan)
   records_is_known= 0;
   chain_ptr= chain;
 
+  init_alloc_root(&blobroot, BLOB_MEMROOT_ALLOC_SIZE, 0);
+
   DBUG_RETURN(0);
 }
 
@@ -1115,7 +1140,7 @@ void ha_tina::position(const uchar *record)
 int ha_tina::rnd_pos(uchar * buf, uchar *pos)
 {
   DBUG_ENTER("ha_tina::rnd_pos");
-  ha_statistic_increment(&SSV::ha_read_rnd_next_count);
+  ha_statistic_increment(&SSV::ha_read_rnd_count);
   current_position= (off_t)my_get_ptr(pos,ref_length);
   DBUG_RETURN(find_current_row(buf));
 }
@@ -1179,6 +1204,7 @@ int ha_tina::rnd_end()
   off_t file_buffer_start= 0;
   DBUG_ENTER("ha_tina::rnd_end");
 
+  free_root(&blobroot, MYF(0));
   records_is_known= 1;
 
   if ((chain_ptr - chain)  > 0)
@@ -1350,6 +1376,8 @@ int ha_tina::repair(THD* thd, HA_CHECK_OPT* check_opt)
   /* set current position to the beginning of the file */
   current_position= next_position= 0;
 
+  init_alloc_root(&blobroot, BLOB_MEMROOT_ALLOC_SIZE, 0);
+
   /* Read the file row-by-row. If everything is ok, repair is not needed. */
   while (!(rc= find_current_row(buf)))
   {
@@ -1358,6 +1386,8 @@ int ha_tina::repair(THD* thd, HA_CHECK_OPT* check_opt)
     current_position= next_position;
   }
 
+  free_root(&blobroot, MYF(0));
+
   my_free((char*)buf, MYF(0));
 
   if (rc == HA_ERR_END_OF_FILE)
@@ -1535,6 +1565,9 @@ int ha_tina::check(THD* thd, HA_CHECK_OPT* check_opt)
   local_saved_data_file_length= share->saved_data_file_length;
   /* set current position to the beginning of the file */
   current_position= next_position= 0;
+
+  init_alloc_root(&blobroot, BLOB_MEMROOT_ALLOC_SIZE, 0);
+
   /* Read the file row-by-row. If everything is ok, repair is not needed. */
   while (!(rc= find_current_row(buf)))
   {
@@ -1543,6 +1576,8 @@ int ha_tina::check(THD* thd, HA_CHECK_OPT* check_opt)
     current_position= next_position;
   }
   
+  free_root(&blobroot, MYF(0));
+
   my_free((char*)buf, MYF(0));
   thd_proc_info(thd, old_proc_info);
 

storage/csv/ha_tina.h

@@ -82,6 +82,7 @@ class ha_tina: public handler
   uint32 chain_size;
   uint local_data_file_version;  /* Saved version of the data file used */
   bool records_is_known;
+  MEM_ROOT blobroot;
 
 private:
   bool get_write_pos(off_t *end_pos, tina_set *closest_hole);