Merge ramayana.hindu.god:/home/tsmith/m/bk/50

into  ramayana.hindu.god:/home/tsmith/m/bk/build/50

9 files changed, 164 insertions, 85 deletions

diff --git a/sql/log.cc b/sql/log.cc
index e66d965c613..0376facb473 100644
--- a/sql/log.cc
+++ b/sql/log.cc
@@ -122,6 +122,20 @@ static int binlog_prepare(THD *thd, bool all)
   return 0;
 }
 
+/**
+  This function is called once after each statement.
+
+  It has the responsibility to flush the transaction cache to the
+  binlog file on commits.
+
+  @param thd   The client thread that executes the transaction.
+  @param all   true if this is the last statement before a COMMIT
+               statement; false if either this is a statement in a
+               transaction but not the last, or if this is a statement
+               not inside a BEGIN block and autocommit is on.
+
+  @see handlerton::commit
+*/
 static int binlog_commit(THD *thd, bool all)
 {
   IO_CACHE *trans_log= (IO_CACHE*)thd->ha_data[binlog_hton.slot];
@@ -134,7 +148,15 @@ static int binlog_commit(THD *thd, bool all)
     // we're here because trans_log was flushed in MYSQL_LOG::log_xid()
     DBUG_RETURN(0);
   }
-  if (all)
+  /*
+    Write commit event if at least one of the following holds:
+     - the user sends an explicit COMMIT; or
+     - the autocommit flag is on, and we are not inside a BEGIN.
+    However, if the user has not sent an explicit COMMIT, and we are
+    either inside a BEGIN or run with autocommit off, then this is not
+    the end of a transaction and we should not write a commit event.
+  */
+  if (all || !(thd->options & (OPTION_NOT_AUTOCOMMIT | OPTION_BEGIN)))
   {
     Query_log_event qev(thd, STRING_WITH_LEN("COMMIT"), TRUE, FALSE);
     qev.error_code= 0; // see comment in MYSQL_LOG::write(THD, IO_CACHE)
@@ -144,6 +166,22 @@ static int binlog_commit(THD *thd, bool all)
     DBUG_RETURN(binlog_end_trans(thd, trans_log, &invisible_commit));
 }
 
+/**
+  This function is called when a transaction involving a transactional
+  table is rolled back.
+
+  It has the responsibility to flush the transaction cache to the
+  binlog file. However, if the transaction does not involve
+  non-transactional tables, nothing needs to be logged.
+
+  @param thd   The client thread that executes the transaction.
+  @param all   true if this is the last statement before a COMMIT
+               statement; false if either this is a statement in a
+               transaction but not the last, or if this is a statement
+               not inside a BEGIN block and autocommit is on.
+
+  @see handlerton::rollback
+*/
 static int binlog_rollback(THD *thd, bool all)
 {
   int error=0;
@@ -1817,9 +1855,11 @@ uint MYSQL_LOG::next_file_id()
   IMPLEMENTATION
     - To support transaction over replication, we wrap the transaction
       with BEGIN/COMMIT or BEGIN/ROLLBACK in the binary log.
-      We want to write a BEGIN/ROLLBACK block when a non-transactional table
-      was updated in a transaction which was rolled back. This is to ensure
-      that the same updates are run on the slave.
+      If a transaction that only involves transactional tables is
+      rolled back, we do not binlog it. However, we write a
+      BEGIN/ROLLBACK block when a non-transactional table was updated
+      in a transaction which was rolled back. This is to ensure that
+      the same updates are run on the slave.
 */
 
 bool MYSQL_LOG::write(THD *thd, IO_CACHE *cache, Log_event *commit_event)
@@ -1837,32 +1877,34 @@ bool MYSQL_LOG::write(THD *thd, IO_CACHE *cache, Log_event *commit_event)
     byte header[LOG_EVENT_HEADER_LEN];
 
     /*
-      Log "BEGIN" at the beginning of the transaction.
-      which may contain more than 1 SQL statement.
+      Log "BEGIN" at the beginning of every transaction.  Here, a
+      transaction is either a BEGIN..COMMIT block or a single
+      statement in autocommit mode.
     */
-    if (thd->options & (OPTION_NOT_AUTOCOMMIT | OPTION_BEGIN))
-    {
-      Query_log_event qinfo(thd, STRING_WITH_LEN("BEGIN"), TRUE, FALSE);
-      /*
-        Imagine this is rollback due to net timeout, after all statements of
-        the transaction succeeded. Then we want a zero-error code in BEGIN.
-        In other words, if there was a really serious error code it's already
-        in the statement's events, there is no need to put it also in this
-        internally generated event, and as this event is generated late it
-        would lead to false alarms.
-        This is safer than thd->clear_error() against kills at shutdown.
-      */
-      qinfo.error_code= 0;
-      /*
-        Now this Query_log_event has artificial log_pos 0. It must be adjusted
-        to reflect the real position in the log. Not doing it would confuse the
-	slave: it would prevent this one from knowing where he is in the
-	master's binlog, which would result in wrong positions being shown to
-	the user, MASTER_POS_WAIT undue waiting etc.
-      */
-      if (qinfo.write(&log_file))
-	goto err;
-    }
+    Query_log_event qinfo(thd, STRING_WITH_LEN("BEGIN"), TRUE, FALSE);
+    /*
+      Imagine this is rollback due to net timeout, after all
+      statements of the transaction succeeded. Then we want a
+      zero-error code in BEGIN.  In other words, if there was a
+      really serious error code it's already in the statement's
+      events, there is no need to put it also in this internally
+      generated event, and as this event is generated late it would
+      lead to false alarms.
+
+      This is safer than thd->clear_error() against kills at shutdown.
+    */
+    qinfo.error_code= 0;
+    /*
+      Now this Query_log_event has artificial log_pos 0. It must be
+      adjusted to reflect the real position in the log. Not doing it
+      would confuse the slave: it would prevent this one from
+      knowing where he is in the master's binlog, which would result
+      in wrong positions being shown to the user, MASTER_POS_WAIT
+      undue waiting etc.
+    */
+    if (qinfo.write(&log_file))
+      goto err;
+
     /* Read from the file used to cache the queries .*/
     if (reinit_io_cache(cache, READ_CACHE, 0, 0, 0))
       goto err;
diff --git a/sql/log_event.cc b/sql/log_event.cc
index d22973d12a3..a950094a018 100644
--- a/sql/log_event.cc
+++ b/sql/log_event.cc
@@ -1862,7 +1862,7 @@ void Query_log_event::print(FILE* file, PRINT_EVENT_INFO* print_event_info)
 {
   print_query_header(file, print_event_info);
   my_fwrite(file, (byte*) query, q_len, MYF(MY_NABP | MY_WME));
-  fprintf(file, "%s\n", print_event_info->delimiter);
+  fprintf(file, "\n%s\n", print_event_info->delimiter);
 }
 #endif /* MYSQL_CLIENT */
 
@@ -3793,6 +3793,7 @@ Xid_log_event(const char* buf,
 #ifndef MYSQL_CLIENT
 bool Xid_log_event::write(IO_CACHE* file)
 {
+  DBUG_EXECUTE_IF("do_not_write_xid", return 0;);
   return write_header(file, sizeof(xid)) ||
          my_b_safe_write(file, (byte*) &xid, sizeof(xid));
 }
@@ -5180,12 +5181,12 @@ void Execute_load_query_log_event::print(FILE* file,
     fprintf(file, " INTO");
     my_fwrite(file, (byte*) query + fn_pos_end, q_len-fn_pos_end,
         MYF(MY_NABP | MY_WME));
-    fprintf(file, "%s\n", print_event_info->delimiter);
+    fprintf(file, "\n%s\n", print_event_info->delimiter);
   }
   else
   {
     my_fwrite(file, (byte*) query, q_len, MYF(MY_NABP | MY_WME));
-    fprintf(file, "%s\n", print_event_info->delimiter);
+    fprintf(file, "\n%s\n", print_event_info->delimiter);
   }
 
   if (!print_event_info->short_form)
diff --git a/sql/set_var.cc b/sql/set_var.cc
index fe57b46745a..a99b063a97e 100644
--- a/sql/set_var.cc
+++ b/sql/set_var.cc
@@ -1448,6 +1448,7 @@ static void fix_trans_mem_root(THD *thd, enum_var_type type)
 static void fix_server_id(THD *thd, enum_var_type type)
 {
   server_id_supplied = 1;
+  thd->server_id= server_id;
 }
 
 
diff --git a/sql/slave.cc b/sql/slave.cc
index 1509916fe91..4a65e9aaa85 100644
--- a/sql/slave.cc
+++ b/sql/slave.cc
@@ -3348,7 +3348,10 @@ static int exec_relay_log_event(THD* thd, RELAY_LOG_INFO* rli)
       if (rli->slave_skip_counter &&
           !((type_code == INTVAR_EVENT ||
              type_code == RAND_EVENT ||
-             type_code == USER_VAR_EVENT) &&
+             type_code == USER_VAR_EVENT ||
+             type_code == BEGIN_LOAD_QUERY_EVENT ||
+             type_code == APPEND_BLOCK_EVENT ||
+             type_code == CREATE_FILE_EVENT) &&
             rli->slave_skip_counter == 1) &&
 #if MYSQL_VERSION_ID < 50100
           /*
diff --git a/sql/sql_acl.cc b/sql/sql_acl.cc
index 134541368e9..8fdd054eb39 100644
--- a/sql/sql_acl.cc
+++ b/sql/sql_acl.cc
@@ -5333,6 +5333,7 @@ bool mysql_create_user(THD *thd, List <LEX_USER> &list)
   LEX_USER *user_name, *tmp_user_name;
   List_iterator <LEX_USER> user_list(list);
   TABLE_LIST tables[GRANT_TABLES];
+  bool some_users_created= FALSE;
   DBUG_ENTER("mysql_create_user");
 
   /* CREATE USER may be skipped on replication client. */
@@ -5361,6 +5362,7 @@ bool mysql_create_user(THD *thd, List <LEX_USER> &list)
       continue;
     }
 
+    some_users_created= TRUE;
     sql_mode= thd->variables.sql_mode;
     if (replace_user_table(thd, tables[0].table, *user_name, 0, 0, 1, 0))
     {
@@ -5371,7 +5373,10 @@ bool mysql_create_user(THD *thd, List <LEX_USER> &list)
 
   VOID(pthread_mutex_unlock(&acl_cache->lock));
 
-  if (mysql_bin_log.is_open())
+  if (result)
+    my_error(ER_CANNOT_USER, MYF(0), "CREATE USER", wrong_users.c_ptr_safe());
+
+  if (some_users_created && mysql_bin_log.is_open())
   {
     Query_log_event qinfo(thd, thd->query, thd->query_length, 0, FALSE);
     mysql_bin_log.write(&qinfo);
@@ -5379,8 +5384,6 @@ bool mysql_create_user(THD *thd, List <LEX_USER> &list)
 
   rw_unlock(&LOCK_grant);
   close_thread_tables(thd);
-  if (result)
-    my_error(ER_CANNOT_USER, MYF(0), "CREATE USER", wrong_users.c_ptr_safe());
   DBUG_RETURN(result);
 }
 
@@ -5405,6 +5408,7 @@ bool mysql_drop_user(THD *thd, List <LEX_USER> &list)
   LEX_USER *user_name, *tmp_user_name;
   List_iterator <LEX_USER> user_list(list);
   TABLE_LIST tables[GRANT_TABLES];
+  bool some_users_deleted= FALSE;
   DBUG_ENTER("mysql_drop_user");
 
   /* DROP USER may be skipped on replication client. */
@@ -5426,7 +5430,9 @@ bool mysql_drop_user(THD *thd, List <LEX_USER> &list)
     {
       append_user(&wrong_users, user_name);
       result= TRUE;
+      continue;
     }
+    some_users_deleted= TRUE;
   }
 
   /* Rebuild 'acl_check_hosts' since 'acl_users' has been modified */
@@ -5440,7 +5446,7 @@ bool mysql_drop_user(THD *thd, List <LEX_USER> &list)
   DBUG_PRINT("info", ("thd->net.last_errno: %d", thd->net.last_errno));
   DBUG_PRINT("info", ("thd->net.last_error: %s", thd->net.last_error));
 
-  if (mysql_bin_log.is_open())
+  if (some_users_deleted && mysql_bin_log.is_open())
   {
     Query_log_event qinfo(thd, thd->query, thd->query_length, 0, FALSE);
     mysql_bin_log.write(&qinfo);
@@ -5473,6 +5479,7 @@ bool mysql_rename_user(THD *thd, List <LEX_USER> &list)
   LEX_USER *user_to, *tmp_user_to;
   List_iterator <LEX_USER> user_list(list);
   TABLE_LIST tables[GRANT_TABLES];
+  bool some_users_renamed= FALSE;
   DBUG_ENTER("mysql_rename_user");
 
   /* RENAME USER may be skipped on replication client. */
@@ -5506,7 +5513,9 @@ bool mysql_rename_user(THD *thd, List <LEX_USER> &list)
     {
       append_user(&wrong_users, user_from);
       result= TRUE;
+      continue;
     }
+    some_users_renamed= TRUE;
   }
   
   /* Rebuild 'acl_check_hosts' since 'acl_users' has been modified */
@@ -5514,7 +5523,10 @@ bool mysql_rename_user(THD *thd, List <LEX_USER> &list)
 
   VOID(pthread_mutex_unlock(&acl_cache->lock));
 
-  if (mysql_bin_log.is_open())
+  if (result)
+    my_error(ER_CANNOT_USER, MYF(0), "RENAME USER", wrong_users.c_ptr_safe());
+  
+  if (some_users_renamed && mysql_bin_log.is_open())
   {
     Query_log_event qinfo(thd, thd->query, thd->query_length, 0, FALSE);
     mysql_bin_log.write(&qinfo);
@@ -5522,8 +5534,6 @@ bool mysql_rename_user(THD *thd, List <LEX_USER> &list)
 
   rw_unlock(&LOCK_grant);
   close_thread_tables(thd);
-  if (result)
-    my_error(ER_CANNOT_USER, MYF(0), "RENAME USER", wrong_users.c_ptr_safe());
   DBUG_RETURN(result);
 }
 
diff --git a/sql/sql_parse.cc b/sql/sql_parse.cc
index 8af79d77fa1..45dac4d5127 100644
--- a/sql/sql_parse.cc
+++ b/sql/sql_parse.cc
@@ -2067,7 +2067,6 @@ bool dispatch_command(enum enum_server_command command, THD *thd,
       unregister_slave(thd,1,1);
       /*  fake COM_QUIT -- if we get here, the thread needs to terminate */
       error = TRUE;
-      net->error = 0;
       break;
     }
 #endif
diff --git a/sql/sql_repl.cc b/sql/sql_repl.cc
index 903d254db8f..5bbff69f197 100644
--- a/sql/sql_repl.cc
+++ b/sql/sql_repl.cc
@@ -1229,9 +1229,6 @@ bool change_master(THD* thd, MASTER_INFO* mi)
       DBUG_RETURN(TRUE);
     }
   }
-  mi->rli.group_master_log_pos = mi->master_log_pos;
-  DBUG_PRINT("info", ("master_log_pos: %lu", (ulong) mi->master_log_pos));
-
   /*
     Coordinates in rli were spoilt by the 'if (need_relay_log_purge)' block,
     so restore them to good values. If we left them to ''/0, that would work;
@@ -1243,6 +1240,7 @@ bool change_master(THD* thd, MASTER_INFO* mi)
     That's why we always save good coords in rli.
   */
   mi->rli.group_master_log_pos= mi->master_log_pos;
+  DBUG_PRINT("info", ("master_log_pos: %lu", (ulong) mi->master_log_pos));
   strmake(mi->rli.group_master_log_name,mi->master_log_name,
 	  sizeof(mi->rli.group_master_log_name)-1);
 
@@ -1355,6 +1353,11 @@ bool mysql_show_binlog_events(THD* thd)
     if ((file=open_binlog(&log, linfo.log_file_name, &errmsg)) < 0)
       goto err;
 
+    /*
+      to account binlog event header size
+    */
+    thd->variables.max_allowed_packet += MAX_LOG_EVENT_HEADER;
+
     pthread_mutex_lock(log_lock);
 
     /*
@@ -1365,7 +1368,6 @@ bool mysql_show_binlog_events(THD* thd)
        This code will fail on a mixed relay log (one which has Format_desc then
        Rotate then Format_desc).
     */
-
     ev = Log_event::read_log_event(&log,(pthread_mutex_t*)0,description_event);
     if (ev)
     {
@@ -1556,37 +1558,52 @@ err:
   DBUG_RETURN(TRUE);
 }
 
-
+/**
+   Load data's io cache specific hook to be executed
+   before a chunk of data is being read into the cache's buffer
+   The fuction instantianates and writes into the binlog
+   replication events along LOAD DATA processing.
+   
+   @param file  pointer to io-cache
+   @return 0
+*/
 int log_loaded_block(IO_CACHE* file)
 {
+  DBUG_ENTER("log_loaded_block");
   LOAD_FILE_INFO *lf_info;
-  uint block_len ;
-
-  /* file->request_pos contains position where we started last read */
-  char* buffer = (char*) file->request_pos;
-  if (!(block_len = (char*) file->read_end - (char*) buffer))
-    return 0;
-  lf_info = (LOAD_FILE_INFO*) file->arg;
+  uint block_len;
+  /* buffer contains position where we started last read */
+  char* buffer= (char*) my_b_get_buffer_start(file);
+  uint max_event_size= current_thd->variables.max_allowed_packet;
+  lf_info= (LOAD_FILE_INFO*) file->arg;
   if (lf_info->last_pos_in_file != HA_POS_ERROR &&
-      lf_info->last_pos_in_file >= file->pos_in_file)
-    return 0;
-  lf_info->last_pos_in_file = file->pos_in_file;
-  if (lf_info->wrote_create_file)
-  {
-    Append_block_log_event a(lf_info->thd, lf_info->thd->db, buffer,
-                             block_len, lf_info->log_delayed);
-    mysql_bin_log.write(&a);
-  }
-  else
+      lf_info->last_pos_in_file >= my_b_get_pos_in_file(file))
+    DBUG_RETURN(0);
+  
+  for (block_len= my_b_get_bytes_in_buffer(file); block_len > 0;
+       buffer += min(block_len, max_event_size),
+       block_len -= min(block_len, max_event_size))
   {
-    Begin_load_query_log_event b(lf_info->thd, lf_info->thd->db,
-                                 buffer, block_len,
-                                 lf_info->log_delayed);
-    mysql_bin_log.write(&b);
-    lf_info->wrote_create_file = 1;
-    DBUG_SYNC_POINT("debug_lock.created_file_event",10);
+    lf_info->last_pos_in_file= my_b_get_pos_in_file(file);
+    if (lf_info->wrote_create_file)
+    {
+      Append_block_log_event a(lf_info->thd, lf_info->thd->db, buffer,
+                               min(block_len, max_event_size),
+                               lf_info->log_delayed);
+      mysql_bin_log.write(&a);
+    }
+    else
+    {
+      Begin_load_query_log_event b(lf_info->thd, lf_info->thd->db,
+                                   buffer,
+                                   min(block_len, max_event_size),
+                                   lf_info->log_delayed);
+      mysql_bin_log.write(&b);
+      lf_info->wrote_create_file= 1;
+      DBUG_SYNC_POINT("debug_lock.created_file_event",10);
+    }
   }
-  return 0;
+  DBUG_RETURN(0);
 }
 
 #endif /* HAVE_REPLICATION */
diff --git a/sql/sql_string.cc b/sql/sql_string.cc
index 606a9ddb26d..75e47dd0c8e 100644
--- a/sql/sql_string.cc
+++ b/sql/sql_string.cc
@@ -307,8 +307,8 @@ bool String::copy_aligned(const char *str,uint32 arg_length, uint32 offset,
     return TRUE;
   
   /*
-    Note, this is only safe for little-endian UCS-2.
-    If we add big-endian UCS-2 sometimes, this code
+    Note, this is only safe for big-endian UCS-2.
+    If we add little-endian UCS-2 sometimes, this code
     will be more complicated. But it's OK for now.
   */
   bzero((char*) Ptr, offset);
diff --git a/sql/sql_view.cc b/sql/sql_view.cc
index 297edd0d90d..dd0b92a06f8 100644
--- a/sql/sql_view.cc
+++ b/sql/sql_view.cc
@@ -1424,6 +1424,8 @@ bool mysql_drop_view(THD *thd, TABLE_LIST *views, enum_drop_mode drop_mode)
   String non_existant_views;
   char *wrong_object_db= NULL, *wrong_object_name= NULL;
   bool error= FALSE;
+  bool some_views_deleted= FALSE;
+  bool something_wrong= FALSE;
 
   VOID(pthread_mutex_lock(&LOCK_open));
   for (view= views; view; view= view->next_local)
@@ -1462,33 +1464,37 @@ bool mysql_drop_view(THD *thd, TABLE_LIST *views, enum_drop_mode drop_mode)
     }
     if (my_delete(path, MYF(MY_WME)))
       error= TRUE;
+    some_views_deleted= TRUE;
     query_cache_invalidate3(thd, view, 0);
     sp_cache_invalidate();
   }
-  if (mysql_bin_log.is_open())
-  {
-    thd->clear_error();
-    Query_log_event qinfo(thd, thd->query, thd->query_length, 0, FALSE);
-    mysql_bin_log.write(&qinfo);
-  }
-
-  VOID(pthread_mutex_unlock(&LOCK_open));
 
-  if (error)
-  {
-    DBUG_RETURN(TRUE);
-  }
   if (wrong_object_name)
   {
     my_error(ER_WRONG_OBJECT, MYF(0), wrong_object_db, wrong_object_name, 
              "VIEW");
-    DBUG_RETURN(TRUE);
   }
   if (non_existant_views.length())
   {
     my_error(ER_BAD_TABLE_ERROR, MYF(0), non_existant_views.c_ptr());
+  }
+
+  something_wrong= error || wrong_object_name || non_existant_views.length();
+  if (some_views_deleted || !something_wrong)
+  {
+    if (!something_wrong)
+      thd->clear_error();
+    Query_log_event qinfo(thd, thd->query, thd->query_length, 0, FALSE);
+    mysql_bin_log.write(&qinfo);
+  }
+
+  VOID(pthread_mutex_unlock(&LOCK_open));
+
+  if (something_wrong)
+  {
     DBUG_RETURN(TRUE);
   }
+
   send_ok(thd);
   DBUG_RETURN(FALSE);
 }