Bug#25411 (trigger code truncated), PART I

The issue found with bug 25411 is due to the function skip_rear_comments()
which damages the source code while implementing a work around.
The root cause of the problem is in the lexical analyser, which does not
process special comments properly.
For special comments like :
[1] aaa /*!50000 bbb */ ccc
since 5.0 is a version older that the current code, the parser is in lining
the content of the special comment, so that the query to process is
[2] aaa bbb ccc
However, the text of the query captured when processing a stored procedure,
stored function or trigger (or event in 5.1), can be after rebuilding it:
[3] aaa bbb */ ccc
which is wrong.

To fix bug 25411 properly, the lexical analyser needs to return [2] when
in lining special comments.
In order to implement this, some preliminary cleanup is required in the code,
which is implemented by this patch.

Before this change, the structure named LEX (or st_lex) contains attributes
that belong to lexical analysis, as well as attributes that represents the
abstract syntax tree (AST) of a statement.
Creating a new LEX structure for each statements (which makes sense for the
AST part) also re-initialized the lexical analysis phase each time, which
is conceptually wrong.

With this patch, the previous st_lex structure has been split in two:
- st_lex represents the Abstract Syntax Tree for a statement. The name "lex"
has not been changed to avoid a bigger impact in the code base.
- class lex_input_stream represents the internal state of the lexical
  analyser, which by definition should *not* be reinitialized when parsing
  multiple statements from the same input stream.

This change is a pre-requisite for bug 25411, since the implementation of
lex_input_stream will later improve to deal properly with special comments,
and this processing can not be done with the current implementation of
sp_head::reset_lex and sp_head::restore_lex, which interfere with the lexer.

This change set alone does not fix bug 25411.


sql/item_func.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/log_event.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/mysql_priv.h:
  Refactoring, separate lex_input_stream from st_lex.
sql/slave.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sp.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sp_head.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sp_head.h:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_class.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_class.h:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_lex.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_lex.h:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_parse.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_prepare.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_trigger.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_view.cc:
  Refactoring, separate lex_input_stream from st_lex.
sql/sql_yacc.yy:
  Refactoring, separate lex_input_stream from st_lex.

1 files changed, 70 insertions, 33 deletions

diff --git a/sql/sql_parse.cc b/sql/sql_parse.cc
index 1b8bfd38fc4..36ae5ca4a31 100644
--- a/sql/sql_parse.cc
+++ b/sql/sql_parse.cc
@@ -1236,6 +1236,7 @@ pthread_handler_t handle_bootstrap(void *arg)
   THD *thd=(THD*) arg;
   FILE *file=bootstrap_file;
   char *buff;
+  const char* found_semicolon= NULL;
 
   /* The following must be called before DBUG_ENTER */
   thd->thread_stack= (char*) &thd;
@@ -1312,7 +1313,7 @@ pthread_handler_t handle_bootstrap(void *arg)
     */
     thd->query_id=next_query_id();
     thd->set_time();
-    mysql_parse(thd,thd->query,length);
+    mysql_parse(thd, thd->query, length, & found_semicolon);
     close_thread_tables(thd);			// Free tables
 
     if (thd->is_fatal_error)
@@ -1789,17 +1790,19 @@ bool dispatch_command(enum enum_server_command command, THD *thd,
     char *packet_end= thd->query + thd->query_length;
     /* 'b' stands for 'buffer' parameter', special for 'my_snprintf' */
     const char *format= "%.*b";
+    const char* found_semicolon= NULL;
+
     mysql_log.write(thd,command, format, thd->query_length, thd->query);
     DBUG_PRINT("query",("%-.4096s",thd->query));
 
     if (!(specialflag & SPECIAL_NO_PRIOR))
       my_pthread_setprio(pthread_self(),QUERY_PRIOR);
 
-    mysql_parse(thd,thd->query, thd->query_length);
+    mysql_parse(thd, thd->query, thd->query_length, & found_semicolon);
 
-    while (!thd->killed && thd->lex->found_semicolon && !thd->net.report_error)
+    while (!thd->killed && found_semicolon && !thd->net.report_error)
     {
-      char *next_packet= thd->lex->found_semicolon;
+      char *next_packet= (char*) found_semicolon;
       net->no_send_error= 0;
       /*
         Multiple queries exits, execute them individually
@@ -1824,7 +1827,7 @@ bool dispatch_command(enum enum_server_command command, THD *thd,
       thd->set_time(); /* Reset the query start time. */
       /* TODO: set thd->lex->sql_command to SQLCOM_END here */
       VOID(pthread_mutex_unlock(&LOCK_thread_count));
-      mysql_parse(thd, next_packet, length);
+      mysql_parse(thd, next_packet, length, & found_semicolon);
     }
 
     if (!(specialflag & SPECIAL_NO_PRIOR))
@@ -1845,7 +1848,7 @@ bool dispatch_command(enum enum_server_command command, THD *thd,
     LEX_STRING conv_name;
 
     /* used as fields initializator */
-    lex_start(thd, 0, 0);
+    lex_start(thd);
 
     statistic_increment(thd->status_var.com_stat[SQLCOM_SHOW_FIELDS],
 			&LOCK_status);
@@ -1882,7 +1885,10 @@ bool dispatch_command(enum enum_server_command command, THD *thd,
       break;
     /* init structures for VIEW processing */
     table_list.select_lex= &(thd->lex->select_lex);
-    mysql_init_query(thd, (uchar*)"", 0);
+
+    lex_start(thd);
+    mysql_reset_thd_for_next_command(thd);
+
     thd->lex->
       select_lex.table_list.link_in_list((byte*) &table_list,
                                          (byte**) &table_list.next_local);
@@ -5687,20 +5693,6 @@ bool my_yyoverflow(short **yyss, YYSTYPE **yyvs, ulong *yystacksize)
 }
 
 
-/****************************************************************************
-  Initialize global thd variables needed for query
-****************************************************************************/
-
-void
-mysql_init_query(THD *thd, uchar *buf, uint length)
-{
-  DBUG_ENTER("mysql_init_query");
-  lex_start(thd, buf, length);
-  mysql_reset_thd_for_next_command(thd);
-  DBUG_VOID_RETURN;
-}
-
-
 /*
  Reset THD part responsible for command processing state.
 
@@ -5887,21 +5879,54 @@ void mysql_init_multi_delete(LEX *lex)
   mysql_test_parse_for_slave() in this same file.
 */
 
-void mysql_parse(THD *thd, char *inBuf, uint length)
+/**
+  Parse a query.
+  @param thd Current thread
+  @param inBuf Begining of the query text
+  @param length Length of the query text
+  @param [out] semicolon For multi queries, position of the character of
+  the next query in the query text.
+*/
+
+void mysql_parse(THD *thd, const char *inBuf, uint length,
+                 const char ** found_semicolon)
 {
   DBUG_ENTER("mysql_parse");
 
   DBUG_EXECUTE_IF("parser_debug", turn_parser_debug_on(););
 
-  mysql_init_query(thd, (uchar*) inBuf, length);
-  if (query_cache_send_result_to_client(thd, inBuf, length) <= 0)
+  /*
+    Warning.
+    The purpose of query_cache_send_result_to_client() is to lookup the
+    query in the query cache first, to avoid parsing and executing it.
+    So, the natural implementation would be to:
+    - first, call query_cache_send_result_to_client,
+    - second, if caching failed, initialise the lexical and syntactic parser.
+    The problem is that the query cache depends on a clean initialization
+    of the thd and thd->lex structures, which happen to be implemented
+    by:
+    - lex_start()
+    - mysql_reset_thd_for_next_command()
+    So, initializing the lexical analyser *before* using the query cache
+    is required for the cache to work properly.
+    FIXME: cleanup the dependencies in the code to simplify this.
+  */
+  Lex_input_stream lip(thd, inBuf, length);
+  thd->m_lip= &lip;
+  lex_start(thd);
+  mysql_reset_thd_for_next_command(thd);
+
+  if (query_cache_send_result_to_client(thd, (char*) inBuf, length) <= 0)
   {
     LEX *lex= thd->lex;
-    
+
     sp_cache_flush_obsolete(&thd->sp_proc_cache);
     sp_cache_flush_obsolete(&thd->sp_func_cache);
-    
-    if (!MYSQLparse((void *)thd) && ! thd->is_fatal_error)
+
+    int err= MYSQLparse(thd);
+    *found_semicolon= lip.found_semicolon;
+
+    if (!err && ! thd->is_fatal_error)
     {
 #ifndef NO_EMBEDDED_ACCESS_CHECKS
       if (mqh_used && thd->user_connect &&
@@ -5924,8 +5949,8 @@ void mysql_parse(THD *thd, char *inBuf, uint length)
             PROCESSLIST.
             Note that we don't need LOCK_thread_count to modify query_length.
           */
-          if (lex->found_semicolon &&
-              (thd->query_length= (ulong)(lex->found_semicolon - thd->query)))
+          if (lip.found_semicolon &&
+              (thd->query_length= (ulong)(lip.found_semicolon - thd->query)))
             thd->query_length--;
           /* Actually execute the query */
 	  mysql_execute_command(thd);
@@ -5952,6 +5977,12 @@ void mysql_parse(THD *thd, char *inBuf, uint length)
     thd->cleanup_after_query();
     DBUG_ASSERT(thd->change_list.is_empty());
   }
+  else
+  {
+    /* There are no multi queries in the cache. */
+    *found_semicolon= NULL;
+  }
+
   DBUG_VOID_RETURN;
 }
 
@@ -5972,8 +6003,13 @@ bool mysql_test_parse_for_slave(THD *thd, char *inBuf, uint length)
   bool error= 0;
   DBUG_ENTER("mysql_test_parse_for_slave");
 
-  mysql_init_query(thd, (uchar*) inBuf, length);
-  if (!MYSQLparse((void*) thd) && ! thd->is_fatal_error &&
+  Lex_input_stream lip(thd, inBuf, length);
+  thd->m_lip= &lip;
+  lex_start(thd);
+  mysql_reset_thd_for_next_command(thd);
+  int err= MYSQLparse((void*) thd);
+
+  if (!err && ! thd->is_fatal_error &&
       all_tables_not_ok(thd,(TABLE_LIST*) lex->select_lex.table_list.first))
     error= 1;                  /* Ignore question */
   thd->end_statement();
@@ -7032,8 +7068,9 @@ bool check_simple_select()
   if (lex->current_select != &lex->select_lex)
   {
     char command[80];
-    strmake(command, lex->yylval->symbol.str,
-	    min(lex->yylval->symbol.length, sizeof(command)-1));
+    Lex_input_stream *lip= thd->m_lip;
+    strmake(command, lip->yylval->symbol.str,
+	    min(lip->yylval->symbol.length, sizeof(command)-1));
     my_error(ER_CANT_USE_OPTION_HERE, MYF(0), command);
     return 1;
   }