From 4bfd2121776d0ce7ebd6761b807f083da87cb185 Mon Sep 17 00:00:00 2001 From: Jorgen Loland Date: Mon, 15 Nov 2010 16:18:04 +0100 Subject: Bug#54812: assert in Diagnostics_area::set_ok_status during EXPLAIN Before the patch, send_eof() of some subclasses of select_result (e.g., select_send::send_eof()) could handle being called after an error had occured while others could not. The methods that were not well-behaved would trigger an ASSERT on debug builds. Release builds were not affected. Consider the following query as an example for how the ASSERT could be triggered: A user without execute privilege on f() does SELECT MAX(key1) INTO @dummy FROM t1 WHERE f() < 1; resulting in "ERROR 42000: execute command denied to user..." The server would end the query by calling send_eof(). The fact that the error had occured would make the ASSERT trigger. select_dumpvar::send_eof() was the offending method in the bug report, but the problem also applied to other subclasses of select_result. This patch uniforms send_eof() of all subclasses of select_result to handle being called after an error has occured. --- sql/sql_insert.cc | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'sql/sql_insert.cc') diff --git a/sql/sql_insert.cc b/sql/sql_insert.cc index 1f8da3fab5c..a81c9ae15a0 100644 --- a/sql/sql_insert.cc +++ b/sql/sql_insert.cc @@ -3506,6 +3506,9 @@ bool select_insert::send_eof() error= (thd->locked_tables_mode <= LTM_LOCK_TABLES ? table->file->ha_end_bulk_insert() : 0); + if (!error && thd->is_error()) + error= thd->stmt_da->sql_errno(); + table->file->extra(HA_EXTRA_NO_IGNORE_DUP_KEY); table->file->extra(HA_EXTRA_WRITE_CANNOT_REPLACE); @@ -4049,7 +4052,7 @@ bool select_create::send_eof() { bool tmp=select_insert::send_eof(); if (tmp) - abort(); + abort_result_set(); else { /* @@ -4081,7 +4084,7 @@ void select_create::abort_result_set() DBUG_ENTER("select_create::abort_result_set"); /* - In select_insert::abort() we roll back the statement, including + In select_insert::abort_result_set() we roll back the statement, including truncating the transaction cache of the binary log. To do this, we pretend that the statement is transactional, even though it might be the case that it was not. -- cgit v1.2.1