MySQL Bugs: #44703: freeing memory across module boundaries

Bug #44703	freeing memory across module boundaries
Submitted:	6 May 2009 18:23	Modified:	7 May 2009 9:13
Reporter:	Matthias Brantner	Email Updates:
Status:	Verified	Impact on me:	None
Category:	Connector / C++	Severity:	S4 (Feature request)
Version:		OS:	Any
Assigned to:	Assigned Account	CPU Architecture:	Any
Tags:	memory management

Description:
I was just browsing the MySQL Connector documentation for C++ (http://dev.mysql.com/doc/refman/5.1/en/connector-cpp-getting-started-examples.html) and realized that using the connector as a (dynamic) library might cause problems. This is due to the fact the memory that is allocated by the MySQL Connector (e.g. sql::mysql::MySQL_Driver::get_mysql_driver_instance() or sql::Connection::connector) is freed by the user.

Taking the example from the website:

---------------
sql::mysql::MySQL_Driver *driver;
sql::Connection *con;

driver = sql::mysql::MySQL_Driver::get_mysql_driver_instance();
con = driver->connect("tcp://127.0.0.1:3306", "user", "password");

delete con;
---------------

The memory pointed to by the variable con is allocated by the connector and released by the user of the connector. As far as I know, this is not a good practise because freeing memory that is allocated in a different runtime/on a different heap might crash.

Fixing this might also fix the last known issue in Section 21.6.8 (http://dev.mysql.com/doc/refman/5.1/en/connector-cpp-bugs.html): "To avoid potential crashes the build configuration...".

How to repeat:
Build the examples in debug mode on Windows and link against a C standard library that was compiled in release mode (see  Section 21.6.8 of the documentation).

Suggested fix:
Introduce either smart pointers (whose destructors call delete within the library itself) or functions that the user need to call in order to free the memory.

Hi Matthias and thanks for your report.

I agree with your analysis thus setting to verified. Could you just lower the severity to S4 (feature request) if you don't mind as this is a feature request after all.

Ulf?

lowering severity since it's a feature request.

Good feature request. Christian Koch pointed out something similar at http://lists.mysql.com/connector-cplusplus/16 . Let's collect suggestions and see how they may fit into our schedule.