MySQL Bugs: #72311: SQLCloseCursor Does Not Return Correct Error

Bug #72311	SQLCloseCursor Does Not Return Correct Error
Submitted:	10 Apr 2014 19:25	Modified:	20 Mar 16:49
Reporter:	Lawrenty Novitsky	Email Updates:
Status:	Closed	Impact on me:	None
Category:	Connector / ODBC	Severity:	S3 (Non-critical)
Version:	5.3.2	OS:	Any
Assigned to:		CPU Architecture:	Any

Description:
SQLCloseCursor simply forwards request to SQLFreeStmt(SQL_CLOSE)

But "SQLCloseCursor returns SQLSTATE 24000 (Invalid cursor state) if no cursor is open. Calling SQLCloseCursor is equivalent to calling SQLFreeStmt with the SQL_CLOSE option, with the exception that SQLFreeStmt with SQL_CLOSE has no effect on the application if no cursor is open on the statement, while SQLCloseCursor returns SQLSTATE 24000 (Invalid cursor state)."

(http://msdn.microsoft.com/en-us/library/ms709301%28v=vs.85%29.aspx)

How to repeat:
SQLExecDirect(hstmt, <any ddl statement>);
SQLCloseCursor(hstmt);

Supposed to return SQL_ERROR and 24000 sql state

Suggested fix:
Add check of resultset existence to SQLClosrCursor

Posted by developer:
 
We think the expected behavior of SQLCloseCursor() should be this:

1. If no cursor is open for the statement then report error 24000 (Invalid cursor state) and do nothing more.

2. Otherwise do the same thing as SQLFreeStmt(SQL_CLOSE), which should either succeed or report relevant error.

Posted by developer:
 
The fix for the problem is as follows: SQLCloseCursor() function returns errors each time it is called when no result set is available for the statement.
It gives SQL State: 24000 and error message: Invalid cursor state.
The patch has been pushed into the source tree.

Posted by developer:
 
Added the following entry to the C/ODBC 9.3.0 changelog: 

"If the SQLCloseCursor() method was called when no result set was available, no error was returned. With this patch, the method returned SQL State: 24000 and error message Invalid cursor state in the situation."