RE: ECPG regression with DECLARE STATEMENT support - Mailing list pgsql-hackers

Dear Rushabh, Michael,

I attached a simple bug-fixing patch.

Could you review it?

An added logic is:

1.     Send a close statement to a backend process regardless of the existence of a cursor.

2.     If ecpg_do function returns false, raise “cursor is invalid” error.

3.     Remove cursor from application.

I already checked this patch passes regression tests and Rushabh’s test code.

Best Regards,