On 01/29/2015 09:36 AM, Michael Paquier wrote:
> IMO, this behavior should be properly documented in the section of
> UseServerSidePrepare, aka there is one case where server-side prepare
> is used with NULL values even if this parameter is disabled.
Added a notice about that. Not sure if it's makes much sense to a user
who's not familiar with the FE/BE protoocl, but anyway..
> The header of this file is incorrect, DisallowPremature does not exist
> anymore with your patch and this test is made to test if premature
> calls of SQLNumResultCols does not influence prepared queries:
> --- /dev/null
> +++ b/test/src/premature-test.c
> @@ -0,0 +1,201 @@
> +/*
> + * This test case tests the so-called "premature execution" behaviour, which
> + * can be influenced with the DisallowPremature setting.
> + */
Fixed that. I also refactored and expanded the test case a bit, and
fixed another bug that I bumped into.
Committed with those fixes. Thanks for the review!
- Heikki