Bruce Momjian wrote:
>
> OK, I have created a context diff of your changes, which is attached.
> Can someone comment on this? The patch appears to deal with
> STMT_PREMATURE differently than our current code.
>
> In fact, it seems to disable this code:
>
> /*
> * If the statement is premature, it means we already executed it from
> * an SQLPrepare/SQLDescribeCol type of scenario. So just return
> * success.
> */
> if (stmt->prepare && stmt->status == STMT_PREMATURE)
>
Though this fixes the bug reported by Keith, STMT_PREMATURE
loses its meaning and SC_pre_execute() does a non-essential
work as a result. Possibly SQLPrepare() must be rewritten
though I'm not sure.
regards,
Hiroshi Inoue