On Tue, Sep 25, 2001 at 03:05:47PM -0400, Greg Stark wrote: > Tim Bunce <[EMAIL PROTECTED]> writes: > > > > Is this reverting to the previous behaviour? > > > > Is there a simply way to recognize this condition? > > > A wrapper function would be fine. > > Tim. > > Any chance of answering my other two questions? Will it signal an error, and > what will the error look like? Sorry, time is in short supply. I think both were answered by the description in the change log and the natural (documented) implications of it. > I'll add a third question, "What in heaven's name is the rationale behind > signaling an error when the query was executed successfully?" At least for Oracle in the cases people have described to me the OCI_SUCCESS_WITH_INFO is telling them something 'serious'. But I'd agree that the behaviour is far from ideal. I'll ponder it in more depth later >>todo. Tim.