On Thu, Apr 7, 2022 at 11:40 PM Greg Stark <stark@mit.edu> wrote:
> That doesn't seem to be entirely inconsistent with what Tom describes.
> Instead of "throw an error" the function would return an error and
> possibly some extra info which the caller would use to handle the
> error appropriately.
I don't personally see how we're going to come out ahead with that
approach, but if you or Tom or someone else want to put something
together, that's fine with me. I'm not stuck on this approach, I just
don't see how we come out ahead with the type of thing you're talking
about. I mean we could return the error text, but it's only to a
handful of places, so it just doesn't really seem like a win over what
the patch is already doing.
--
Robert Haas
EDB: http://www.enterprisedb.com