Robert Haas [] wrote:
>> Where you get into trouble there is that you might run CREATE
>> from that session

Yes. I can see this problem. And, while I can imagine resolving it with
context belonging to the extension, separate from the current session's
context, any resolution gets to be pretty complex. Probably complex
enough that the resolution is worse than living with the problem as part
of the cost of the feature. Which means, it remains one of the arguments
against it.

>> Again, I'm not trying to rain down fire and brimstone
>> on your idea here and I clearly see the utility of it.

I do not feel this at all (though, since this is e-mail, it is helpful
that you state it explicitly). And, I have not felt like this was the
case at any point in these discussions. I have consistently received
thoughtful and remarkably good responses containing solid points. 

While I did not recognize all the impacts, and I probably still do not,
I never thought this was a trivial issue with no arguments against it
(not least, that what I was asking for is not standard compliant). Every
modal behavior of any sort in any software adds pain. At the very least
it increases the regression testing burden. And, this is a mode with
pretty fundamental impact. It has to be worth a lot to somebody to be
worth having. And, of course, it also has to work.

Ian Lewis (

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to