Re: TODO item: Allow more complex user/database default GUC settings

On Mon, Sep 21, 2009 at 12:23 PM, Alvaro Herrera
<alvherre(at)commandprompt(dot)com> wrote:
> Robert Haas escribió:
>
>> So here's the followup question - do you intend to do one of those
>> things for this CommitFest, or should we mark this as Returned with
>> Feedback once Bernd posts the rest of his review?
>
> What feedback is it supposed to be returned with?  Precisely what I
> wanted is some feedback on the general idea.  Brendan's "I like the
> approach" is good, but is it enough to deter a later veto from someone
> else?

Well, you've hit there on one of the things that we don't always do
well. Many a patch author has posted an idea, received no feedback,
proceeded to implementation, and then the knives come out. On a good
day, the CommitFest process ensures that every patch gets a second
opinion, but it doesn't guarantee that a third opinion won't come
crawling out of the woodwork at a later date. In this respect, you're
actually operating at a slight advantage relative to most of us,
because you can post your revised patch and commit it if no one
objects too strongly, whereas I (for example) have to convince one of
about two people - Tom or Peter, for nearly anything I'm likely to
develop - to take an affirmative action on my behalf.

This whole phenomenon of proposals to which no objection was made at
the outset later getting flak for one reason or another is, I think, a
source of much frustration and discourages people from putting effort
into projects they might otherwise be willing to undertake. But I
haven't the least idea how to fix it, and I can't offer you any
guarantees with respect to the present situation either.

...Robert