On Wed, Sep 6, 2017 at 12:26 PM, Simon Riggs <simon@2ndquadrant.com> wrote:
>>> I'd personally be fine with --no-whatever for any whatever that might
>>> be a subsidiary property of database objects. We've got
>>> --no-security-labels, --no-tablespaces, --no-owner, and
>>> --no-privileges already, so what's wrong with --no-comments?
>>>
>>> (We've also got --no-publications; I think it's arguable whether that
>>> is the same kind of thing.)
>>
>> And --no-subscriptions in the same bucket.
>
> Yes, it is. I was suggesting that we remove those as well.
That seems like a non-starter to me. I have used those options many
times to solve real problems, and I'm sure other people have as well.
We wouldn't have ended up with all of these options if users didn't
want to control such things.
> But back to the main point which is that --no-comments discards ALL
> comments simply to exclude one pointless and annoying comment. That
> runs counter to our stance that we do not allow silent data loss. I
> want to solve the problem too. I accept that not everyone uses
> comments, but if they do, spilling them all on the floor is a user
> visible slip up that we should not be encouraging. Sorry y'all.
/me shrugs.
I agree that there could be better solutions to the original problem,
but I think there's no real argument that a user can't exclude
comments from a backup if they wish to do so. As the OP already
pointed out, it can still be done without the switch; it's just more
annoying.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company