Mark Dilger <mark.dil...@enterprisedb.com> writes:
> On Sep 11, 2020, at 12:54 PM, Robert Haas <robertmh...@gmail.com> wrote:
>> On Fri, Sep 11, 2020 at 3:23 PM Mark Dilger
>> <mark.dil...@enterprisedb.com> wrote:
>>> Another option would be to have pg_dump take a strictness mode option. I
>>> don't think the option should have anything to do with postfix operators
>>> specifically, but be more general like --dump-incompatible-objects vs.
>>> --omit-incompatible-objects vs. --error-on-incompatible-objects vs.
>>> --do-your-best-to-fixup-incompatible-objects, with one of those being the
>>> default (and with all of them having better names). If
>>> --error-on-incompatible-objects were the default, that would behave as
>>> Robert recommended upthread.
>>> I can totally see an objection to the added complexity of such options, so
>>> I'm really just putting this out on the list for comment.
>> I'm not opposed to Tom's proposal. I just wanted to raise the issue
>> for discussion.
> Ah, ok. I don't feel any need for changes, either. I'll leave the
> patch as it stands now.
We're in violent agreement it seems.
At some point it might be worth doing something like what Mark suggests
above, but this patch shouldn't be tasked with it. In any case, since
pg_dump does not know what the target server version really is, it's
going to be hard for it to authoritatively distinguish what will work
or not.
regards, tom lane
