Re: pageinspect's infomask and infomask2 as smallint

Robert Haas <robertmhaas@gmail.com> writes:
> On Tue, Feb 15, 2011 at 10:53 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> What risk? �And at least we'd be trying to do it cleanly, in a manner
>> that should work for at least 99% of users. �AFAICT, Heikki's proposal
>> is "break it for everyone, and damn the torpedoes".

> I must be confused.  I thought Heikki's proposal was "fix it in 9.1,
> because incompatibilities are an expected part of major release
> upgrades, but don't break it in 9.0 and prior, because it's not
> particularly important and we don't want to change behavior or risk
> breaking things in minor releases".

No, nobody was proposing changing it before 9.1 (or at least I didn't
think anybody was).  What's under discussion is how much effort to put
into making a 9.0-to-9.1 upgrade go smoothly for people who have the
function installed.
        regards, tom lane