Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> Tom Lane wrote:
>> Note that if the line type had been in use in the field, this would
>> break pg_upgrade'ability of databases containing line values; but
>> it seems unlikely that there are any (they'd have had to be compiled
>> with -DENABLE_LINE_TYPE).
> Should we add a test to pg_upgrade --check mode?
I doubt it's worth it. Anybody who was enabling the old code is pretty
well screwed no matter what. They can't upgrade with a plain pg_dump
either, since the I/O format is different.
What we do need, probably, is a warning about the incompatibility in the
release notes.
regards, tom lane