Re: pg_migrator and an 8.3-compatible tsvector data type

On Mon, Jun 1, 2009 at 11:18 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Greg Stark <stark@enterprisedb.com> writes:
>> On Mon, Jun 1, 2009 at 4:03 PM, Robert Haas <robertmhaas@gmail.com> wrote:
>>> We really need to figure out an approach that lets us keep the old
>>> datatypes around under a different name while making the original name
>>> be the new version of the datatype.  That way people can migrate and
>>> be up, and deal with the need to rewrite their tables at a later time.
>
>> I do agree that having to rewrite the whole table isn't really
>> "upgrade-in-place".
>
> It's certainly the case that there is a lot more work to do before
> pg_migrator could support everything that we reasonably want to be
> able to do in a version update.  As I see it, the reason it's getting
> revived now is that 8.3->8.4 happens to be an update where most of what
> it can't (yet) do isn't necessary.  That means we can get it out there,
> get the bugs out of the functionality it does have, and most importantly
> try to set an expectation that future updates will also have some degree
> of update-in-place capability.  If we wait till it's perfect then
> nothing will ever happen at all in this space.

I agree.  I remain doubtful that dumping and reloading the schema is
the best way to go, but it's certainly a worthwhile experiment,
because (a) I might easily be wrong and (b) we'll hopefully learn some
things that will be useful going forward.

...Robert