> On Sun, May 28, 2006 at 09:43:46AM -0700, Alan Hodgson wrote:
>> On Saturday 27 May 2006 19:26, elein <elein at varlena.com> wrote:
>> > I've got a customer with tons of inherited tables.
>> > Is there any reason why slony-I does not work with
>> > inheritance?  I thought I heard that, but would like
>> > very much to believe it is not true.
>> >
>> > Anything I should watch out for with inheritance and slony-I
>> > if it does work?
>>
>> The only problem I've seen doing this is sometimes the column ordering
>> gets
>> out of whack depending when you add them on the master vs. copying the
>> schema to the slave(s).
>>
>> This shouldn't be a problem now that Slony specifies the column ordering
>> during COPY, though.  I haven't had any problems since that was changed.
>>
>
> Good to know!  Any other gotchas anyone has run into may be helpful.
> If it looks good, I am going to test on a wild schema with
> tons of child tables (not my design!) and will report back.
>
> Thank you all for perltools...

If that "wild schema" is publishable, it would be a most excellent idea to
add this as an additional test to the set of regression tests for Slony-I.

I think "desire for an inheritance case" has been around for a while; I
haven't had a case for needing it, so I certainly haven't gotten around to
it.