Tom Lane wrote:
>
> Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> > Tom Lane wrote:
> >> Actually, your trial required column dropped-ness to be checked in
> >> many more places than the proposed approach does.
>
> > Have you ever really checked my trial implementation ?
>
> Well, I've certainly stumbled over it in places like relcache.c
> and preptlist.c, which IMHO should not have to know about this...
> and I have little confidence that there are not more places that
> would have needed fixes if the change had gotten any wide use.
> You were essentially assuming that it was okay for pg_attribute.attnum
> to not agree with indexes into tuple descriptors, which seems very
> shaky to me.
I already explained it to you once in the thread Re: [HACKERS]
RFC: Restructuring pg_aggregate. How many times should I
explain the same thing ?
My trial implementation is essentially the same as adding
isdropped pg_attribute column. There's no strangeness in
my implementation.
The reason why I adopted negative attnos is as follows.
I also explained it more than twice.
1) It doesn't need initdb. It was very conveneient for the TRIAL implementation.
2) It's more sensitive about oversights of modification than isdropped column implementation.
regards,
Hiroshi Inouehttp://w2422.nsk.ne.jp/~inoue/