This is still an open item for 9.0, and we should also backport it to 8.4.
Which do we take on? Or is there better idea?
Alvaro's idea:
> > > Maybe a better solution is to have some kind of notion of a default-only
> > > entry, which is sufficient to insert the default into the struct but
> > > isn't accepted as a user-settable item.
Itagaki's idea:
> It just fills fillfactor (and analyze_threshold)
> to default values for TOAST relations.
Regards,
---
Takahiro Itagaki
NTT Open Source Software Center