Tom Lane wrote:
>
> I think the proposed LONG type is a hack, and I'd rather see us solve
> the problem correctly. ISTM that allowing a tuple to be divided into
> "primary" and "continuation" tuples, all stored in the same relation
> file, would be a much more general answer and not significantly harder
> to implement than a LONG datatype as Jan is describing it.
Actually they seem to be two _different_ problems -
1) we may need bigger tuples for several reasons (I would also suggest
making index tuples twice as long as data tuples to escape the problem
of indexing text fields above 4K (2K?)
2) the LOB support should be advanced to a state where one could reasonably
use them for storing more than a few LOBs without making everything else to
crawl, even on filesystems that don't use indexes on filenames (like ext2)
After achieving 2) support could be added for on-demand migrating of LONG
types to LOBs
I guess that Jans suggestion is just a quick hack for avoiding fixing LOBs.
-----------------------
Hannu