At 10:00 4/07/00 -0400, Bruce Momjian wrote:
>
>Yes, this type of thing is clearly needed. We need to be able to UPDATE
>the oid column too.
>
Updating OIDs sounds like a *very* scary thing to do. My motivation for
heap_create_oid is very specific - I want to be able to do a valid database
backup/restore, and need to handle the current (bad) BLOB implementation. I
would sincerely hope that (a) heap_create_oid would be marked in some way
as 'internal use only', and (b) go away in 7.2 or 7.3 when TOASTED BLOBs
come along...
FWIW, I now have a working lo_create_oid, that does what is expected.
When (and if) the new pg_restore is accepted, I'll probably have a look at
implementing BLOB backup, but there might be some very justifiable
resistance to heap_create_oid...so before I go much further I'd appreciate
a barometer reading on the likely acceptability of heap_create_oid.
----------------------------------------------------------------
Philip Warner | __---_____
Albatross Consulting Pty. Ltd. |----/ - \
(A.C.N. 008 659 498) | /(@) ______---_
Tel: (+61) 0500 83 82 81 | _________ \
Fax: (+61) 0500 83 82 82 | ___________ |
Http://www.rhyme.com.au | / \| | --________--
PGP key available upon request, | /
and from pgp5.ai.mit.edu:11371 |/