Hi,
we finished first stage of our proposal for changing of index AM tables
(see for reference http://fts.postgresql.org/db/mw/msg.html?mid=1029290)
I attached 3 files:
1. patch_72_systbl.gz - patch to current CVS
2. btree_gist.tar.gz - contrib/btree_gist module - implementation of Btree using GiST with
support of int4 and timestamp types.
3. test.tar.gz - test suite for brave (not for applying !)
Regression tests and our tests passed fine
Patch is for today CVS, please apply them asap to avoid possible
conflicts.
Now we're going to 2nd stage of our proposal.
We plan to remove pg_index.indislossy (now we have pg_amop.amopreqcheck)
and pg_index.indhaskeytype (it's just don't used, all functionality is in
pg_opclass.opckeytypenow)
question:
Do we need to normalize pg_amop and pg_amproc tables ?
Tom was concerned (http://fts.postgresql.org/db/mw/msg.html?mid=1025860)
about possible performance degradation.
We think it's possible to leave tables as is. Of course it'd require
some attention when updating these tables.
Regards, Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83