mainline XLOG replay.  Shouldn't generic_xlog.c take the same trouble?
Yes, it should. Alexander now works on it.

2. Unless I'm missing something, contrib/bloom is making no effort
to ensure that bloom index pages can be distinguished from other pages
by pg_filedump.  That's fine if your expectation is that bloom will always
be a toy with no use in production; but otherwise, not so much.  You need
to make sure that the last two bytes of the page special space contain a
uniquely identifiable code; compare spgist_page_id in SPGiST indexes.

Now contrib/bloom is a canonical example how to implement yourown index and how to use generic WAL interface. And it is an useful toy: in some cases it could help in production system, patch attached. I'm a little dissatisfied with patch because I had to add unused field to BloomPageOpaqueData, in opposite case size of BloomPageOpaqueData struct equals 6 bytes but special size is MAXALIGNED, so, last two bytes will be unused. If unused field is not a problem, I will push this patch.

Teodor Sigaev                                   E-mail:

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to