From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
---|---|
To: | Andres Freund <andres(at)anarazel(dot)de> |
Cc: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Subject: | _bt_delitems_delete() should use XLogRegisterBufData(), not XLogRegisterData() |
Date: | 2020-01-01 21:00:59 |
Message-ID: | CAH2-Wzkz4TjmezzfAbaV1zYrh=fr0bCpzuJTvBe5iUQ3aUPsCQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
The RelationNeedsWAL() code block within _bt_delitems_delete() has had
the following comment for many years now:
/*
* We need the target-offsets array whether or not we store the whole
* buffer, to allow us to find the latestRemovedXid on a standby
* server.
*/
XLogRegisterData((char *) itemnos, nitems * sizeof(OffsetNumber));
However, we don't actually need to do it that way these days. We won't
go on to determine a latestRemovedXid on a standby as of commit
558a9165e08 (that happens on the primary instead), so the comment
seems wrong.
Rather than just changing the comment, I propose that we tweak the
behavior of _bt_delitems_delete() to match its sibling function
_bt_delitems_vacuum(). That is, it should use XLogRegisterBufData(),
not XLogRegisterData(). This is cleaner, and ought to be a minor win.
Attached patch shows what I have in mind. The new comment block has
been copied from _bt_delitems_vacuum().
--
Peter Geoghegan
Attachment | Content-Type | Size |
---|---|---|
0001-Associate-LP_DEAD-offsets-with-WAL-record-s-buffer.patch | application/x-patch | 2.0 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Vik Fearing | 2020-01-01 21:12:09 | Re: WIP: System Versioned Temporal Table |
Previous Message | David Fetter | 2020-01-01 19:09:18 | Re: backup manifests |