(1) Keep a pin on the scan's current buffer and mark buffer. This
avoids the need to do a ReadBuffer() for each tuple produced by the
scan. Since ReadBuffer() is expensive, this is a significant win.

(2) Convert a ReleaseBuffer(); ReadBuffer() pair into
ReleaseAndReadBuffer(). Surely not a huge win, but it saves a lock
acquire/release...

(3) Remove a bunch of duplicated code in rtget.c; make rtnext() handle
both the "initial result" and "subsequent result" cases.

(4) Add support for index tuple killing

(5) Remove rtscancache(): it is dead code, for the same reason that
gistscancache() is dead code (an index scan ought not be invoked with
NoMovementScanDirection).

The end result is about a 10% improvement in rtree index scan perf,
according to contrib/rtree_gist/bench.

good style and to satisfy sparse. From Alvaro Herrera.

its that time ... tag it for release

interval datatypes as well as timestamp.

correct.  It's in addition to the previous patch.

David Fetter

an untyped literal in the CASE's test expression. This adds test
coverage for a bug that was fixed by Tom on January 12.

tells what arguments go to pl/perl triggers.

David Fetter

Ian Barwick

Alvaro

Euler Taveira de Oliveira

Euler Taveira de Oliveira

the right place in the docs, and gettext()ify it.

per Andrew Dunstan.  Also, don't override the user's value of PGHOST
in the 'make installcheck' case.  I think the latter was an ill-considered
workaround for the Windows code back when libpq didn't properly default
to localhost on Unix-socket-less platforms.

failure in SelectConfigFiles().  Cosmetic issue, but ...