This was a thinko introduced in a patch from last February; it results
in memory leakage if an SRF is shut down before the actual end of query,
because subsequent code will be running in a longer-lived context than
it's expecting to be.

regardless of plan changes.  Per intermittent buildfarm failures on
"pigeon" and others.

to drop a table that is referenced by an open cursor.  Fix unstable
ecpg regression test result that was produced by this oversight.

but I think this is enough to turn the buildfarm green again.

RHS that can't be unique-ified --- join_is_legal has to check that before
deciding to build a join, else we'll have an unimplementable joinrel.
Per report from Greg Stark.

a mismatching installation.  Pick a default port number calculated from the
version number, and try a few times with other numbers if that one doesn't
work.  Check if we can connect to the port before starting our own postmaster,
to detect some other server already running there.  To simplify the code,
drop --temp-port option and use --port for both temp-install and pre-installed
case.

the sample contains just a one tsvector, containing only one lexeme.

as soon as the first page fills up, and is marked as (almost) full,
though.

PL/pgSQL has much to do with SQL/PSM.

business in core.

that the presented list of equivalent operations is meant to be the
primary definition of what it does.  Per comment from Guillaume Smet.

DestReceiver created during postquel_start needs to be destroyed during
postquel_end.  In a moment of brain fade I had assumed this would be taken
care of by FreeQueryDesc, but it's not (and shouldn't be).

interval style into ecpg.

versions of gcc.

page build target.  This covers from-source, Debian, and Fedora
installation variants.

results (ie, an empty "broken" buffer) if memory overrun occurs anywhere
along the way to filling the buffer.  The previous coding would just silently
discard portions of the intended buffer contents, as exhibited in trouble
report from Sam Mason.  Also, tweak psql's main loop to correctly detect
and report such overruns.  There's probably much more that should be done
in this line, but this is a start.

misdocumented as not being implemented.  In reality it has worked since
the release of 8.2.

by hand.  As an added bonus, the new code is smaller and more understandable,
and the ugly loops are gone.

This had been discussed all along but never implemented.  It became clear that
it really needed to be fixed after a bug report by Pavan Deolasee.

libpq.  As noted by Peter, adding this variable created a risk of unexpected
connection failures when talking to older server versions, and since it
doesn't do anything you can't do with PGOPTIONS, it doesn't seem really
necessary.  Removing it does occasion a few extra lines in pg_regress.c,
but saving a getenv() call per libpq connection attempt is perhaps worth
that anyway.

complaints about empty object files on some platforms, eg Darwin.

of our way to ignore the exit status).  If this ever really happens, worst
case you will not see the "score" at the bottom of the test run.

it out for now so the buildfarm recovers.

The information on why the shared libraries are built the way they are
was not relevant to end users and has been made a mailing list archive
link in Makefile.shlib.

This uses the function fnmatch() which is not available on all platforms
(notably Windows), so import the implementation from NetBSD into src/port.

Jim Cox and Peter Eisentraut

the bgwriter immediately.  This covers the case where the bgwriter is still
starting up, as seen in a recent buildfarm failure.  In future it might also
assist with clean recovery after a bgwriter termination and restart ---
right now the postmaster treats early bgwriter exit as a system crash,
but that might not always be so.