Copy-editing for previous patch, plus fixing some longstanding markup
issues and oversights (like not mentioning that failures will set the
PQerrorMessage string).

INT64CONST macro. Fix lo_hton64 and lo_ntoh64 not to use int32_t and
uint32_t.

addition.

4TB large objects (standard 8KB BLCKSZ case).  For this purpose new
libpq API lo_lseek64, lo_tell64 and lo_truncate64 are added.  Also
corresponding new backend functions lo_lseek64, lo_tell64 and
lo_truncate64 are added. inv_api.c is changed to handle 64-bit
offsets.

Patch contributed by Nozomi Anzai (backend side) and Yugo Nagata
(frontend side, docs, regression tests and example program). Reviewed
by Kohei Kaigai. Committed by Tatsuo Ishii with minor editings.

Use the terms "simple bind" and "search+bind" consistently do
distinguish the two modes (better than first mode and second mode in
any case).  They were already used in some places, now it's just more
prominent.

Split up the list of options into one for common options and one for
each mode, for clarity.

Add configuration examples for either mode.

because it is not correct.

Instead of continuing if the next character is not an array boundary get_data()
used to continue only on finding a boundary so it was not able to read any
element after the first.

The regular backend's main loop handles signal handling and error recovery
better than the current WAL sender command loop does. For example, if the
client hangs and a SIGTERM is received before starting streaming, the
walsender will now terminate immediately, rather than hang until the
connection times out.

Per testing of previous patch.

This makes the naming inside plpgsql consistent and distinguishes the
file from the backend's gram.y file.  It will also allow easier
refactoring of the bison make rules later on.

Our getnameinfo() replacement implementation in getaddrinfo.c failed
unless NI_NUMERICHOST and NI_NUMERICSERV were given as flags, because
it doesn't resolve host names, only numeric IPs.  But per standard,
when those flags are not given, an implementation can still degrade to
not returning host names, so this restriction is unnecessary.  When we
remove it, we can eliminate some code in postmaster.c that apparently
tried to work around that.

The initial transition value is stored as a text string and not fed to the
transition type's input function until runtime (so that values such as
"now" don't get frozen at creation time).  Previously, CREATE AGGREGATE
didn't do anything with it but that, which meant that even erroneous values
would be accepted and not complained of until the aggregate is used.  This
seems unhelpful, and it's confused at least one user, as in Rhys Stewart's
recent report.  It seems worth taking a few more cycles to invoke the input
function and verify that the value is acceptable.  We can't do this if the
transition type is polymorphic, but in normal aggregates we know the actual
transition type so we can call the right input function.

The previous coding of the YYLLOC_DEFAULT macro behaved strangely for empty
productions, assigning the previous nonterminal's location as the parse
location of the result.  The usefulness of that was (at best) debatable
already, but the real problem is that in list-generating nonterminals like
	OptFooList: /* EMPTY */ { ... } | OptFooList Foo { ... } ;
the initially-identified location would get copied up, so that even a
nonempty list would be given a bogus parse location.  Document how to work
around that, and do so for OptSchemaEltList, so that the error condition
just added for CREATE SCHEMA IF NOT EXISTS produces a sane error cursor.
So far as I can tell, there are currently no other cases where the
situation arises, so we don't need other instances of this coding yet.

These reference pages still claimed that you have to be superuser to create
a database or schema owned by a different role.  That was true before 8.1,
but it was changed in commits aa111062 and
f91370cd to allow assignment of ownership
to any role you are a member of.  However, at the time we were thinking of
that primarily as a change to the ALTER OWNER rules, so the need to touch
these two CREATE ref pages got missed.

Per discussion, schema-element subcommands are not allowed together with
this option, since it's not very obvious what should happen to the element
objects.

Fabrízio de Royes Mello

Remove duplicate implementation of catalog munging and miscellaneous
privilege and consistency checks.  Instead rely on already existing data
in objectaddress.c to do the work.

Author: KaiGai Kohei
Tweaked by me
Reviewed by Robert Haas

examine_simple_variable supposed that any RTE_SUBQUERY rel it gets pointed
at must have been planned already.  However, this isn't a safe assumption
because we must do selectivity estimation while generating indexscan paths,
and that code might look at join clauses involving a rel that the loop in
set_base_rel_sizes() hasn't reached yet.  The simplest fix is to play dumb
in such a situation, that is give up trying to extract any stats for the
Var.  This could possibly be improved by making a separate pass over the
RTE list to plan each unflattened subquery before we start the main
planning work --- but that would be pretty invasive and it doesn't seem
worth it, for now at least.  (We couldn't just break set_base_rel_sizes()
into two loops: the prescan would need to handle all subquery rels in the
query, not only those in the current join subproblem.)

This bug was introduced in commit 1cb108ef,
although I think that subsequent changes may have exposed it more than it
was originally.  Per bug #7580 from Maxim Boguk.

Apparently this was considered in the original code (see commit
cec3b0a9) but I failed to notice that such entries would always be
skipped by the database check at the start of the loop.

Per bugs #7578 by Nikolay, #6116 by tushar.qa@gmail.com.

This allows logging only some fraction of transactions, greatly reducing
the amount of log generated.

Tomas Vondra, reviewed by Robert Haas and Jeff Janes.

You can now get the number of rows processed by a COPY statement in a
PL/pgSQL function with "GET DIAGNOSTICS x = ROW_COUNT".

Pavel Stehule, reviewed by Amit Kapila, with some editing by me.

The comment explaining the naming of timeline history files was wrong, and
the history file was not being arhived.

Pointed out by Fujii Masao.

Backpatch to 9.2.

On some platforms these functions return NULL, rather than the more common
practice of returning a pointer to a zero-sized block of memory.  Hack our
various wrapper functions to hide the difference by substituting a size
request of 1.  This is probably not so important for the callers, who
should never touch the block anyway if they asked for size 0 --- but it's
important for the wrapper functions themselves, which mistakenly treated
the NULL result as an out-of-memory failure.  This broke at least pg_dump
for the case of no user-defined aggregates, as per report from
Matthew Carrington.

Back-patch to 9.2 to fix the pg_dump issue.  Given the lack of previous
complaints, it seems likely that there is no live bug in previous releases,
even though some of these functions were in place before that.

Instead of having each object type implement the catalog munging
independently, centralize knowledge about how to do it and expand the
existing table in objectaddress.c with enough data about each object
type to support this operation.

Author: KaiGai Kohei
Tweaks by me
Reviewed by Robert Haas

We had a number of variants on the theme of "malloc or die", with the
majority named like "pg_malloc", but by no means all.  Standardize on the
names pg_malloc, pg_malloc0, pg_realloc, pg_strdup.  Get rid of pg_calloc
entirely in favor of using pg_malloc0.

This is an essentially cosmetic change, so no back-patch.  (I did find
a couple of places where psql and pg_dump were using plain malloc or
strdup instead of the pg_ versions, but they don't look significant
enough to bother back-patching.)

Fujii Masao

objects does not match between the old and new clusters.

Backpatch to 9.2.

entries are not dumped.   This fixes an error caused by
droping/recreating the information_schema, but other failures were also
possible.

Backpatch to 9.2.

comparison;  also report the old/new values if they don't match.

Backpatch to 9.2.

timeval.t_sec is of type time_t, which is not always compatible with long.
I'm not sure if this was just harmless warning or a real bug, but this
fixes it, anyway.

Hopefully this makes the *BSD buildfarm animals happy.

This is just refactoring, to make the functions accessible outside xlog.c.
A followup patch will make use of that, to allow fetching timeline history
files over streaming replication.

This affects date_in(), and a couple of other funcions that use DecodeDate().

Hitoshi Harada

don't accidentally remove it.

The error messages they generate are not portable enough.

Also, since the only point of the alter_generic_1 expected file was to
cover platforms with no collation support, it's now useless, so remove
it.