The old value of 32MB has been around for a very long time, and in the
meantime typical system memories have become vastly larger.  Also, now
that we no longer depend on being able to fit the entirety of our
shared memory segment into the system's limit on System V shared
memory, there's a much better chance of the higher limit actually
proving productive.

Per recent discussion on pgsql-hackers.

Amit Kapila, reviewed by Shigeru Hanada and Peter Eisentraut,
with some modifications by me.

This makes it possible for the master to track how much data has
actually been written my pg_receivexlog - and not just how much
has been sent towards it.

This ensures that a standby such as pg_receivexlog will not be selected
as sync standby - which would cause the master to block waiting for
a location that could never happen.

Fujii Masao

This hasn't been true since 9.1, when the default was changed to -1.
Remove the reference completely, keeping the discussion of the parameter
and it's shared memory effects on the config page.

pgfoundry is deprectaed and no longer accepting new projects,
so we really shouldn't be directing people there.

Also remove special references to downloads off pgfoundry since they are
not correct - downloads are done through the main website.

This commit improves the comments in pg_wchar.h and creates #define symbols
for some formerly hard-coded values.  No substantive code changes.

Tatsuo Ishii and Tom Lane

Per bug #6593, REASSIGN OWNED fails when the affected role has created
an extension.  Even though the user related to the extension is not
nominally the owner, its OID appears on pg_shdepend and thus causes
problems when the user is to be dropped.

This commit adds code to change the "ownership" of the extension itself,
not of the contained objects.  This is fine because it's currently only
called from REASSIGN OWNED, which would also modify the ownership of the
contained objects.  However, this is not sufficient for a working ALTER
OWNER implementation extension.

Back-patch to 9.1, where extensions were introduced.

Bug #6593 reported by Emiliano Leporati.

Since the request size will now be ~48 bytes regardless of how
shared_buffers et. al. are set, much of this advice is no longer
relevant.

Commit f11e8be3 broke this; it was right
in Peter's original patch, but I messed it up before committing.

This avoids discriminating against wal_sync_method = open_sync or
open_datasync.

Fujii Masao, reviewed by Andres Freund

A thinko in commit 029dfdf1 caused the year
519 to be handled differently from either adjacent year, which was not the
intention AFAICS.  Report and diagnosis by Marc Cousin.

In passing, remove redundant re-tests of year value.

Per gripe from Tom Lane.

Instead of letting every backend participating in a group commit wait
independently, have the first one that becomes ready to flush WAL wait
for the configured delay, and let all the others wait just long enough
for that first process to complete its flush.  This greatly increases
the chances of being able to configure a commit_delay setting that
actually improves performance.

As a side consequence of this change, commit_delay now affects all WAL
flushes, rather than just commits.  There was some discussion on
pgsql-hackers about whether to rename the GUC to, say, wal_flush_delay,
but in the absence of consensus I am leaving it alone for now.

Peter Geoghegan, with some changes, mostly to the documentation, by me.

Per testing by Andres Freund, this improves replication performance
and reduces replication latency and latency jitter.  I was a bit
concerned about moving more work into XLogInsert, but testing seems
to show that it's not a problem in practice.

Along the way, improve comments for WaitLatchOrSocket.

Andres Freund.  Review and stylistic cleanup by me.

When (re) loading the typcache comparison cache for an enum type's values,
use an up-to-date MVCC snapshot, not the transaction's existing snapshot.
This avoids problems if we encounter an enum OID that was created since our
transaction started.  Per report from Andres Freund and diagnosis by Robert
Haas.

To ensure this is safe even if enum comparison manages to get invoked
before we've set a transaction snapshot, tweak GetLatestSnapshot to
redirect to GetTransactionSnapshot instead of throwing error when
FirstSnapshotSet is false.  The existing uses of GetLatestSnapshot (in
ri_triggers.c) don't care since they couldn't be invoked except in a
transaction that's already done some work --- but it seems just conceivable
that this might not be true of enums, especially if we ever choose to use
enums in system catalogs.

Note that the comparable coding in enum_endpoint and enum_range_internal
remains GetTransactionSnapshot; this is perhaps debatable, but if we
changed it those functions would have to be marked volatile, which doesn't
seem attractive.

Back-patch to 9.1 where ALTER TYPE ADD VALUE was added.

Commit 7357558f introduced "(void) token;"
into the READ_TEMP_LOCALS() macro, to suppress complaints from gcc 4.6
when the value of token was not used anywhere in a particular node-read
function.  However, this just moved the warning around: inspection of
buildfarm results shows that some compilers are now complaining that token
is being read before it's set.  Revert the READ_TEMP_LOCALS() macro change
and instead put "(void) token;" into READ_NODE_FIELD(), which is the
principal culprit for cases where the warning might occur.  In principle we
might need the same in READ_BITMAPSET_FIELD() and/or READ_LOCATION_FIELD(),
but it seems unlikely that a node would consist only of such fields, so
I'll leave them alone for now.

Solaris Studio warns about this, and some compilers might think it's an
outright syntax error.

The original coding had it as "PGShmemHeader *", but that doesn't offer any
notational benefit because we don't dereference it.  And it was resulting
in compiler warnings on some platforms, notably buildfarm member
castoroides, where mmap() and munmap() are evidently declared to take and
return "char *".

If a CHECK constraint or index definition contained a whole-row Var (that
is, "table.*"), an attempt to copy that definition via CREATE TABLE LIKE or
table inheritance produced incorrect results: the copied Var still claimed
to have the rowtype of the source table, rather than the created table.

For the LIKE case, it seems reasonable to just throw error for this
situation, since the point of LIKE is that the new table is not permanently
coupled to the old, so there's no reason to assume its rowtype will stay
compatible.  In the inheritance case, we should ideally allow such
constraints, but doing so will require nontrivial refactoring of CREATE
TABLE processing (because we'd need to know the OID of the new table's
rowtype before we adjust inherited CHECK constraints).  In view of the lack
of previous complaints, that doesn't seem worth the risk in a back-patched
bug fix, so just make it throw error for the inheritance case as well.

Along the way, replace change_varattnos_of_a_node() with a more robust
function map_variable_attnos(), which is capable of being extended to
handle insertion of ConvertRowtypeExpr whenever we get around to fixing
the inheritance case nicely, and in the meantime it returns a failure
indication to the caller so that a helpful message with some context can be
thrown.  Also, this code will do the right thing with subselects (if we
ever allow them in CHECK or indexes), and it range-checks varattnos before
using them to index into the map array.

Per report from Sergey Konoplev.  Back-patch to all supported branches.

Change things so that something like initdb --auth-local=peer
--auth-host=md5 does not cause a "must specify a password" error,
like initdb -A md5 does.

If the record header is garbled, we're now quite likely to notice it before
we try to make a bogus memory allocation and run out of memory. That can
still happen, if the xlog record is split across pages (we cannot verify
the record header until reading the next page in that scenario), but this
reduces the chances. An out-of-memory is treated as a corrupt record
anyway, so this isn't a correctness issue, just a case of giving a better
error message.

Per Amit Kapila's suggestion.

Noted by Andres Freund.  Also improve a couple of comments.

This bug was introduced by commit 20d98ab6e4110087d1816cd105a40fcc8ce0a307,
so backpatch this to 9.0-9.2 like that one.

This fixes bug #6710, reported by Tarvi Pillessaar

Now the log file not only contains the output from commands executed by
system(), but also what command it was in the first place.  This
arrangement makes debugging a lot simpler.

This is for convenience, now that adding recursive targets is much
easier than it used to be when the NLS stuff was initially added.

The LISTEN/NOTIFY subsystem got confused if SimpleLruZeroPage failed,
which would typically happen as a result of a write() failure while
attempting to dump a dirty pg_notify page out of memory.  Subsequently,
all attempts to send more NOTIFY messages would fail with messages like
"Could not read from file "pg_notify/nnnn" at offset nnnnn: Success".
Only restarting the server would clear this condition.  Per reports from
Kevin Grittner and Christoph Berg.

Back-patch to 9.0, where the problem was introduced during the
LISTEN/NOTIFY rewrite.

snprintf counts trailing NUL towards the char limit.  Failing to account
for that was causing an invalid value to be passed to pg_resetxlog -l,
aborting the upgrade process.

On old HPUX this has to be #defined to -1.  It might be that other values
are required on other dinosaur systems, but we'll worry about that when
and if we get reports.

Spotted by Amit Kapila

The $(or) make function was introduced in GNU make 3.81, so the
previous coding didn't work in 3.80.  Write it differently, and
improve the variable naming to make more sense in the new coding.

Per an observation by Thom Brown that my previous commit made an
overly large shmem allocation crash the server, on Linux.

Except when compiling with EXEC_BACKEND, we'll now allocate only a tiny
amount of System V shared memory (as an interlock to protect the data
directory) and allocate the rest as anonymous shared memory via mmap.
This will hopefully spare most users the hassle of adjusting operating
system parameters before being able to start PostgreSQL with a
reasonable value for shared_buffers.

There are a bunch of documentation updates needed here, and we might
need to adjust some of the HINT messages related to shared memory as
well.  But it's not 100% clear how portable this is, so before we
write the documentation, let's give it a spin on the buildfarm and
see what turns red.

This has apparently been wrong since event_source was added.

Alexander Lakhin

The callers of UtilityContainsQuery want it to return a non-utility Query
if it returns anything at all.  However, since we made CREATE TABLE
AS/SELECT INTO into a utility command instead of a variant of SELECT,
a command like "EXPLAIN SELECT INTO" results in two nested utility
statements.  So what we need UtilityContainsQuery to do is drill down
to the bottom non-utility Query.

I had thought of this possibility in setrefs.c, and fixed it there by
looping around the UtilityContainsQuery call; but overlooked that the call
sites in plancache.c have a similar issue.  In those cases it's
notationally inconvenient to provide an external loop, so let's redefine
UtilityContainsQuery as recursing down to a non-utility Query instead.

Noted by Rushabh Lathia.  This is a somewhat cleaned-up version of his
proposed patch.