Author: Sawada Masahiko

Most of the Windows buildfarm members (bowerbird, hamerkop, currawong,
jacana, brolga) are unhappy with yesterday's abbreviated keys patch,
although there are some (narwhal, frogmouth) that seem OK with it.
Since there's no obvious pattern to explain why some are working and
others are failing, just disable this across-the-board on Windows for
now.  This is a bit unfortunate since the optimization will be a big
win in some cases, but we can't leave the buildfarm broken.

This dumps the predefined preprocessor macros

Per buildfarm failure on bowerbird after last night's commit
4ea51cdf.

Peter Geoghegan

This results in a very substantial reduction in disk space usage during
"make check-world", since that sequence involves creation of numerous
temporary installations.  It should also help a bit in the buildfarm, even
though the buildfarm script doesn't create as many temp installations,
because the current script misses deleting some of them; and anyway it
seems better to do this once in one place rather than expecting that
script to get it right every time.

In 9.4 and HEAD, also undo the unwise choice in commit b1aebbb6
to report strerror(errno) after a rmtree() failure.  rmtree has already
reported that, possibly for multiple failures with distinct errnos; and
what's more, by the time it returns there is no good reason to assume
that errno still reflects the last reportable error.  So reporting errno
here is at best redundant and at worst badly misleading.

Back-patch to all supported branches, so that future revisions of the
buildfarm script can rely on this behavior.

Per discussion, change the log level of this message to be LOG not WARNING.
The main point of this change is to avoid causing buildfarm run failures
when the stats collector is exceptionally slow to respond, which it not
infrequently is on some of the smaller/slower buildfarm members.

This change does lose notice to an interactive user when his stats query
is looking at out-of-date stats, but the majority opinion (not necessarily
that of yours truly) is that WARNING messages would probably not get
noticed anyway on heavily loaded production systems.  A LOG message at
least ensures that the problem is recorded somewhere where bulk auditing
for the issue is possible.

Also, instead of an untranslated "pgstat wait timeout" message, provide
a translatable and hopefully more understandable message "using stale
statistics instead of current ones because stats collector is not
responding".  The original text was written hastily under the assumption
that it would never really happen in practice, which we now know to be
unduly optimistic.

Back-patch to all active branches, since we've seen the buildfarm issue
in all branches.

As noted by Tom Lane the improvements in 4b4b680c had the problem
that in some situations we searched, entered and modified entries in
the private refcount hash while holding a spinlock. I had tried to
keep the logic entirely local to PinBuffer_Locked(), but that's not
really possible given it's called with a spinlock held...

Besides being disadvantageous from a performance point of view, this
also has problems with error handling safety. If we failed inserting
an entry into the hashtable due to an out of memory error, we'd error
out with a held spinlock. Not good.

Change the way private refcounts are manipulated: Before a buffer can
be tracked an entry has to be reserved using
ReservePrivateRefCountEntry(); then, if a entry is not found using
GetPrivateRefCountEntry(), it can be entered with
NewPrivateRefCountEntry().

Also take advantage of the fact that PinBuffer_Locked() currently is
never called for buffers that already have been pinned by the current
backend and don't search the private refcount entries for preexisting
local pins. That results in a small, but measurable, performance
improvement.

Additionally make ReleaseBuffer() always call UnpinBuffer() for shared
buffers. That avoids duplicating work in an eventual UnpinBuffer()
call that already has been done in ReleaseBuffer() and also saves some
code.

Per discussion with Tom Lane.

Discussion: 15028.1418772313@sss.pgh.pa.us

This commit extends the SortSupport infrastructure to allow operator
classes the option to provide abbreviated representations of Datums;
in the case of text, we abbreviate by taking the first few characters
of the strxfrm() blob.  If the abbreviated comparison is insufficent
to resolve the comparison, we fall back on the normal comparator.
This can be much faster than the old way of doing sorting if the
first few bytes of the string are usually sufficient to resolve the
comparison.

There is the potential for a performance regression if all of the
strings to be sorted are identical for the first 8+ characters and
differ only in later positions; therefore, the SortSupport machinery
now provides an infrastructure to abort the use of abbreviation if
it appears that abbreviation is producing comparatively few distinct
keys.  HyperLogLog, a streaming cardinality estimator, is included in
this commit and used to make that determination for text.

Peter Geoghegan, reviewed by me.

Etsuro Fujita

Pointed out by Alan Mogi (bug #12571)

Amit Langote

This was previously only done for libpq, not it's done for all shared
libraries.
Reviewed-by: Michael Paquier <michael.paquier@gmail.com>

The code used sizeof(ItemPointerData) where sizeof(ItemIdData) is correct,
since we're trying to account for a tuple's line pointer.  Spotted by
Tomonari Katsumata (bug #12584).

Although this mistake is of very long standing, no back-patch, since it's
a relatively harmless error and changing it would risk changing default
planner behavior in stable branches.  (I don't see any change in regression
test outputs here, but the buildfarm may think differently.)

Elaborated from an idea by Andres Freund.

pg_dump.c:dumDatabase() called ArchiveEntry() with the results of a a
query that was PQclear()ed a couple lines earlier.

Backpatch to 9.2 where security labels for shared objects where
introduced.

Relying on the normal shared latch simplifies interrupt/signal
handling because we can rely on all signal handlers setting the proc
latch. That in turn allows us to avoid the use of
ImmediateInterruptOK, which arguably isn't correct because
WaitLatchOrSocket isn't declared to be immediately interruptible.

Also change sections that wait on the walsender's latch to notice
interrupts quicker/more reliably and make them more consistent with
each other.

This is part of a larger "get rid of ImmediateInterruptOK" series.

Discussion: 20150115020335.GZ5245@awork2.anarazel.de

Up to now, EXPLAIN has contented itself with printing the sort expressions
in a Sort or Merge Append plan node.  This patch improves that by
annotating the sort keys with COLLATE, DESC, USING, and/or NULLS FIRST/LAST
whenever nondefault sort ordering options are used.  The output is now a
reasonably close approximation of an ORDER BY clause equivalent to the
plan's ordering.

Marius Timmer, Lukas Kreft, and Arne Scheffer; reviewed by Mike Blackwell.
Some additional hacking by me.

Currently, a backend will reset it's PGXACT->xmin value when it doesn't
have any registered snapshots left. That covered the common case that a
transaction in read committed mode runs several queries, one after each
other, as there would be no snapshots active between those queries.
However, if you hold cursors across each of the query, we didn't get a
chance to reset xmin.

To make that better, keep all the registered snapshots in a pairing heap,
ordered by xmin so that it's always quick to find the snapshot with the
smallest xmin. That allows us to advance PGXACT->xmin whenever the oldest
snapshot is deregistered, even if there are others still active.

Per discussion originally started by Jeff Davis back in 2009 and more
recently by Robert Haas.

For no significant extra complexity, we can cache knowledge that the
target page is lossy, and save a hash_search per iteration in that
case as well.  This probably makes little difference, since the extra
rechecks that must occur when pages are lossy are way more expensive
than anything we can save here ... but we might as well do it if we're
going to cache anything.

When adding a large number of tuples to a TID bitmap using
tbm_add_tuples() sometimes a lot of time was spent looking up a page's
entry in the bitmap's internal hashtable.

Improve efficiency by caching the last accessed page, while iterating
over the passed in tuples, hoping consecutive tuples will often be on
the same page.  In many cases that's a good bet, and in the rest the
added overhead isn't big.

Discussion: 54479A85.8060309@sigaev.ru

Author: Teodor Sigaev
Reviewed-By: David Rowley

Previous fix mapped "Norwegian (Bokmål)" locale, which contains a non-ASCII
character, to the pure ASCII alias "norwegian-bokmal". However, it turns
out that more recent versions of the CRT library, in particular MSVCR110
(Visual Studio 2012), changed the behaviour of setlocale() so that if
you pass "norwegian-bokmal" to setlocale, it returns "Norwegian_Norway".

That meant trouble, when setlocale(..., NULL) first returned
"Norwegian (Bokmål)_Norway", which we mapped to "norwegian-bokmal_Norway",
but another call to setlocale(..., "norwegian-bokmal_Norway") returned
"Norwegian_Norway". That caused PostgreSQL to think that they are different
locales, and therefore not compatible. That caused initdb to fail at
CREATE DATABASE.

Older CRT versions seem to accept "Norwegian_Norway" too, so change the
mapping to return "Norwegian_Norway" instead of "norwegian-bokmal".

Backpatch to 9.2 like the previous attempt. We haven't made a release that
includes the previous fix yet, so we don't need to worry about changing the
locale of existing clusters from "norwegian-bokmal" to "Norwegian_Norway".
(Doing any mapping like this at all requires changing the locale of
existing databases; the release notes need to include instructions for
that).

Commit 894459e5 revealed this option to
be broken for NLS builds on Darwin, but "make -C contrib/unaccent check"
and the buildfarm client rely on it.  Fix that configuration by
redefining the option to imply LANG=C on Darwin.  In passing, use LANG=C
instead of LANG=en on Windows; since only postmaster startup uses that
value, testers are unlikely to notice the change.  Back-patch to 9.0,
like the predecessor commit.

Up to now, the "child" executor state trees generated for EvalPlanQual
rechecks have simply shared the ResultRelInfo arrays used for the original
execution tree.  However, this leads to dangling-pointer problems, because
ExecInitModifyTable() is all too willing to scribble on some fields of the
ResultRelInfo(s) even when it's being run in one of those child trees.
This trashes those fields from the perspective of the parent tree, because
even if the generated subtree is logically identical to what was in use in
the parent, it's in a memory context that will go away when we're done
with the child state tree.

We do however want to share information in the direction from the parent
down to the children; in particular, fields such as es_instrument *must*
be shared or we'll lose the stats arising from execution of the children.
So the simplest fix is to make a copy of the parent's ResultRelInfo array,
but not copy any fields back at end of child execution.

Per report from Manuel Kniep.  The added isolation test is based on his
example.  In an unpatched memory-clobber-enabled build it will reliably
fail with "ctid is NULL" errors in all branches back to 9.1, as a
consequence of junkfilter->jf_junkAttNo being overwritten with $7f7f.
This test cannot be run as-is before that for lack of WITH syntax; but
I have no doubt that some variant of this problem can arise in older
branches, so apply the code change all the way back.

The flag is supposed to be copied from the record. Same issue with
track_commit_timestamps, but that's master-only.

Report and fix by Petr Jalinek. Backpatch to 9.4, where wal_log_hints was
added.

The folly of the previous arrangement was just demonstrated: there's no
convenient way to add fields to ExplainState without breaking ABI, even
if callers have no need to touch those fields.  Since we might well need
to do that again someday in back branches, let's change things so that
only explain.c has to have sizeof(ExplainState) compiled into it.  This
costs one extra palloc() per EXPLAIN operation, which is surely pretty
negligible.

As of 9.3, ruleutils.c goes to some lengths to ensure that table and column
aliases used in its output are unique.  Of course this takes more time than
was required before, which in itself isn't fatal.  However, EXPLAIN was set
up so that recalculation of the unique aliases was repeated for each
subexpression printed in a plan.  That results in O(N^2) time and memory
consumption for large plan trees, which did not happen in older branches.

Fortunately, the expensive work is the same across a whole plan tree,
so there is no need to repeat it; we can do most of the initialization
just once per query and re-use it for each subexpression.  This buys
back most (not all) of the performance loss since 9.2.

We need an extra ExplainState field to hold the precalculated deparse
context.  That's no problem in HEAD, but in the back branches, expanding
sizeof(ExplainState) seems risky because third-party extensions might
have local variables of that struct type.  So, in 9.4 and 9.3, introduce
an auxiliary struct to keep sizeof(ExplainState) the same.  We should
refactor the APIs to avoid such local variables in future, but that's
material for a separate HEAD-only commit.

Per gripe from Alexey Bashtanov.  Back-patch to 9.3 where the issue
was introduced.

Previously, read() might have returned a length equal to the buffer
length, and then the subsequent store to buf[len] would write a
zero-byte one byte past the end.  This doesn't seem likely to be
a security issue, but there's some chance it could result in
pg_standby misbehaving.

Spotted by Coverity; patch by Michael Paquier, reviewed by me.

The possibly, depending on compiler settings, generated warning was
"warning: `S_UNLOCK' redefined".

The hppa spinlock implementation doesn't follow the rules of s_lock.h
and provides a gcc specific implementation outside of the the part of
the file that's supposed to do that.  It does so to avoid duplication
between the HP compiler and gcc. That unfortunately means that
S_UNLOCK is already defined when the HPPA specific section is reached.

Undefine the generic fallback S_UNLOCK definition inside the HPPA
section. That's far from pretty, but has the big advantage of being
simple. If somebody is interested to fix this in a prettier way...

This presumably got broken in the course of 0709b7ee.

Discussion: 20150114225919.GY5245@awork2.anarazel.de

Per complaint from Tom Lane.

Michael Paquier

Spotted by Coverity.  This isn't likely to matter in practice, but
there's no harm in fixing it.

Michael Paquier

To do so, move InitializeLatchSupport() into the new common process
initialization functions, and add a new global variable MyLatch.

MyLatch is usable as soon InitPostmasterChild() has been called
(i.e. very early during startup). Initially it points to a process
local latch that exists in all processes. InitProcess/InitAuxiliaryProcess
then replaces that local latch with PGPROC->procLatch. During shutdown
the reverse happens.

This is primarily advantageous for two reasons: For one it simplifies
dealing with the shared process latch, especially in signal handlers,
because instead of having to check for MyProc, MyLatch can be used
unconditionally. For another, a later patch that makes FEs/BE
communication use latches, now can rely on the existence of a latch,
even before having gone through InitProcess.

Discussion: 20140927191243.GD5423@alap3.anarazel.de

Previously, configure would add any switches that it chose of its own
accord to the end of the user-specified CFLAGS string.  Since most
compilers process these left-to-right, this meant that configure's choices
would override the user-specified flags in case of conflicts.  We'd rather
that worked the other way around, so adjust the logic to put the user's
string at the end not the beginning.

There does not seem to be a need for a similar behavior change for CPPFLAGS
or LDFLAGS: in those, the earlier switches tend to win (think -I or -L
behavior) so putting the user's string at the front is fine.

Backpatch to 9.4 but not earlier.  I'm not planning to run buildfarm member
guar on older branches, and it seems a bit risky to change this behavior
in long-stable branches.

Autoconf has known about automatically selecting -Ae when needed for
quite some time now, so remove the redundant addition in template/hpux.
Noted while setting up buildfarm member pademelon.

Since commit 626eb021 has introduced the auxiliary process
infrastructure, bootstrap_signals() was never used when forked from
postmaster.

Remove the IsUnderPostmaster specific code, and add a appropriate
assertion.

Move common code, that was duplicated in every postmaster child/every
standalone process, into two functions in miscinit.c.  Not only does
that already result in a fair amount of net code reduction but it also
makes it much easier to remove more duplication in the future. The
prime motivation wasn't code deduplication though, but easier addition
of new common code.

Commit b94ce6e8 reordered postmaster's startup sequence so that the
tempfile directory is only cleaned up after all the necessary state
for pg_ctl is collected.  Unfortunately the chosen location is after
the syslogger has been started; which normally is fine, except for
!WIN32 EXEC_BACKEND builds, which pass information to children via
files in the temp directory.

Move the call to RemovePgTempFiles() to just before the syslogger has
started. That's the first child we fork.

Luckily EXEC_BACKEND is pretty much only used by endusers on windows,
which has a separate method to pass information to children. That
means the real world impact of this bug is very small.

Discussion: 20150113182344.GF12272@alap3.anarazel.de

Backpatch to 9.1, just as the previous commit was.

Also use lower-case for a few more features, to be consistent with the
others and with the SQL spec.

Similar warnings from backend were silenced earlier by commit c8315930,
but there were a few more contrib/hstore.

Michael Paquier

Since their introduction latches have required barriers in SetLatch
and ResetLatch - but when they were introduced there wasn't any
barrier abstraction. Instead latches were documented to rely on the
callsites to provide barrier semantics.

Now that the barrier support looks halfway complete, add the necessary
barriers to both latch implementations.

Also remove a now superflous lock acquisition from syncrep.c and a
superflous (and insufficient) barrier from freelist.c. There might be
other cases that can now be simplified, but those are the only ones
I've seen on a quick scan.

We might want to backpatch this at some later point, but right now the
barrier infrastructure in the backbranches isn't totally on par with
master.

Discussion: 20150112154026.GB2092@awork2.anarazel.de

So far WaitLatchOrSocket() required to pass in WL_SOCKET_READABLE as
that solely was used to indicate error conditions, like EOF. Waiting
for WL_SOCKET_WRITEABLE would have meant to busy wait upon socket
errors.

Adjust the API to signal errors by returning the socket as readable,
writable or both, depending on WL_SOCKET_READABLE/WL_SOCKET_WRITEABLE
being specified.  It would arguably be nicer to return WL_SOCKET_ERROR
but that's not possible on platforms and would probably also result in
more complex callsites.

This previously had explicitly been forbidden in e42a21b9, as
there was no strong use case at that point. We now are looking into
making FE/BE communication use latches, so changing this makes sense.

There also are some portability concerns because there cases of older
platforms where select(2) is known to, in violation of POSIX, not
return a socket as writable after the peer has closed it.  So far the
platforms where that's the case provide a working poll(2). If we find
one where that's not the case, we'll need to add a workaround for that
platform.

Discussion: 20140927191243.GD5423@alap3.anarazel.de
Reviewed-By: Heikki Linnakangas, Noah Misch