Erroneously omitted in 261c7d4b

If POSIX shared memory is supported and works, we prefer it.
Otherwise, we prefer System V, except on Windows, where we use
the implementation specific to that platform.

The buildfarm pointed out the problem.

Fix based on suggestion by Robert Haas.

Andrew Tipton.

Per buildfarm.

This allows decorating pg_fatal() with noreturn compiler hints, leading
to better diagnostics.
Reviewed-by: Marko Tiikkaja <marko@joh.to>

Change the input/output format to {A,B,C}, to match the internal
representation.

Complete the implementations of line_in, line_out, line_recv, line_send.
Remove comments and error messages about the line type not being
implemented.  Add regression tests for existing line operators and
functions.
Reviewed-by: rui hua <365507506hua@gmail.com>
Reviewed-by: Álvaro Herrera <alvherre@2ndquadrant.com>
Reviewed-by: Jeevan Chalke <jeevan.chalke@enterprisedb.com>

Patch by myself and Amit Kapila.  Design help from Noah Misch.  Review
by Andres Freund.

REFRESH MATERIALIZED VIEW CONCURRENTLY was broken for any matview
containing a column of a type without a default btree operator
class.  It also did not produce results consistent with a non-
concurrent REFRESH or a normal view if any column was of a type
which allowed user-visible differences between values which
compared as equal according to the type's default btree opclass.
Concurrent matview refresh was modified to use the new operators
to solve these problems.

Documentation was added for record comparison, both for the
default btree operator class for record, and the newly added
operators.  Regression tests now check for proper behavior both
for a matview with a box column and a matview containing a citext
column.

Reviewed by Steve Singer, who suggested some of the doc language.

Backpatch through 8.4

Per suggestion by Amit Langote

Previously set_default_effective_cache_size() could not handle fork,
non-fork, and bootstrap cases.

It previously reported failure at REPEATABLE READ and SERIALIZABLE
transaction isolation levels for make installcheck.

Suggestion from Richard Neill

Suggestion from Pavel Stehule

The TYPEALIGN macro, and the related ones like MAXALIGN, don't work with
values larger than intptr_t, because TYPEALIGN casts the argument to
intptr_t to do the arithmetic. That's not a problem when dealing with
pointers or lengths or offsets related to pointers, but the XLogInsert
scaling patch added a call to MAXALIGN with an XLogRecPtr argument.

To fix, add wider variants of the macros, called TYPEALIGN64 and MAXALIGN64,
which are just like the existing variants but work with uint64 instead of
intptr_t.

Report and patch by David Rowley, analysis by Andres Freund.

1. In heap_hot_search_buffer(), the PredicateLockTuple() call is passed
wrong offset number. heapTuple->t_self is set to the tid of the first
tuple in the chain that's visited, not the one actually being read.

2. CheckForSerializableConflictIn() uses the tuple's t_ctid field
instead of t_self to check for exiting predicate locks on the tuple. If
the tuple was updated, but the updater rolled back, t_ctid points to the
aborted dead tuple.

Reported by Hannu Krosing. Backpatch to 9.1.

Otherwise, we access already-freed memory.  Oops.

Report by Michael Paquier.  Fix by me.

This option provides more detailed error messages when STRICT is used
and the number of rows returned is not one.

Marko Tiikkaja, reviewed by Ian Lawrence Barwick

If a tuple was frozen while its predicate locks mattered,
read-write dependencies could be missed, resulting in failure to
detect conflicts which could lead to anomalies in committed
serializable transactions.

This field was added to the tag when we still thought that it was
necessary to carry locks forward to a new version of an updated
row.  That was later proven to be unnecessary, which allowed
simplification of the code, but elimination of xmin from the tag
was missed at the time.

Per report and analysis by Heikki Linnakangas.
Backpatch to 9.1.

Cédric Villemain and me.

Reviewed by Fabien COELHO.

Clamp the minimum sleep time during immediate shutdown or crash to a
minimum of zero, not a maximum of one second.  The previous code could
result in a negative sleep time, leading to failure in select() calls.

Also, on crash recovery, reset AbortStartTime as soon as SIGKILL is sent
or abort processing has commenced instead of waiting until the startup
process completes.  Reset AbortStartTime as soon as SIGKILL is sent,
too, to avoid doing that repeatedly.

Per trouble report from Jeff Janes on
CAMkU=1xd3=wFqZwwuXPWe4BQs3h1seYo8LV9JtSjW5RodoPxMg@mail.gmail.com

Author: MauMau

Isolate transaction latency (elapsed time between submitting first
command and receiving response to last command) from client-side delays
pertaining to the --rate schedule.  Under --rate, report schedule lag as
defined in the documentation.  Report latency standard deviation
whenever we collect the measurements to do so.  All of these changes
affect --progress messages and the final report.

Fabien COELHO, reviewed by Pavel Stehule.

Oversight in commit 4a87f308.

Fabien COELHO

Mention that link mode uses less disk space, and uses junction points on
Windows.

Backpatch to 9.3.

Also, make sure they pass on all transaction isolation levels.

Previously, isolationtester would forbid returning tuples in
session-specific teardown (but not global teardown), as well as in
global setup.  Allow these places to return tuples, too.

Issue error for SET LOCAL/CONSTRAINTS/TRANSACTION outside a transaction
block, as they have no effect.

Per suggestion from Morten Hustveit