== PostgreSQL Weekly News - July 5, 2020 ==

== PostgreSQL Weekly News - July 5, 2020 ==

Foss4G has been cancelled. The next FOSS4G event will take place in August 2021
in Buenos Ares, Argentina.
https://2020.foss4g.org/cancellation/

Person of the week: https://postgresql.life/post/andrew_dunstan/

== PostgreSQL Product News ==

ORC Foreign Data Wrapper 1.0.0, a FDW for files in ORC format, released.
https://github.com/HighgoSoftware/orc_fdw/releases/tag/ORC_FDW-REL-1-0

pgAdmin4 4.23, a web- and native GUI control center for PostgreSQL, released.
https://www.pgadmin.org/docs/pgadmin4/dev/release_notes_4_23.html

pg_probackup 2.4.1, a utility to manage backup and recovery of PostgreSQL
database clusters, released.
https://github.com/postgrespro/pg_probackup/releases/tag/2.4.1

CYPEX 1.0, an application building toolkit for PostgreSQL, released.
https://www.cybertec-postgresql.com/en/products/cypex/

== PostgreSQL Jobs for July ==

http://archives.postgresql.org/pgsql-jobs/2020-07/

== PostgreSQL Local ==

PGDay Ukraine will take place September 5th, 2020 in Lviv at the Bank Hotel.
https://pgday.org.ua/

pgDay Israel 2020 will take place on September 10, 2020 in Tel Aviv.
http://pgday.org.il/

PGDay Austria will take place September 18, 2020 at Schloss Schoenbrunn
(Apothekertrakt) in Vienna.
https://pgday.at/en/

PostgreSQL Conference Europe 2020 will be held on October 20-23, 2020 in Berlin,
Germany. The CfP is open through July 31, 2020 at https://2020.pgconf.eu/callforpapers
https://2020.pgconf.eu/

PG Day Russia will take place in Saint Petersburg on July 9, 2021.
https://pgday.ru/en/2020/

== PostgreSQL in the News ==

Planet PostgreSQL: http://planet.postgresql.org/

PostgreSQL Weekly News is brought to you this week by David Fetter

Submit news and announcements by Sunday at 3:00pm PST8PDT to david(at)fetter(dot)org(dot)

== Applied Patches ==

Michaël Paquier pushed:

- Refactor ObjectAddress field assignments for type dependencies. The logic used
to build the set of dependencies needed for a type is rather repetitive with
direct assignments for each ObjectAddress field. This refactors the code to
use the macro ObjectAddressSet() instead, to do the same work. There are more
areas of the backend code that could use this macro, but these are left for a
follow-up patch that will partially rework the way dependencies are recorded
as well. Type dependencies are left out of the follow-up patch, so they are
refactored separately here. Extracted from a larger patch by the same author.
Author: Daniel Gustafsson Discussion:
https://potgr.es/m/20190213182737.mxn6hkdxwrzgxk35@alap3.anarazel.de
https://git.postgresql.org/pg/commitdiff/68de1440c79d75e529ff8c7395d698252370f992

- Prevent compilation of frontend-only files in src/common/ with backend. Any
frontend-only file of src/common/ should include a protection to prevent such
code to be included in the backend compilation. fe_memutils.c and
restricted_token.c have been doing that, while file_utils.c (since bf5bb2e)
and logging.c (since fc9a62a) forgot it. Reviewed-by: Daniel Gustafsson
Discussion: https://postgr.es/m/20200625080757.GI130132@paquier.xyz
https://git.postgresql.org/pg/commitdiff/324435eb14e4f41cd430f96c9b13ad9b160e45e4

- Fix removal of files generated by TAP tests for SSL. 001_ssltests.pl and
002_scram.pl both generated an extra file for a client key used in the tests
that were not removed. In Debian, this causes repeated builds to fail. The
code refactoring done in 4dc6355 broke the cleanup done in 001_ssltests.pl,
and the new tests added in 002_scram.pl via d6e612f forgot the removal of one
file. While on it, fix a second issue introduced in 002_scram.pl where we use
the same file name in 001 and 002 for the temporary client key whose
permissions are changed in the test, as using the same file name in both tests
could cause failures with parallel jobs of src/test/ssl/ if one test removes a
file still needed by the second test. Reported-by: Felix Lechner Author:
Daniel Gustafsson, Felix Lechner Reviewed-by: Tom Lane, Michael Paquier
Discussion:
https://postgr.es/m/CAFHYt543sjX=Cm_aEeoejStyP47C+Y3+Wh6WbirLXsgUMaw7iw@mail.gmail.com
Backpatch-through: 13
https://git.postgresql.org/pg/commitdiff/c4342c932abb6792f6c7d8528b6224ca4d5e1a84

- Refactor creation of normal dependency records when creating extension. When
creating an extension, the same type of dependency is used when registering a
dependency to a schema and required extensions. This improves the code so as
those dependencies are not recorded one-by-one, but grouped together. Note
that this has as side effect to remove duplicate dependency entries, even if
it should not happen in practice as extensions listed as required in a control
file should be listed only once. Extracted from a larger patch by the same
author. Author: Daniel Dustafsson Discussion:
https://postgr.es/m/20200629065535.GA183079@paquier.xyz
https://git.postgresql.org/pg/commitdiff/684b4f29b729e5b135d256802d6e712222b9551f

- Refactor ObjectAddress field assignments in more places. This is a follow-up
commit similar to 68de144, with more places in the backend code simplified
with the macros able to assign values to the fields of ObjectAddress. The
code paths changed here could be transitioned later into using more grouping
when inserting dependency records, simplifying this future work. Author:
Daniel Gustafsson, Michael Paquier Discussion:
https://postgr.es/m/20190213182737.mxn6hkdxwrzgxk35@alap3.anarazel.de
https://git.postgresql.org/pg/commitdiff/4315e8c23b9a897e12fcf91de7bfd734621096bf

- Move description of libpqwalreceiver hooks out of the replication's README.
src/backend/replication/README includes since 32bc08b a basic description of
the WAL receiver hooks available in walreceiver.h for a module like
libpqwalreceiver, but the README has never been updated to reflect changes
done to the hooks, so the contents of the README have rotten with the time.
This commit moves the description from the README to walreceiver.h, where it
will be hard to miss that a description update or addition is needed depending
on the modifications done to the hooks. Each hook now includes a description
of what it does in walreceiver.h, and the replication's README mentions
walreceiver.h. Thanks also to Amit Kapila for the discussion. Author:
Michael Paquier Reviewed-by: Peter Eisentraut Discussion:
https://postgr.es/m/20200502024606.GA471944@paquier.xyz
https://git.postgresql.org/pg/commitdiff/641dd167a3424f97eed8416b5ef1c21dbbf8cf12

- doc: Fix incorrect reference to textout in plpgsql examples. This error has
survived for 22 years, and has been introduced by da63386. Reported-by: Erwin
Brandstetter Discussion:
https://postgr.es/m/CAGHENJ57wogGOvGXo5LgWYcqswxafLck8ELqHDR+zrkTPgs_OQ@mail.gmail.com
Backpatch-through: 9.5
https://git.postgresql.org/pg/commitdiff/8cea55e99b766fd654431f24987a774622594c9b

Peter Eisentraut pushed:

- Clean up grammar a bit. Simplify the grammar specification of substring() and
overlay() a bit, simplify and update some comments. Reviewed-by: Pavel
Stehule <pavel(dot)stehule(at)gmail(dot)com> Reviewed-by: Vik Fearing
<vik(at)postgresfriends(dot)org> Reviewed-by: Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr>
Discussion:
https://www.postgresql.org/message-id/flat/a15db31c-d0f8-8ce0-9039-578a31758adb%402ndquadrant.com
https://git.postgresql.org/pg/commitdiff/aafefb4dcbf79e8cb1439e888a9cdb3dfefa7657

- Add current substring regular expression syntax. SQL:1999 had syntax
SUBSTRING(text FROM pattern FOR escapechar) but this was replaced in SQL:2003
by the more clear SUBSTRING(text SIMILAR pattern ESCAPE escapechar) but
this was never implemented in PostgreSQL. This patch adds that new syntax as
an alternative in the parser, and updates documentation and tests to indicate
that this is the preferred alternative now. Reviewed-by: Pavel Stehule
<pavel(dot)stehule(at)gmail(dot)com> Reviewed-by: Vik Fearing <vik(at)postgresfriends(dot)org>
Reviewed-by: Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr> Discussion:
https://www.postgresql.org/message-id/flat/a15db31c-d0f8-8ce0-9039-578a31758adb%402ndquadrant.com
https://git.postgresql.org/pg/commitdiff/78c887679d7632c1211f85eb95723f3226bf1b46

- pgstattuple: Have pgstattuple_approx accept TOAST tables. TOAST tables have a
visibility map and a free space map, so they can be supported by
pgstattuple_approx just fine. Add test cases to show how various pgstattuple
functions accept TOAST tables. Also add similar tests to pg_visibility, which
already accepted TOAST tables correctly but had no test coverage for them.
Reviewed-by: Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at> Discussion:
https://www.postgresql.org/message-id/flat/27c4496a-02b9-dc87-8f6f-bddbef54e0fe(at)2ndquadrant(dot)com
https://git.postgresql.org/pg/commitdiff/ee0202d552791f14bc407ce299628ce8d50eebe3

- Rename enable_incrementalsort for clarity. Author: James Coleman
<jtc331(at)gmail(dot)com> Discussion:
https://www.postgresql.org/message-id/flat/df652910-e985-9547-152c-9d4357dc3979%402ndquadrant.com
https://git.postgresql.org/pg/commitdiff/e61225ffab3f8726ecd053b1c22ff249c417f9ba

- doc: Spell checking.
https://git.postgresql.org/pg/commitdiff/90b2d8c1adc86d1fc46f83c73444a1fb5b28dd3c

Tom Lane pushed:

- Avoid using %c printf format for potentially non-ASCII characters. Since %c
only passes a C "char" to printf, it's incapable of dealing with multibyte
characters. Passing just the first byte of such a character leads to an
output string that is visibly not correctly encoded, resulting in undesirable
behavior such as encoding conversion failures while sending error messages to
clients. We've lived with this issue for a long time because it was
inconvenient to avoid in a portable fashion. However, now that we always use
our own snprintf code, it's reasonable to use the %.*s format to print just
one possibly-multibyte character in a string. (We previously avoided that
obvious-looking answer in order to work around glibc's bug #6530, cf commits
54cd4f045 and ed437e2b2.) Hence, run around and fix a bunch of places that
used %c to report a character found in a user-supplied string. For
simplicity, I did not touch places that were emitting non-user-facing debug
messages, or reporting catalog data that should always be ASCII. (It's also
unclear how useful this approach could be in frontend code, where it's less
certain that we know what encoding we're dealing with.) In passing, improve a
couple of poorly-written error messages in pageinspect/heapfuncs.c. This is a
longstanding issue, but I'm hesitant to back-patch because of the impact on
translatable message strings. In any case this fix would not work reliably
before v12. Tom Lane and Quan Zongliang Discussion:
https://postgr.es/m/a120087c-4c88-d9d4-1ec5-808d7a7f133d@gmail.com
https://git.postgresql.org/pg/commitdiff/16e3ad5d143795b05a21dc887c2ab384cce4bcb8

- Mop up some no-longer-necessary hacks around printf %.*s format. Commit
54cd4f045 added some kluges to work around an old glibc bug, namely that %.*s
could misbehave if glibc thought any characters in the supplied string were
incorrectly encoded. Now that we use our own snprintf.c implementation, we
need not worry about that bug (even if it still exists in the wild). Revert a
couple of particularly ugly hacks, and remove or improve assorted comments.
Note that there can still be encoding-related hazards here: blindly clipping
at a fixed length risks producing wrongly-encoded output if the clip splits a
multibyte character. However, code that's doing correct multibyte-aware
clipping doesn't really need a comment about that, while code that isn't needs
an explanation why not, rather than a red-herring comment about an obsolete
bug. Discussion: https://postgr.es/m/279428.1593373684@sss.pgh.pa.us
https://git.postgresql.org/pg/commitdiff/c410af098c46949e36607eb13689e697fa2def97

- Remove support for timezone "posixrules" file. The IANA tzcode library has a
feature to read a time zone file named "posixrules" and apply the
daylight-savings transition dates and times therein, when it is given a
POSIX-style time zone specification that lacks an explicit transition rule.
However, there's a problem with that code: it doesn't work for dates past the
Y2038 time_t rollover. (Effectively, all times beyond that point are treated
as standard time.) The IANA crew regard this feature as legacy, so their plan
is to remove it not fix it. The time frame in which that will happen is
unclear, but presumably it'll happen well before 2038. Moreover, effective
with the next IANA data update (probably this fall), the recommended default
will be to not install a "posixrules" file in the first place. The time frame
in which tzdata packagers might adopt that suggestion is likewise unclear, but
at least some platforms will probably do it in the next year or so. While we
could ignore that recommendation so far as PG-supplied tzdata trees are
concerned, builds using --with-system-tzdata will be subject to whatever the
platform's tzdata packager decides to do. Thus, whether or not we do
anything, some increasing fraction of Postgres users will be exposed to the
behavior observed when there is no "posixrules" file; and if we do nothing,
we'll have essentially no control over the timing of that change. The best
thing to do to ameliorate the uncertainty seems to be to proactively remove
the posixrules-reading feature. If we do that in a scheduled release then at
least we can release-note the behavioral change, rather than having users be
surprised by it after a routine tzdata update. The change in question is
fairly minor anyway: to be affected, you have to be using a POSIX-style
timezone spec, it has to not have an explicit rule, and it has to not be one
of the four traditional continental-USA zone names (EST5EDT, CST6CDT, MST7MDT,
or PST8PDT), as those are special-cased. Since the default "posixrules" file
provides USA DST rules, the number of people who are likely to find such a
zone spec useful is probably quite small. Moreover, the fallback behavior
with no explicit rule and no "posixrules" file is to apply current USA rules,
so the only thing that really breaks is the DST transitions in years before
2007 (and you get the countervailing fix that transitions after 2038 will be
applied). Now, some installations might have replaced the "posixrules" file,
allowing e.g. EU rules to be applied to a POSIX-style timezone spec. That
won't work anymore. But it's not exactly clear why this solution would be
preferable to using a regular named zone. In any case, given the Y2038 issue,
we need to be pushing users to stop depending on this. Back-patch into v13;
it hasn't been released yet, so it seems OK to change its behavior.
(Personally I think we ought to back-patch further, but I've been outvoted.)
Discussion: https://postgr.es/m/1390.1562258309@sss.pgh.pa.us Discussion:
https://postgr.es/m/20200621211855.6211-1-eggert@cs.ucla.edu
https://git.postgresql.org/pg/commitdiff/ea57e531b9487e042131ca1151a3ef5d655f40ec

- Inline plpgsql's exec_stmt() into exec_stmts(). This saves one level of C
function call per plpgsql statement executed, and permits a tiny additional
optimization of not saving and restoring estate->err_stmt for each statement
in a block. The net effect seems nearly un-measurable on x86_64, but there's
a clear win on aarch64, amounting to two or three percent in a loop over a few
simple plpgsql statements. To do this, we have to get rid of the other
existing call sites for exec_stmt(). Replace them with exec_toplevel_block(),
which is just defined to do what exec_stmts() does, but for a single
PLpgSQL_stmt_block statement. Hard-wiring the expectation of which statement
type applies here allows us to skip the dispatch switch, making this not much
uglier than the previous factorization. Amit Khandekar, tweaked a bit by me
Discussion:
https://postgr.es/m/CAJ3gD9eBNrmUD7WBBLG8ohaZ485H9y+4eihQTgr+K8Lhka3vcQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/1f902d499eda862a98e881a2227b8d2cf3565374

- Fix temporary tablespaces for shared filesets some more. Commit ecd9e9f0b
fixed the problem in the wrong place, causing unwanted side-effects on the
behavior of GetNextTempTableSpace(). Instead, let's make SharedFileSetInit()
responsible for subbing in the value of MyDatabaseTableSpace when the default
tablespace is called for. The convention about what is in the
tempTableSpaces[] array is evidently insufficiently documented, so try to
improve that. It also looks like SharedFileSetInit() is doing the wrong thing
in the case where temp_tablespaces is empty. It was hard-wiring use of the
pg_default tablespace, but it seems like using MyDatabaseTableSpace is more
consistent with what happens for other temp files. Back-patch the reversion
of PrepareTempTablespaces()'s behavior to 9.5, as ecd9e9f0b was. The changes
in SharedFileSetInit() go back to v11 where that was introduced. (Note there
is net zero code change before v11 from these two patch sets, so nothing to
release-note.) Magnus Hagander and Tom Lane Discussion:
https://postgr.es/m/CABUevExg5YEsOvqMxrjoNvb3ApVyH+9jggWGKwTDFyFCVWczGQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/f7b5988cc0d879d6cc9d481ba898e55747a53bbe

- Clamp total-tuples estimates for foreign tables to ensure planner sanity.
After running GetForeignRelSize for a foreign table, adjust rel->tuples to be
at least as large as rel->rows. This prevents bizarre behavior in
estimate_num_groups() and perhaps other places, especially in the scenario
where rel->tuples is zero because pg_class.reltuples is (suggesting that
ANALYZE has never been run for the table). As things stood, we'd end up
estimating one group out of any GROUP BY on such a table, whereas the default
group-count estimate is more likely to result in a sane plan. Also, clarify
in the documentation that GetForeignRelSize has the option to override the
rel->tuples value if it has a better idea of what to use than what is in
pg_class.reltuples. Per report from Jeff Janes. Back-patch to all supported
branches. Patch by me; thanks to Etsuro Fujita for review Discussion:
https://postgr.es/m/CAMkU=1xNo9cnan+Npxgz0eK7394xmjmKg-QEm8wYG9P5-CcaqQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/ca5e93f769e1bf5e90d080cfcc8c0368ef649a7c

- Inline the fast path of plpgsql's exec_cast_value(). In the common case where
this function isn't actually asked to perform any type conversion, there's
nothing it has to do beyond comparing the arguments. Arrange for that part to
be inlined into callers, with the slower path remaining out-of-line. This
seems to be good for several percent speedup on simple cases, with only
minimal code bloat. Amit Khandekar Discussion:
https://postgr.es/m/CAJ3gD9eBNrmUD7WBBLG8ohaZ485H9y+4eihQTgr+K8Lhka3vcQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/fe2e206cdb00a2d5dcebc3c8e8119017392d9781

Peter Geoghegan pushed:

- nbtree: Correct inaccurate split location comment. Minor oversight in commit
fab25024338.
https://git.postgresql.org/pg/commitdiff/f7a476f0d656bbc673474b9165cd05fa548c18c9

- nbtree: Rename _bt_search() variables. Make some of the variable names in
_bt_search() consistent with corresponding variables within _bt_getstackbuf().
This naming scheme is clearer because the variable names always express a
relationship between the currently locked buffer/page and some other page.
https://git.postgresql.org/pg/commitdiff/e25d462a38f853621a2ad529638d463836230766

- Initialize work_mem using current guc.c default. Do the same for the
maintenance_work_mem global variable. Oversight in commit 848ae330a49, which
increased the previous defaults for work_mem and maintenance_work_mem by 4X.
https://git.postgresql.org/pg/commitdiff/947456a823d6b0973b68c6b38c8623a0504054e7

Fujii Masao pushed:

- Add +(pg_lsn,numeric) and -(pg_lsn,numeric) operators. By using these
operators, the number of bytes can be added into and subtracted from LSN.
Bump catalog version. Author: Fujii Masao Reviewed-by: Kyotaro Horiguchi,
Michael Paquier, Asif Rehman Discussion:
https://postgr.es/m/ed9f7f74-e996-67f8-554a-52ebd3779b3b@oss.nttdata.com
https://git.postgresql.org/pg/commitdiff/9bae7e4cde7c9786ee61dac4a3e032b346350a88

- Change default of pg_stat_statements.track_planning to off. Since v13
pg_stat_statements is allowed to track the planning time of statements when
track_planning option is enabled. Its default was on. But this feature could
cause more terrible spinlock contentions in pg_stat_statements. As a result of
this, Robins Tharakan reported that v13 beta1 showed ~45% performance drop at
high DB connection counts (when compared with v12.3) during fully-cached
SELECT-only test using pgbench. To avoid this performance regression by the
default setting, this commit changes default of
pg_stat_statements.track_planning to off. Back-patch to v13 where
pg_stat_statements.track_planning was introduced. Reported-by: Robins
Tharakan Author: Fujii Masao Reviewed-by: Julien Rouhaud Discussion:
https://postgr.es/m/2895b53b033c47ccb22972b589050dd9@EX13D05UWC001.ant.amazon.com
https://git.postgresql.org/pg/commitdiff/d1763ea8c9c32837d373a196ed0c2e1256a55824

- doc: Correct description of restart_lsn in pg_replication_slots. Previously
the document explained that restart_lsn indicates the LSN of oldest WAL won't
be automatically removed during checkpoints. But since v13 this was no longer
true thanks to max_slot_wal_keep_size. Back-patch to v13 where
max_slot_wal_keep_size was added. Author: Fujii Masao Discussion:
https://postgr.es/m/6497f1e9-3148-c5da-7e49-b2fddad9a42f@oss.nttdata.com
https://git.postgresql.org/pg/commitdiff/8f9b6d40570bd8991f18a089a8445cc5275c1f49

Bruce Momjian pushed:

- doc: change pg_upgrade wal_level to be not minimal. Previously it was
specified to be only replica. Discussion:
https://postgr.es/m/20200618180058.GK7349@momjian.us Backpatch-through: 9.5
https://git.postgresql.org/pg/commitdiff/2a06cb86dbacdf19a19d5bfd00d5156584f2188a

- doc: clarify that storage parameter values are optional. In a few cases, the
documented syntax specified storage parameter values as required.
Reported-by: galiev_mr(at)taximaxim(dot)ru Discussion:
https://postgr.es/m/159283163235.684.4482737698910467437@wrigleys.postgresql.org
Backpatch-through: 9.5
https://git.postgresql.org/pg/commitdiff/aa90d9957bd2d9ce39f9078d15914a37d99af072

Michael Meskes pushed:

- Fix ecpg crash with bytea and cursor variables. Author: Jehan-Guillaume de
Rorthais <jgdr(at)dalibo(dot)com>
https://git.postgresql.org/pg/commitdiff/e576f71fbe713d9357d1b480b4993043f7fcc73a

David Rowley pushed:

- Further adjustments to Hashagg EXPLAIN ANALYZE output. The "Disk Usage" and
"HashAgg Batches" properties in the EXPLAIN ANALYZE output for HashAgg were
previously only shown if the number of batches was greater than 0. Here we
change this so that these properties are always shown for EXPLAIN ANALYZE
formats other than "text". The idea here is that since the HashAgg could have
spilled to disk if there had been more data or groups to aggregate, then it's
relevant that we're clear in the EXPLAIN ANALYZE output when no spilling
occurred in this particular execution of the given plan. For the "text"
EXPLAIN format, we still hide these properties when no spilling occurs. This
EXPLAIN format is designed to be easy for humans to read. To maintain the
readability we have a higher threshold for which properties we display for
this format. Discussion:
https://postgr.es/m/CAApHDvo_dmNozQQTmN-2jGp1vT%3Ddxx7Q0vd%2BMvD1cGpv2HU%3DSg%40mail.gmail.com
Backpatch-through: 13, where the hashagg spilling code was added.
https://git.postgresql.org/pg/commitdiff/40efbf8706cdd96e06bc4d1754272e46d9857875

Amit Kapila pushed:

- Improve vacuum error context handling. Use separate functions to save and
restore error context information as that made code easier to understand.
 Also, make it clear that the index information required for error context is
sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through:
13, where it was introduced Discussion:
https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/a69e041d0c91759fb60ab52e7e21e3ec3752c69b

Magnus Hagander pushed:

- Fix temporary tablespaces for shared filesets. A likely copy/paste error in
98e8b480532 from back in 2004 would cause temp tablespace to be reset to
InvalidOid if temp_tablespaces was set to the same value as the primary
tablespace in the database. This would cause shared filesets (such as for
parallel hash joins) to ignore them, putting the temporary files in the
default tablespace instead of the configured one. The bug is in the old code,
but it appears to have been exposed only once we had shared filesets.
Reviewed-By: Daniel Gustafsson Discussion:
https://postgr.es/m/CABUevExg5YEsOvqMxrjoNvb3ApVyH+9jggWGKwTDFyFCVWczGQ@mail.gmail.com
Backpatch-through: 9.5
https://git.postgresql.org/pg/commitdiff/ecd9e9f0bc141550726b2205dd6f5745a58e9ecd

Joe Conway pushed:

- Read until EOF vice stat-reported size in read_binary_file.
read_binary_file(), used by SQL functions pg_read_file() and friends, uses
stat to determine file length to read, when not passed an explicit length as
an argument. This is problematic, for example, if the file being read is a
virtual file with a stat-reported length of zero. Arrange to read until EOF,
or StringInfo data string lenth limit, is reached instead. Original complaint
and patch by me, with significant review, corrections, advice, and code
optimizations by Tom Lane. Backpatched to v11. Prior to that only paths
relative to the data and log dirs were allowed for files, so no "zero length"
files were reachable anyway. Reviewed-By: Tom Lane Discussion:
https://postgr.es/m/flat/969b8d82-5bb2-5fa8-4eb1-f0e685c5d736%40joeconway.com
Backpatch-through: 11
https://git.postgresql.org/pg/commitdiff/96d1f423f95d15c74989336e0776e1d55561a3f1

- Fix "ignoring return value" complaints from commit 96d1f423f9. The cfbot and
some BF animals are complaining about the previous read_binary_file commit
because of ignoring return value of ‘fread’. So let's make everyone happy by
testing the return value even though not strictly needed. Reported by Justin
Pryzby, and suggested patch by Tom Lane. Backpatched to v11 same as the
previous commit. Reported-By: Justin Pryzby Reviewed-By: Tom Lane Discussion:
https://postgr.es/m/flat/969b8d82-5bb2-5fa8-4eb1-f0e685c5d736%40joeconway.com
Backpatch-through: 11
https://git.postgresql.org/pg/commitdiff/1d05627fcf54b26e0cbd7527f9858f165d117813

== Pending Patches ==

David Rowley sent in three more revisions of a patch to keep elog(ERROR) and
ereport(ERROR) calls in the cold path.

Andy Fan sent in another revision of a patch to track uniqueness with UniqueKey.

Ashutosh Bapat sent in a patch to Assert if ExecOpenIndices() is called twice on
same result relation.

Peter Eisentraut sent in another revision of a patch to allow CURRENT_ROLE where
CURRENT_USER is accepted.

Fabien COELHO sent in a patch to pgbench to fix some of the underlying problems
reported by delaying the :var to $1 substitution to the last possible moments,
so that what variables are actually defined is known.

Daisuke Higuchi and Kyotaro HORIGUCHI traded patches to ensure that the complete
segment is archived immediately on standby.

Peter Eisentraut sent in another revision of a patch to pg_dump which
reorganizes dumpFunc() and dumpAgg() to reduce duplication.

Felix Lechner and Daniel Gustafsson traded patches to fix file removal in SSL
tests.

Michaël Paquier sent in two more revisions of a patch to use heap_multi_insert()
for pg_attribute/depend insertions.

Bharath Rupireddy sent in a patch to speed up COPY FROM ... BINARY by reading in
RAW_BUF_SIZE bytes at a time and processing that, rather than hitting storage at
many points in between.

Atsushi Torikoshi sent in three more revisions of a patch to add a function
exposing memory usage of local backend.

Ants Aasma and Fujii Masao traded patches to fix an issue where track_planning
was causing performance regressions.

John Naylor sent in another revision of a patch to enable truncating timestamps
on a wider variety of intervals.

Kyotaro HORIGUCHI sent in a patch to make sure standby archives all segments,
and another to make sure standby archives all segments immediately.

Mark Dilger sent in another revision of a patch to de-support postfix operators.

Vik Fearing sent in a patch to implement <null treatment> for window functions.

Peter Eisentraut sent in a patch to add support for writing CREATE FUNCTION and
CREATE PROCEDURE statements for language SQL with a function body that conforms
to the SQL standard.

David Steele sent in another revision of a patch to remove the exclusive backup
mode, which has been deprecated since PostgreSQL 9.6.

Peter Eisentraut sent in two revisions of a patch to fix -Wcast-function-type
warnings.

Georgios Kokolatos sent in another revision of a patch to Include AM in
listTables output.

Michaël Paquier sent in another revision of a patch to fix the TAP tests on
Windows around the issue of symlinks.

Dilip Kumar and Amit Kapila traded patches to fix an infelicity between
logical_work_mem and logical streaming of large in-progress transactions.

Peter Geoghegan sent in a PoC patch to teach nbtree to delete old duplicate
versions from unique indexes.

Vigneshwaran C sent in another revision of a patch to implement parallel COPY.

Vigneshwaran C sent in a patch to make COPY FROM more efficient by arranging to
call CopyGetData less frequently, removing an unused function pointer in
CopyMultiInsertInfoNextFreeSlot, and avoids unneccessarily clearing EOL and
converting to server encoding for the header line, whose contents are never
actually examined.

Amit Langote sent in two revisions of a patch to reduce ModifyTable overheads in
generic plans.

Julien Rouhaud sent in two more revisions of a patch to implement collation
versioning.

Tsahi Zidenberg sent in a patch to support outline-atomics on aarch64.

Stephen Frost sent in a patch to give a somewhat clearer error message when we
detect that the user has asked us to follow a timeline switch to a new timeline
before we've reached consistency.

Bilva Sanaba sent in a patch to support COPY TO callback functions.

Mark Dilger sent in two more revisions of a patch to clean up inappropriate uses
of the name "relkind."

Bertrand Drouvot sent in another revision of a patch to add standby recovery
conflicts info.

Amul Sul sent in a patch to put ipc.c back inside the 80-column limit.

Kyotaro HORIGUCHI sent in a patch to replace min_safe_lsn with "distance" in
pg_replication_slots.

Kyotaro HORIGUCHI sent in another revision of a patch to allow wait event sets
to be registered to resource owners, add infrastructure for asynchronous
execution, and use same in the PostgreSQL FDW.

Kyotaro HORIGUCHI sent in another revision of a patch to remove the page-read
callback from XLogReaderState.

James Sewell sent in a patch to replace *sigprocmask with *pthread_sigmask*
calls in the BGWorker system to make it possible to spawn threads safely in
background workers.

David Rowley sent in another revision of a patch to allow estimate_num_groups()
to pass back further details about the estimation, allow users of simplehash.h
to perform direct deletions, and add a new Result Cache executor node.

Bharath Rupireddy sent in a patch to retry cached remote connections in the
PostgreSQL FDW in case the remote backend is killed.

Daniel Gustafsson sent in another revision of a patch to implement progress
tracking for COPY.

Daniel Gustafsson sent in another revision of a patch to use the internal
pthreads reimplementation only when building with MSVC.

Pavel Borisov sent in another revision of a patch to fix a bug that manifested
as GIN index search sometimes losing results.

Pavel Stěhule sent in another revision of a patch to implement a
unicode_unescape() function.

Peter Geoghegan sent in another revision of a patch to add Valgrind buffer
access instrumentation and add nbtree Valgrind buffer lock checks.

James Coleman sent in two revisions of a patch to use unint64 instead of long
for space used variables in incremental sort.

Michaël Paquier sent in another revision of a patch to add flag to
format_type_extended to enforce NULL-ness, refactor the format procedure and
operator APIs to be more modular, and eliminate user-visible cache lookup errors
for objaddr SQL functions.

Masahiko Sawada sent in a PoC patch to reset spilled txn statistics in
pg_stat_replication.

Konstantin Knizhnik sent in two more revisions of a patch to implement a
built-in connection pooler.

David Cramer sent in a patch to add binary support to the pgoutput plugin.

Tomáš Vondra sent in another revision of a patch to implement BRIN multi-range
indexes.

Tomáš Vondra sent in a patch to improve estimation of OR clauses using extended
statistics.

Peter Eisentraut sent in a patch to adjust cycle detection examples and tests.

Nikolay Shaplov sent in another revision of a patch to split StdRdOptions into
HeapOptions and ToastOptions.

Vigneshwaran C sent in a patch to fix situations where Parallel worker hangs
while handling errors.

Laurenz Albe sent in another revision of a patch to introduce a new RelInfoList
structure, add basic functionality for aggregate pushdown, and use partial paths
as additional input for grouped paths.

Daniel Gustafsson sent in another revision of a patch to support NSS as a libpq
TLS backend.

Fujii Masao sent in a patch to document the overhead caused by track planning.

Peter Eisentraut sent in a patch to avoid hiding the "Up" link when it is the
same as "Home".

Peter Geoghegan sent in a patch to add a GUC that limits memory use for hash
tables.

Vigneshwaran C sent in two revisions of a patch to remove an nused function
parameter in parallel vacuum & reorder buffer.

Justin Pryzby sent in another revision of a patch to implement some secondary
index access optimizations, and avoid index scan inconsistent with partition
constraint.

Fabien COELHO sent in another revision of a patch to measure pgbench time in
microseconds, and use pthread barriers.

Movead Li sent in another revision of a patch to add functions to get origin
information from commit_ts.

Pavel Stěhule sent in another revision of a patch to implement
string_to_table().

Pavel Stěhule sent in another revision of a patch to implement schema variables.

Anna Akenteva sent in two revisions of a patch to implement ALTER TABLE ...
ALTER CONSTRAINT ... USING INDEX ...

Paul A Jungwirth sent in another revision of a patch to implement multiranges.

Pavel Stěhule sent in another revision of a patch to add a --filter option to
pg_dump.