== PostgreSQL Weekly News - January 13, 2019 ==

From: David Fetter <david(at)fetter(dot)org>
To: PostgreSQL Announce <pgsql-announce(at)postgresql(dot)org>
Subject: == PostgreSQL Weekly News - January 13, 2019 ==
Date: 2019-01-13 21:45:45
Message-ID: 20190113214545.GA27163@fetter.org
Views: Raw Message | Whole Thread | Download mbox
Lists: pgsql-announce

== PostgreSQL Weekly News - January 13, 2019 ==

The Turkish PostgreSQL conference will take place January 24,
2019 in Ankara.

== PostgreSQL Product News ==

pgAdmin4 4.0, a web- and native GUI control center for PostgreSQL, released.

== PostgreSQL Jobs for January ==


== PostgreSQL Local ==

FOSDEM PGDay 2019, a one day conference held before the main FOSDEM event will
be held in Brussels, Belgium, on Feb 1st, 2019.

Prague PostgreSQL Developer Day 2019 (P2D2 2019) is a two-day
conference that will be held on February 13-14, 2019 in Prague, Czech Republic.

PGConf India 2019 will be on February 13-15, 2019 in Bengaluru, Karnataka.

pgDay Paris 2019 will be held in Paris, France on March 12, 2019
at 199bis rue Saint-Martin.

PGConf APAC 2019 will be held in Singapore March 19-21, 2019.

The German-speaking PostgreSQL Conference 2019 will take place on May 10, 2019
in Leipzig. The CfP is open until February 26, 2019 at http://2019.pgconf.de/cfp

PGDay.IT 2019 will take place May 16th and May 17th in Bologna, Italy. Both the
CfP https://2019.pgday.it/en/blog/cfp and the Call for Workshops
https://2019.pgday.it/en/blog/cfw are open until January 15, 2019.

PGCon 2019 will take place in Ottawa on May 28-31, 2019. The CfP is open
through January 19, 2019 at http://www.pgcon.org/2019/papers.php

Swiss PGDay 2019 will take place in Rapperswil (near Zurich) on June 28, 2019.
The CfP is open January 17, 2019 through April 18, 2019, and registration will
open January 17, 2019.

== 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 ==

Peter Eisentraut pushed:

- pg_basebackup: Use atexit(). Instead of using our custom
disconnect_and_exit(), just register the desired cleanup using atexit() and
use the standard exit() to leave the program. Reviewed-by: Alvaro Herrera
<alvherre(at)2ndquadrant(dot)com> Reviewed-by: Michael Paquier <michael(at)paquier(dot)xyz>

- initdb: Use atexit(). Replace exit_nicely() calls with standard exit() and
register the cleanup actions using atexit(). The coding pattern used here
mirrors existing use in pg_basebackup.c. Reviewed-by: Alvaro Herrera
<alvherre(at)2ndquadrant(dot)com> Reviewed-by: Michael Paquier <michael(at)paquier(dot)xyz>

- isolationtester: Use atexit(). Replace exit_nicely() calls with standard
exit() and register the cleanup actions using atexit(). Reviewed-by: Alvaro
Herrera <alvherre(at)2ndquadrant(dot)com> Discussion:

- Add value 'current' for recovery_target_timeline. This value represents the
default behavior of using the current timeline. Previously, this was
represented by an empty string. (Before the removal of recovery.conf, this
setting could not be chosen explicitly but was used when
recovery_target_timeline was not mentioned at all.) Discussion:
Reviewed-by: David Steele <david(at)pgmasters(dot)net> Reviewed-by: Michael Paquier

- Create INSTALL file using Pandoc. Replace using lynx with using pandoc.
Pandoc creates better looking output and it avoids the delicate
locale/encoding issues of lynx because it always uses UTF-8 for both input and
output. Note: requires Pandoc >=1.13 Discussion:
Reviewed-by: Mi Tar <mmitar(at)gmail(dot)com>

- doc: Correct documentation of install-time environment variables. Since
approximately PostgreSQL 10, it is no longer required that environment
variables at installation time such as PERL, PYTHON, TCLSH be "full path
names", so change that phrasing in the installation instructions. (The exact
time of change appears to differ for PERL and the others, but it works
consistently in PostgreSQL 10.) Also while we're here document the defaults
for PERL and PYTHON, but since the search list for TCLSH is so long, let's
leave that out so we don't need to maintain a copy of that list in the
installation instructions.

- Change default of recovery_target_timeline to 'latest'. This is what one
usually wants for recovery and almost always wants for a standby. Discussion:
Reviewed-by: David Steele <david(at)pgmasters(dot)net> Reviewed-by: Michael Paquier

- configure: Update python search order. Some systems don't ship with "python"
by default anymore, only "python3" or "python2" or some combination, so
include those in the configure search. Discussion:

- Make INSTALL makefile rule more robust. With the previous rule, if pandoc was
missing, a zero-length output file would be created without an error from
make. To improve that, write the rule as two separate commands without a
pipe. Reported-by: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>

Andrew Gierth pushed:

- doc: document that INFO messages always go to client. In passing add a couple
of links to the message severity table. Backpatch because it's always been
this way. Author: Karl O. Pinc <kop(at)meme(dot)com>

Tatsuo Ishii pushed:

- Doc: fix meaning of acronym "btree". Acronym "btree" better means "multi-way
balanced tree" rather than "multi-way binary tree". Discussion:

Michaël Paquier pushed:

- Add --disable-page-skipping and --skip-locked to vacuumdb.
DISABLE_PAGE_SKIPPING is available since v9.6, and SKIP_LOCKED since v12.
They lacked equivalents for vacuumdb, so this closes the gap. Author: Nathan
Bossart Reviewed-by: Michael Paquier, Masahiko Sawada Discussion:

- Fix grammar mistakes in md.c. Author: Kirk Jamison Discussion:

- Update unaccent rules with release 34 of CLDR for Latin-ASCII.xml. This has
required an update of the python script generating the rules, as its format
has changed in release 29. This release has also added new punctuation and
symbols, and a new set of rules has been generated to include them. The way
to find newest versions of Latin-ASCII gets also more clearly documented.
Author: Hugh Ranalli, Michael Paquier Discussion:

- Improve error messages for incorrect types of logical replication targets. If
trying to use something else than a plain table as logical replication target,
a rather-generic error message gets used to report the problem. This can be
confusing when it comes to foreign tables and partitioned tables, so use more
dedicated messages in these cases. Author: Amit Langote Reviewed-by: Peter
Eisentraut, Magnus Hagander, Michael Paquier Discussion:

- Fix error message for logical replication targets. This fixes an oversight
from 373bda6. Noted by Erik Rijkers.

Tom Lane pushed:

- Doc: update our docs about kernel IPC parameters on *BSD. runtime.sgml said
that you couldn't change SysV IPC parameters on OpenBSD except by rebuilding
the kernel. That's definitely wrong in OpenBSD 6.x, and excavation in their
man pages says it changed in OpenBSD 3.3. Update NetBSD and OpenBSD sections
to recommend adjustment of the SEMMNI and SEMMNS settings, which are painfully
small by default on those platforms. (The discussion thread contemplated
recommending that people select POSIX semaphores instead, but the performance
consequences of that aren't really clear, so I'll refrain.) Remove pointless
discussion of SEMMNU and SEMMAP from the FreeBSD section. Minor other
wordsmithing. Discussion: https://postgr.es/m/27582.1546928073@sss.pgh.pa.us

- Update docs & tests to reflect that unassigned OLD/NEW are now NULL. For a
long time, plpgsql has allowed trigger functions to parse references to OLD
and NEW even if the current trigger event type didn't assign a value to one or
the other variable; but actually executing such a reference would fail. The
v11 changes to use "expanded records" for DTYPE_REC variables changed the
behavior so that the unassigned variable now reads as a null composite value.
While this behavioral change was more or less unintentional, it seems that
leaving it like this is better than adding code and complexity to be
bug-compatible with the old way. The change doesn't break any code that worked
before, and it eliminates a gotcha that often required extra code to work
around. Hence, update the docs to say that these variables are "null" not
"unassigned" when not relevant to the event type. And add a regression test
covering the behavior, so that we'll notice if we ever break it again. Per
report from Kristjan Tammekivi. Discussion:

- Reduce the size of the fmgr_builtin_oid_index[] array. This index array was
originally defined to have 10000 entries (ranging up to FirstGenbkiObjectId),
but we really only need entries up to the last existing builtin function OID,
currently 6121. That saves close to 8K of never-accessed space in the server
executable, at the small price of one more fetch in fmgr_isbuiltin(). We
could reduce the array size still further by renumbering a few of the
highest-numbered builtin functions; but there's a small risk of breaking
clients that have chosen to hardwire those function OIDs, so it's not clear if
it'd be worth the trouble. (We should, however, discourage future patches
from choosing function OIDs above 6K as long as there's still lots of space
below that.) Discussion: https://postgr.es/m/12359.1547063064@sss.pgh.pa.us

- Use perfect hashing, instead of binary search, for keyword lookup. We've been
speculating for a long time that hash-based keyword lookup ought to be faster
than binary search, but up to now we hadn't found a suitable tool for
generating the hash function. Joerg Sonnenberger provided the inspiration,
and sample code, to show us that rolling our own generator wasn't a ridiculous
idea. Hence, do that. The method used here requires a lookup table of
approximately 4 bytes per keyword, but that's less than what we saved in the
predecessor commit afb0d0712, so it's not a big problem. The time savings is
indeed significant: preliminary testing suggests that the total time for raw
parsing (flex + bison phases) drops by ~20%. Patch by me, but it owes its
existence to Joerg Sonnenberger; thanks also to John Naylor for review.
Discussion: https://postgr.es/m/20190103163340.GA15803@britannica.bec.de

- Remove unnecessary #include. Discussion:

- Fix C++ compile failures in headers. Avoid using "typeid" as a parameter name
in header files, since that is a C++ keyword. These cases were introduced
recently, in 04fe805a1 and 586b98fdf. Since I'm an incurable neatnik, also
rename these parameters in the underlying function definitions. That's not
really necessary per project rules, but I don't like function declarations
that don't quite agree with the underlying definitions. Per

- Add .gitignore entry for a derived file created by "make distprep". I chanced
to notice that "make distprep" leaves a state of the tree that git complains
about. It's been like this for awhile, but given the lack of complaints it
probably doesn't need back-patching.

- Avoid sharing PARAM_EXEC slots between different levels of NestLoop. Up to
now, createplan.c attempted to share PARAM_EXEC slots for NestLoopParams
across different plan levels, if the same underlying Var was being fed down to
different righthand-side subplan trees by different NestLoops. This was, I
think, more of an artifact of using subselect.c's PlannerParamItem
infrastructure than an explicit design goal, but anyway that was the end
result. This works well enough as long as the plan tree is executing
synchronously, but the feature whereby Gather can execute the parallelized
subplan locally breaks it. An upper NestLoop node might execute for a row
retrieved from a parallel worker, and assign a value for a PARAM_EXEC slot
from that row, while the leader's copy of the parallelized subplan is
suspended with a different active value of the row the Var comes from. When
control eventually returns to the leader's subplan, it gets the wrong answers
if the same PARAM_EXEC slot is being used within the subplan, as reported in
bug #15577 from Bartosz Polnik. This is pretty reminiscent of the problem
fixed in commit 46c508fbc, and the proper fix seems to be the same: don't try
to share PARAM_EXEC slots across different levels of controlling NestLoop
nodes. This requires decoupling NestLoopParam handling from PlannerParamItem
handling, although the logic remains somewhat similar. To avoid bizarre
division of labor between subselect.c and createplan.c, I decided to move all
the param-slot-assignment logic for both cases out of those files and put it
into a new file paramassign.c. Hopefully it's a bit better documented now,
too. A regression test case for this might be nice, but we don't know a test
case that triggers the problem with a suitably small amount of data.
Back-patch to 9.6 where we added Gather nodes. It's conceivable that related
problems exist in older branches; but without some evidence for that, I'll
leave the older branches alone. Discussion:

- Improve missing-program error handling in make_ctags and make_etags. If ctags
(resp. etags) isn't installed, these scripts naturally fail, but the error
messages were less clear than one could wish. It seems worth installing an
explicit test to improve that. Nikolay Shaplov, with suggestions from Michael
Paquier and Andrew Dunstan Discussion:

Álvaro Herrera pushed:

- pgbench: add \cset and \gset commands. These commands allow assignment of
values produced by queries to pgbench variables, where they can be used by
further commands. \gset terminates a command sequence (just like a bare
semicolon); \cset separates multiple queries in a compound command, like an
escaped semicolon (\;). A prefix can be provided to the \-command and is
prepended to the name of each output column to produce the final variable
name. This feature allows pgbench scripts to react meaningfully to the actual
database contents, allowing more powerful benchmarks to be written. Authors:
Fabien Coelho, Álvaro Herrera Reviewed-by: Amit Langote
<Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp> Reviewed-by: Stephen Frost
<sfrost(at)snowman(dot)net> Reviewed-by: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
Reviewed-by: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> Reviewed-by: Tatsuo Ishii
<ishii(at)sraoss(dot)co(dot)jp> Reviewed-by: Rafia Sabih <rafia(dot)sabih(at)enterprisedb(dot)com>
Discussion: https://postgr.es/m/alpine.DEB.2.20.1607091005330.3412@sto

- Don't use address of array as boolean. Per buildfarm

- Move inheritance expansion code into its own file. This commit moves
expand_inherited_tables and underlings from optimizer/prep/prepunionc.c to
optimizer/utils/inherit.c. Also, all of the AppendRelInfo-based expression
manipulation routines are moved to optimizer/utils/appendinfo.c. No
functional code changes. One exception is the introduction of
make_append_rel_info, but that's still just moving around code. Also, stop
including <limits.h> in prepunion.c, which no longer needs it since
3fc6e2d7f5b6. I (Álvaro) noticed this because Amit was copying that to
inherit.c, which likewise doesn't need it. Author: Amit Langote Discussion:

Andrew Dunstan pushed:

- Fix missing values when doing ALTER TABLE ALTER COLUMN TYPE. This was an
oversight in commit 16828d5c. If the table is going to be rewritten, we simply
clear all the missing values from all the table's attributes, since there will
no longer be any rows with the attributes missing. Otherwise, we repackage the
missing value in an array constructed with the new type specifications.
Backpatch to release 11. This fixes bug #15446, reported by Dmitry Molotkov
Reviewed by Dean Rasheed

- Free pre-modification HeapTuple in ALTER TABLE ... TYPE ... This was an
oversight in commit 3b174b1a3. Per offline gripe from Alvaro Herrera
Backpatch to release 11.

- Make DLSUFFIX easily discoverable by build scripts. This will enable things
like the buildfarm client to discover more reliably if certain libraries have
been installed. Discussion:
Backpatch to all live branches.

Amit Kapila pushed:

- Extend pg_stat_statements_reset to reset statistics specific to a.particular
user/db/query. The function pg_stat_statements_reset() is extended to accept
userid, dbid, and queryid as input parameters. Now, it can discard the
statistics gathered so far by pg_stat_statements corresponding to the
specified userid, dbid, and queryid. If no parameter is specified or all the
specified parameters have default value aka 0, it will discard all statistics
as per the old behavior. The new behavior is useful to get the fresh
statistics for a specific user/database/query without resetting all the
existing statistics. Author: Haribabu Kommi, with few additional changes by
me Reviewed-by: Michael Paquier, Amit Kapila and Fujii Masao Discussion:

Andres Freund pushed:

- Remove heapam.h include made superfluous by b60c3975990. Noticed this while
working on another patch. Author: Andres Freund

Noah Misch pushed:

- Make Emacs perl-mode indent more like perltidy. This especially helps braces
that surround code blocks. Back-patch to v11, where commit
56fb890ace8ac0ca955ae0803c580c2074f876f6 first appeared; before that, settings
were even more distant from perltidy. Reviewed by Andrew Dunstan.
Discussion: https://postgr.es/m/20190103055355.GB267595@gust.leadboat.com

== Pending Patches ==

Etsuro Fujita sent in another revision of a patch to refactor the PostgreSQL
FDW's DML execution hook functions.

Amit Khandekar sent in a patch to restore the default_with_oids GUC and add a
check-hook that throws an error when it's set to true.

Jesper Pedersen sent in a patch to highlight the fact that the -j option isn't
passed down to vacuumdb by default.

Nikolay Shaplov sent in another revision of a patch to replace StdRdOptions with
individual binary reloptions representation for each relation kind.

Christoph Berg sent in two revisions of a patch to add more flexibility for
passing custom compiler flags, ensuring that C++ "just works."

Michaël Paquier sent in another revision of a patch to add --skip-locked and
friends to vacuumdb.

John Naylor sent in two more revisions of a patch to avoid creation of the free
space map for small heap relations and conditionally skip transfer of FSMs
during pg_upgrade.

Amit Langote and Michaël Paquier traded patches to error out when foreign or
partitioned tables are used as targets of logical replication.

Michael Banck and Fabien COELHO traded patches to allow enabling and disabling
page checksums offline.

Amit Kapila and Dilip Kumar traded patches to provide interfaces to store and
fetch undo records.

Haribabu Kommi and Amit Kapila traded patches to extend pg_stat_statements_reset
to reset statistics specific to a particular user/db/query.

Amit Langote sent in five more revisions of a patch to speed up planning with

Peter Geoghegan sent in another revision of a patch to make all nbtree entries
unique by having heap TIDs participate in comparisons.

Daisuke Higuchi sent in two revisions of a patch to fix some issues during
Windows start.

Artur Zakirov sent in another revision of a patch to make it possible to use
shared ispell dictionaries.

Konstantin Knizhnik sent in another revision of a patch to implement compression
for libpq.

Surafel Temesgen and Tomáš Vondra traded patches to add PERCENT as an option to

Kyotaro HORIGUCHI sent in another revision of a patch to improve selectivity
estimates for range queries.

Etsuro Fujita and Amit Langote traded patches to add child EC members for only
the non-dummy children.

Masahiko Sawada sent in two more revisions of a patch to add a

Amit Langote sent in a patch to ensure PK-side action triggers for partitions
after being detached and stop tracking foreign key inheritance by dependencies.

Daniel Gustafsson sent in a patch to fix some odd wording in a comment in the
advanced tutorial.

Michaël Paquier sent in another revision of a patch to make WAL receiver startup
rely on GUC context for primary_conninfo and primary_slot_name.

Andreas Karlsson sent in another revision of a patch to inline CTEs.

Chapman Flack, Artur Zakirov, and Pavel Stěhule traded patches to fix a problem

Dilip Kumar sent in another revision of a patch to implement an UNDO worker and
transaction rollback using same.

Nikita Glukhov sent in another revision of a patch to implement kNN searches for
B-tree indexes.

KaiGai Kohei sent in another revision of a patch to fix a bug in the way
extensions can munge paths in the planner.

Nikita Glukhov sent in another revision of a patch to implement JSONPATH.

Peter Eisentraut sent in another revision of a patch to implement a unified
logging system for command-line programs.

Andrew Gierth sent in another revision of a patch to use the Ryu routines to
make floating point output faster and more precise.

Peter Eisentraut sent in another revision of a patch to implement generated
columns, both virtual and stored.

Andres Freund sent in a patch to remove all "INTERFACE ROUTINE" style comments.

David Rowley sent in another revision of a patch to allow lock acquisitions for
partitions to be delayed.

Amit Kapila sent in another revision of a patch to implement logical decoding
for operations on zheap tables.

Dmitry Dolgov sent in two more revisions of a patch to add appropriate
access_method dumps to pg_dump.

Alexey Bashtanov sent in two more revisions of a patch to log bind parameter
values on error.

Michaël Paquier sent in a PoC patch to use O_DIRECT for relations and SLRUs.

David Rowley sent in another revision of a patch to forgo generating
single-subpath Append and MergeAppend paths.

Tomáš Vondra sent in another revision of a patch to implement multivariate
histograms and MCV lists.

Paul Martinez sent in a patch to include all columns in the default names for
foreign key constraints.

Tom Lane sent in a patch to make heap.c's RemoveStatistics send out an sinval
message commanding deletion of negative STATRELATTINH entries that match the OID
of the table being deleted.

James Coleman sent in another revision of a patch to prove IS NOT NULL inference
for ScalarArrayOpExpr.

Browse pgsql-announce by date

  From Date Subject
Next Message Dan Langille 2019-01-13 22:32:03 PGCon 2019 CFP closes on 19 January
Previous Message Akshay Joshi 2019-01-10 10:52:28 pgAdmin 4 v4.0 released