== PostgreSQL Weekly News - October 01 2017 ==
PostgreSQL@SCaLE is a two day, two track event which takes place on
March 8-9, 2018, at Pasadena Convention Center, as part of SCaLE 16X.
The CfP is open through October 31, 2017 at
http://www.socallinuxexpo.org/scale/16x/cfp
pgday.Seoul 2017 will be held in Seoul, South Korea on November 4, 2017.
Korean language information is here:
http://pgday.postgresql.kr/
== PostgreSQL Product News ==
PostGIS 2.4.0, the industry standard geographic information
system package for PostgreSQL, released.
http://postgis.net/2017/09/30/postgis-2.4.0/
Pagila 10.a, a sample database for PostgreSQL, released.
https://github.com/xzilla/pagila
dbForge Data Compare for PostgreSQL v3.0 released.
https://docs.google.com/document/d/1ggVLlKpU7NkozHrj0fUnhSQKpBOqNW330foZPWz4iy4/edit?usp=sharing
== PostgreSQL Jobs for October ==
http://archives.postgresql.org/pgsql-jobs/2017-10/
== PostgreSQL Local ==
PGDay.IT 2017 will take place October 13th, in Milan, Italy.
http://pgday.it
PostgreSQL Conference Europe 2017 will be held on October 24-27, 2017 in the
Warsaw Marriott Hotel, in Warsaw, Poland.
https://2017.pgconf.eu/
PGConf Local: Seattle will be held November 13 - 14, 2017.
https://www.pgconf.us/#Seattle2017
PGDay Australia 2017 will be held on November 17 in Melbourne.
http://2017.pgday.com.au/
2ndQuadrant PostgreSQL Conference 2017 (2Q PGConf, for short) will be hosted on
November 6th & 7th in New York City, and November 9th in Chicago.
http://www.2qpgconf.com/
PostgreSQL Session November 17th, 2017, in Paris, France. The CfP is open until
September 30, 2017 at call-for-paper AT postgresql-sessions DOT org. Details at
http://blog.dalibo.com/2017/08/29/cfp_pgsession9.html
PGConf Local: Austin will be held December 4 - 5, 2017. Call for Papers is
now open at https://www.pgconf.us/conferences/Austin2017
PGConf.ASIA 2017 will take place on December 4-6 2017 in Akihabara, Tokyo,
Japan.
http://www.pgconf.asia/EN/2017/
PGConf India 2018 will be on February 22-23, 2018 in Bengaluru, Karnataka.
Proposals are due via https://goo.gl/forms/F9hRjOIsaNasVOAz2 by October 31st, 2017.
http://pgconf.in/
== 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 EST5EDT. Please send English
language ones to david@fetter.org, German language to pwn@pgug.de, Italian
language to pwn@itpug.org.
== Applied Patches ==
Andrew Dunstan pushed:
- Support building with Visual Studio 2017. Haribabu Kommi, reviewed by Takeshi Ideriha and Christian Ullrich
Backpatchto 9.6 https://git.postgresql.org/pg/commitdiff/f2ab3898f3a25ef431db4ea90a8d128b974dbffe
Tom Lane pushed:
- Make construct_[md_]array return a valid empty array for zero-size input. If construct_array() or
construct_md_array()were given a dimension of zero, they'd produce an array that contains no elements but has positive
dimension.This violates a general expectation that empty arrays should have ndims = 0; in particular, while arrays like
thisprint as empty, they don't compare equal to other empty arrays. Up to now we've expected callers to avoid making
suchcalls and instead be careful to call construct_empty_array() if there would be no elements. But this has always
beenan easily missed case, and we've repeatedly had to fix callers to do it right. In bug #14826, Erwin Brandstetter
pointedout yet another such oversight, in ts_lexize(); and a bit of examination of other call sites found at least two
morewith similar issues. So let's fix the problem centrally and permanently by changing these two functions to
constructa proper zero-D empty array whenever the array would be empty. This renders a few explicit calls of
construct_empty_array()redundant, but the only such place I found that really seemed worth changing was in
ExecEvalArrayExpr(). Although this fixes some very old bugs, no back-patch: the problem is pretty minor and the risk of
changingbehavior seems to outweigh the benefit in stable branches. Discussion:
https://postgr.es/m/20170923125723.1448.39412@wrigleys.postgresql.orgDiscussion:
https://postgr.es/m/20570.1506198383@sss.pgh.pa.us
https://git.postgresql.org/pg/commitdiff/716ea626a88ac510523ab3af5bc779d78eeced58
- Avoid SIGBUS on Linux when a DSM memory request overruns tmpfs. On Linux, shared memory segments created with
shm_open()are backed by swap files created in tmpfs. If the swap file needs to be extended, but there's no tmpfs space
left,you get a very unfriendly SIGBUS trap. To avoid this, force allocation of the full request size when we create
thesegment. This adds a few cycles, but none that we wouldn't expend later anyway, assuming the request isn't hugely
biggerthan the actual need. Make this code #ifdef __linux__, because (a) there's not currently a reason to think the
sameproblem exists on other platforms, and (b) applying posix_fallocate() to an FD created by shm_open() isn't very
portableanyway. Back-patch to 9.4 where the DSM code came in. Thomas Munro, per a bug report from Amul Sul
Discussion:https://postgr.es/m/1002664500.12301802.1471008223422.JavaMail.yahoo@mail.yahoo.com
https://git.postgresql.org/pg/commitdiff/899bd785c0edf376077d3f5d65c316f92c1b64b5
- Use a blacklist to distinguish original from add-on enum values. Commit 15bc038f9 allowed ALTER TYPE ADD VALUE to be
executedinside transaction blocks, by disallowing the use of the added value later in the same transaction, except
underlimited circumstances. However, the test for "limited circumstances" was heuristic and could reject references to
enumvalues that were created during CREATE TYPE AS ENUM, not just later. This breaks the use-case of restoring pg_dump
scriptsin a single transaction, as reported in bug #14825 from Balazs Szilfai. We can improve this by keeping a
"blacklist"table of enum value OIDs created by ALTER TYPE ADD VALUE during the current transaction. Any
visible-but-uncommittedvalue whose OID is not in the blacklist must have been created by CREATE TYPE AS ENUM, and can
beused safely because it could not have a lifespan shorter than its parent enum type. This change also removes the
restrictionthat a renamed enum value can't be used before being committed (unless it was on the blacklist). Andrew
Dunstan,with cosmetic improvements by me. Back-patch to v10. Discussion:
https://postgr.es/m/20170922185904.1448.16585@wrigleys.postgresql.org
https://git.postgresql.org/pg/commitdiff/1635e80d30b16df98aebead12f2b82f17efd9bc8
- Remove heuristic same-transaction test from check_safe_enum_use(). The blacklist mechanism added by the preceding
commitdirectly fixes most of the practical cases that the same-transaction test was meant to cover. What remains is
use-caseslike begin; create type e as enum('x'); alter type e add value 'y'; -- use 'y' somehow commit; However,
becausethe same-transaction test is heuristic, it fails on small variants of that, such as renaming the type or
changingits owner. Rather than try to explain the behavior to users, let's remove it and just have a rule that the
newlyadded value can't be used before being committed, full stop. Perhaps later it will be worth the implementation
effortand overhead to have a more accurate test for type-was-created-in-this-transaction. We'll wait for some field
experiencewith v10 before deciding to do that. Back-patch to v10. Discussion:
https://postgr.es/m/20170922185904.1448.16585@wrigleys.postgresql.org
https://git.postgresql.org/pg/commitdiff/984c92074d84a81dc17e9865fc79e264eb50ad61
- Fix failure-to-read-man-page in commit 899bd785c. posix_fallocate() is not quite a drop-in replacement for
fallocate(),because it is defined to return the error code as its function result, not in "errno". I (tgl) missed this
becauseRHEL6's version seems to set errno as well. That is not the case on more modern Linuxen, though, as per
buildfarmresults. Aside from fixing the return-convention confusion, remove the test for ENOSYS; we expect that glibc
willmask that for posix_fallocate, though it does not for fallocate. Keep the test for EINTR, because POSIX specifies
thatas a possible result, and buildfarm results suggest that it can happen in practice. Back-patch to 9.4, like the
previouscommit. Thomas Munro Discussion:
https://postgr.es/m/1002664500.12301802.1471008223422.JavaMail.yahoo@mail.yahoo.com
https://git.postgresql.org/pg/commitdiff/5ea96efaa0100e96b6b927eb1e67869143e1db4e
- Improve wording of error message added in commit 714805010. Per suggestions from Peter Eisentraut and David
Johnston. Back-patch, like the previous commit. Discussion:
https://postgr.es/m/E1dv9jI-0006oT-Fn@gemulon.postgresql.org
https://git.postgresql.org/pg/commitdiff/9a50a93c7b1427f6182ed1f21ed76da5b1d6a57c
- Revert to 9.6 treatment of ALTER TYPE enumtype ADD VALUE. This reverts commit 15bc038f9, along with the followon
commits1635e80d3 and 984c92074 that tried to clean up the problems exposed by bug #14825. The result was incomplete
becauseit failed to address parallel-query requirements. With 10.0 release so close upon us, now does not seem like
thetime to be adding more code to fix that. I hope we can un-revert this code and add the missing parallel query
supportduring the v11 cycle. Back-patch to v10. Discussion:
https://postgr.es/m/20170922185904.1448.16585@wrigleys.postgresql.org
https://git.postgresql.org/pg/commitdiff/28e07270768518524291d7d7906668eb67f6b8a5
- Fix behavior when converting a float infinity to numeric. float8_numeric() and float4_numeric() failed to consider
thepossibility that the input is an IEEE infinity. The results depended on the platform-specific behavior of
sprintf():on most platforms you'd get something like ERROR: invalid input syntax for type numeric: "inf" but at least
onWindows it's possible for the conversion to succeed and deliver a finite value (typically 1), due to a nonstandard
outputformat from sprintf and lack of syntax error checking in these functions. Since our numeric type lacks the
conceptof infinity, a suitable conversion is impossible; the best thing to do is throw an explicit error before letting
sprintfdo its thing. While at it, let's use snprintf not sprintf. Overrunning the buffer should be impossible if
sprintfdoes what it's supposed to, but this is cheap insurance against a stack smash if it doesn't. Problem reported
byTaiki Kondo. Patch by me based on fix suggestion from KaiGai Kohei. Back-patch to all supported branches.
Discussion:https://postgr.es/m/12A9442FBAE80D4E8953883E0B84E088C8C7A2@BPXM01GP.gisp.nec.co.jp
https://git.postgresql.org/pg/commitdiff/7769fc000aa3b959d3e1c7d7c3c2555aba7722c3
- Marginal improvement for generated code in execExprInterp.c. Avoid the coding pattern "*op->resvalue = f();", as
somecompilers think that requires them to evaluate "op->resvalue" before the function call. Unless there are lots of
freeregisters, this can lead to a useless register spill and reload across the call. I changed all the cases like this
inExecInterpExpr(), but didn't bother in the out-of-line opcode eval subroutines, since those are presumably not as
performance-critical. Discussion: https://postgr.es/m/2508.1506630094@sss.pgh.pa.us
https://git.postgresql.org/pg/commitdiff/136ab7c5a5f54fecea7c28c8550c19123245acf0
- Fix inadequate locking during get_rel_oids(). get_rel_oids used to not take any relation locks at all, but that
stoppedbeing a good idea with commit 3c3bb9933, which inserted a syscache lookup into the function. A concurrent DROP
TABLEcould now produce "cache lookup failed", which we don't want to have happen in normal operation. The best
solutionseems to be to transiently take a lock on the relation named by the RangeVar (which also makes the result of
RangeVarGetRelida lot less spongy). But we shouldn't hold the lock beyond this function, because we don't want VACUUM
tolock more than one table at a time. (That would not be a big problem right now, but it will become one after the
pendingfeature patch to allow multiple tables to be named in VACUUM.) In passing, adjust vacuum_rel and analyze_rel to
documentthat we don't trust the passed RangeVar to be accurate, and allow the RangeVar to possibly be NULL --- which it
isanyway for a whole-database VACUUM, though we accidentally didn't crash for that case. The passed RangeVar is in
factinaccurate when dealing with a child partition, as of v10, and it has been wrong for a whole long time in the case
ofvacuum_rel() recursing to a TOAST table. None of these things present visible bugs up to now, because the passed
RangeVaris in fact only consulted for autovacuum logging, and in that particular context it's always accurate because
autovacuumdoesn't let vacuum.c expand partitions nor recurse to toast tables. Still, this seems like trouble waiting
tohappen, so let's nail the door at least partly shut. (Further cleanup is planned, in HEAD only, as part of the
pendingfeature patch.) Fix some sadly inaccurate/obsolete comments too. Back-patch to v10. Michael Paquier and Tom
LaneDiscussion: https://postgr.es/m/25023.1506107590@sss.pgh.pa.us
https://git.postgresql.org/pg/commitdiff/19de0ab23ccba12567c18640f00b49f01471018d
- Support arrays over domains. Allowing arrays with a domain type as their element type was left un-done in the
originaldomain patch, but not for any very good reason. This omission leads to such surprising results as array_agg()
notworking on a domain column, because the parser can't identify a suitable output type for the polymorphic aggregate.
Inorder to fix this, first clean up the APIs of coerce_to_domain() and some internal functions in parse_coerce.c so
thatwe consistently pass around a CoercionContext along with CoercionForm. Previously, we sometimes passed an
"isExplicit"boolean flag instead, which is strictly less information; and coerce_to_domain() didn't even get that, but
insteadhad to reverse-engineer isExplicit from CoercionForm. That's contrary to the documentation in primnodes.h that
saysthat CoercionForm only affects display and not semantics. I don't think this change fixes any live bugs, but it
makesthings more consistent. The main reason for doing it though is that now build_coercion_expression() receives
ccontext,which it needs in order to be able to recursively invoke coerce_to_target_type(). Next, reimplement
ArrayCoerceExprso that the node does not directly know any details of what has to be done to the individual array
elementswhile performing the array coercion. Instead, the per-element processing is represented by a sub-expression
whoseinput is a source array element and whose output is a target array element. This simplifies life in
parse_coerce.c,because it can build that sub-expression by a recursive invocation of coerce_to_target_type(). The
executornow handles the per-element processing as a compiled expression instead of hard-wired code. The main advantage
ofthis is that we can use a single ArrayCoerceExpr to handle as many as three successive steps per element: base type
conversion,typmod coercion, and domain constraint checking. The old code used two stacked ArrayCoerceExprs to handle
type+ typmod coercion, which was pretty inefficient, and adding yet another array deconstruction to do domain
constraintchecking seemed very unappetizing. In the case where we just need a single, very simple coercion function,
doingthis straightforwardly leads to a noticeable increase in the per-array-element runtime cost. Hence, add an
additionalshortcut evalfunc in execExprInterp.c that skips unnecessary overhead for that specific form of expression.
Theruntime speed of simple cases is within 1% or so of where it was before, while cases that previously required two
levelsof array processing are significantly faster. Finally, create an implicit array type for every domain type, as
wedo for base types, enums, etc. Everything except the array-coercion case seems to just work without further effort.
TomLane, reviewed by Andrew Dunstan Discussion: https://postgr.es/m/9852.1499791473@sss.pgh.pa.us
https://git.postgresql.org/pg/commitdiff/c12d570fa147d0ec273df53de3a2802925d551ba
- Fix pg_dump to assign domain array type OIDs during pg_upgrade. During a binary upgrade, all type OIDs are supposed
tobe assigned by pg_dump based on their values in the old cluster. But now that domains have arrays, there's nothing
tobase the arrays' type OIDs on, if we're upgrading from a pre-v11 cluster. Make pg_dump search for an unused type OID
touse for this purpose. Per buildfarm. Discussion: https://postgr.es/m/E1dyLlE-0002gT-H5@gemulon.postgresql.org
https://git.postgresql.org/pg/commitdiff/2632bcce5e767a2b5901bbef54ae52df061eee72
- Use a longer connection timeout in pg_isready test. Buildfarm members skink and sungazer have both recently failed
thistest, with symptoms indicating that the default 3-second timeout isn't quite enough for those very slow systems.
There'sno reason to be miserly with this timeout, so boost it to 60 seconds. Back-patch to all versions containing
thistest. That may be overkill, because the failure has only been observed in the v10 branch, but I don't feel like
havingto revisit this later. https://git.postgresql.org/pg/commitdiff/4a1c0f3dde765c65e0ccb712c899df16986d09ad
- Update v10 release notes, and set the official release date. Last(?) round of changes for 10.0.
https://git.postgresql.org/pg/commitdiff/5a632a213d43c30940de3328286172c52730a01d
Robert Haas pushed:
- Fix trivial mistake in README. You might think I (Robert) could manage to count to five without messing it up, but
ifyou did, you would be wrong. Amit Kapila Discussion:
http://postgr.es/m/CAA4eK1JxqqcuC5Un7YLQVhOYSZBS+t=3xqZuEkt5RyquyuxpwQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/79a4a665c046af91d4216fe69b535c429039d0d0
- Remove lsn from HashScanPosData. This was intended as infrastructure for weakening VACUUM's locking requirements,
similarto what was done for btree indexes in commit 2ed5b87f96d473962ec5230fd820abfeaccb2069. However, for hash
indexes,it seems that the improvements which are possible are actually extremely marginal. Furthermore, performing the
LSNcross-check will end up skipping cleanup far more often than is necessary; we only care about page modifications due
toa VACUUM, but the LSN check will fail if ANY modification has occurred. So, rather than pressing forward with that
"optimization",just rip the LSN field out. Patch by me, reviewed by Ashutosh Sharma and Amit Kapila Discussion:
http://postgr.es/m/CAA4eK1JxqqcuC5Un7YLQVhOYSZBS+t=3xqZuEkt5RyquyuxpwQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/22c5e73562c53437979efec4c26cd9fff408777c
- Remove replacement selection sort. At the time replacement_sort_tuples was introduced, there were still cases where
replacementselection sort noticeably outperformed using quicksort even for the first run. However, those cases seem to
haveevaporated as a result of further improvements made since that time (and perhaps also advances in CPU technology).
Soremove replacement selection and the controlling GUC entirely. This makes tuplesort.c noticeably simpler and
probablypaves the way for further optimizations someone might want to do later. Peter Geoghegan, with review and
testingby Tomas Vondra and me. Discussion:
https://postgr.es/m/CAH2-WzmmNjG_K0R9nqYwMq3zjyJJK+hCbiZYNGhAy-Zyjs64GQ@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/8b304b8b72b0a60f1968d39f01cf817c8df863ec
- pgbench: If we fail to send a command to the server, fail. This beats the old behavior of busy-waiting hands down.
Oversightin commit 12788ae49e1933f463bc59a6efe46c4a01701b76. Report by Pavan Deolasee. Patch by Fabien Coelho.
Reviewedby Pavan Deolasee. Discussion:
http://postgr.es/m/CABOikdPhfXTypckMC1Ux6Ko+hKBWwUBA=EXsvamXYSg8M9J94w@mail.gmail.com
https://git.postgresql.org/pg/commitdiff/e55d9643ecb87f41185941b54d632641b3852aaa
- psql: Don't try to print a partition constraint we didn't fetch. If \d rather than \d+ is used, then verbose is
falseand we don't ask the server for the partition constraint; so we shouldn't print it in that case either. Maksim
Milyutin,per a report from Jesper Pedersen. Reviewed by Jesper Pedersen and Amit Langote. Discussion:
http://postgr.es/m/2af5fc4d-7bcc-daa8-4fe6-86274bea363c@redhat.com
https://git.postgresql.org/pg/commitdiff/69c16983e103f913ee0dae7f288611de006ba2ba
Peter Eisentraut pushed:
- Handle heap rewrites better in logical replication. A FOR ALL TABLES publication naturally considers all base tables
tobe a candidate for replication. This includes transient heaps that are created during a table rewrite during DDL.
Thiscauses failures on the subscriber side because it will not have a table like pg_temp_16386 to receive data (and if
itdid, it would be the wrong table). The prevent this problem, we filter out any tables that match this naming pattern
andmatch an actual table from FOR ALL TABLES publications. This is only a heuristic, meaning that user tables that
matchthat naming could accidentally be omitted. A more robust solution might require an explicit marking of such
tablesin pg_class somehow. Reported-by: yxq <yxq@o2.pl> Bug: #14785 Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by:Petr Jelinek <petr.jelinek@2ndquadrant.com>
https://git.postgresql.org/pg/commitdiff/ab28feae2bd3d4629bd73ae3548e671c57d785f0
- Sort pg_basebackup options better. The --slot option somehow ended up under options controlling the output, and some
otheroptions were in a nonsensical place or were not moved after recent renamings, so tidy all that up a bit.
https://git.postgresql.org/pg/commitdiff/15a8010ed691f190aad19c0a205f4a17868591e9
- Turn on log_replication_commands in PostgresNode. This is useful for example for the pg_basebackup and related
tests.https://git.postgresql.org/pg/commitdiff/43588f58aa045c736af168267e0f1c5934333e15
- Add some more pg_receivewal tests. Add some more tests for the --create-slot and --drop-slot options, verifying that
theright kind of slot was created and that the slot was dropped. While working on an unrelated patch for
pg_basebackup,some of this was temporarily broken without any tests noticing.
https://git.postgresql.org/pg/commitdiff/fa41461205ae4eb417045825583c3209e5a4f339
- pg_basebackup: Add option to create replication slot. When requesting a particular replication slot, the new
pg_basebackupoption -C/--create-slot creates it before starting to replicate from it. Further refactor the slot
creationlogic to include the temporary slot creation logic into the same function. Add new arguments is_temporary and
preserve_walto CreateReplicationSlot(). Print in --verbose mode that a slot has been created. Author: Michael Banck
<michael.banck@credativ.de>https://git.postgresql.org/pg/commitdiff/3709ca1cf069cee24ef8000cb6a479813b5537df
- Get rid of parameterized marked sections in SGML. Previously, we created a variant of the installation instructions
forproducing the plain-text INSTALL file by marking up certain parts of installation.sgml using SGML parameterized
markedsections. Marked sections will not work anymore in XML, so before we can convert the documentation to XML, we
needa new approach. DocBook provides a "profiling" feature that allows selecting content based on attributes, which
wouldwork here. But it imposes a noticeable overhead when building the full documentation and causes complications
whenbuilding some output formats, and given that we recently spent a fair amount of effort optimizing the documentation
buildtime, it seems sad to have to accept that. So as an alternative, (1) we create our own mini-profiling layer that
adjustsjust the text we want, and (2) assemble the pieces of content that we want in the INSTALL file using XInclude.
Thatway, there is no overhead when building the full documentation and most of the "ugly" stuff in installation.sgml
canbe removed and dealt with out of line.
https://git.postgresql.org/pg/commitdiff/684cf76b83e9dc8aed12aeb9131d2208f61bd31f
- Fix plperl build. The changes in 639928c988c1c2f52bbe7ca89e8c7c78a041b3e2 turned out to require Perl 5.9.3, which is
newerthan our minimum required version. So revert back to the old code for the normal case and only use the new
variantwhen both coverage and vpath are used. As the minimum Perl version moves forward, we can drop the old code
sometime.https://git.postgresql.org/pg/commitdiff/65c865620237bf1964757436a36c40af591d30fb
- Improve vpath support in plperl build. Run xsubpp with the -output option instead of redirecting stdout. That
ensuresthat the #line directives in the output file point to the right place in a vpath build. This in turn fixes an
errorin coverage builds that it can't find the source files. Refactor the makefile rules while we're here.
Reviewed-by:Michael Paquier <michael.paquier@gmail.com>
https://git.postgresql.org/pg/commitdiff/639928c988c1c2f52bbe7ca89e8c7c78a041b3e2
- Run only top-level recursive lcov. This is the way lcov was intended to be used. It is much faster and more robust
andmakes the makefiles simpler than running it in each subdirectory. The previous coding ran gcov before lcov, but
thatis useless because lcov/geninfo call gcov internally and use that information. Moreover, this led to complications
andfailures during parallel make. This separates the two targets: You either use "make coverage" to get textual
outputfrom gcov or "make coverage-html" to get an HTML report via lcov. (Using both is still problematic because they
writethe same output files.) Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
https://git.postgresql.org/pg/commitdiff/504923a0ed5c75775196c8ed0cd59b15d55cd39b
- Have lcov exclude external files. Call lcov with --no-external option to exclude external files (for example, system
headerswith inline functions) from output. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
https://git.postgresql.org/pg/commitdiff/66fd86a6a3d2ac9772f977ec43af190ea3fe6ddb
- Remove SGML marked sections. For XML compatibility, replace marked sections <![IGNORE[ ]]> with comments <!-- -->.
Insome cases it seemed better to remove the ignored text altogether, and in one case the text should not have been
ignored.https://git.postgresql.org/pg/commitdiff/22d9764646d03ac7d3419c4fd0effd256568c922
- Add lcov --initial. By just running lcov on the produced .gcda data files, we don't account for source files that
arenot touched by tests at all. To fix that, run lcov --initial to create a base line info file with all zero
counters,and merge that with the actual counters when creating the final report. Reviewed-by: Michael Paquier
<michael.paquier@gmail.com>https://git.postgresql.org/pg/commitdiff/4bb5a2536bcff5dfef9242818979faaa0659b1af
- Add PostgreSQL version to coverage output. Also make overriding the title easier. That helps telling where the
reportcame from and labeling different variants of a report. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
https://git.postgresql.org/pg/commitdiff/d2773f9bcd980cf6ed720928cd0700196608ef19
- Add background worker type. Add bgw_type field to background worker structure. It is intended to be set to the same
valuefor all workers of the same type, so they can be grouped in pg_stat_activity, for example. The backend_type
columnin pg_stat_activity now shows bgw_type for a background worker. The ps listing also no longer calls out that a
processis a background worker but just show the bgw_type. That way, being a background worker is more of an
implementationdetail now that is not shown to the user. However, most log messages still refer to 'background worker
"%s"';otherwise constructing sensible and translatable log messages would become tricky. Reviewed-by: Michael Paquier
<michael.paquier@gmail.com>Reviewed-by: Daniel Gustafsson <daniel@yesql.se>
https://git.postgresql.org/pg/commitdiff/5373bc2a0867048bb78f93aede54ac1309b5e227
- psql: Update \d sequence display. For \d sequencename, the psql code just did SELECT * FROM sequencename to get the
informationto display, but this does not contain much interesting information anymore in PostgreSQL 10, because the
metadatahas been moved to a separate system catalog. This patch creates a newly designed sequence display that is not
merelyan extension of the general relation/table display as it was previously. Reviewed-by: Fabien COELHO
<coelho@cri.ensmp.fr>https://git.postgresql.org/pg/commitdiff/2a14b9609df1de4f2eb5a97aff674aaad033a7e6
- Use Py_RETURN_NONE where suitable. This is more idiomatic style and available as of Python 2.4, which is our
minimum.https://git.postgresql.org/pg/commitdiff/0008a106d4f84206a96fc1fb09a1e6b09f1627ec
- Add list of acknowledgments to release notes. This contains all individuals mentioned in the commit messages during
PostgreSQL10 development. current through babf18579455e85269ad75e1ddb03f34138f77b6 Discussion:
https://www.postgresql.org/message-id/flat/54ad0e42-770e-dfe1-123e-bce9361ad452%402ndquadrant.com
https://git.postgresql.org/pg/commitdiff/655c938fb0aacde297d4c53daf97ff82a3b90fad
Noah Misch pushed:
- Don't recommend "DROP SCHEMA information_schema CASCADE". It drops objects outside information_schema that depend on
objectsinside information_schema. For example, it will drop a user-defined view if the view query refers to
information_schema. Discussion: https://postgr.es/m/20170831025345.GE3963697@rfd.leadboat.com
https://git.postgresql.org/pg/commitdiff/59597e6485847ae40eab2e80ff04af3e8663f2d8
Dean Rasheed pushed:
- Improve the CREATE POLICY documentation. Provide a correct description of how multiple policies are combined,
clarifywhen SELECT permissions are required, mention SELECT FOR UPDATE/SHARE, and do some other more minor tidying up.
Reviewedby Stephen Frost Discussion:
https://postgr.es/m/CAEZATCVrxyYbOFU8XbGHicz%2BmXPYzw%3DhfNL2XTphDt-53TomQQ%40mail.gmail.comBack-patch to 9.5.
https://git.postgresql.org/pg/commitdiff/af44cbd5ecd7e1db0ae4bce75c8f1bce14b1d6db
Álvaro Herrera pushed:
- Fix freezing of a dead HOT-updated tuple. Vacuum calls page-level HOT prune to remove dead HOT tuples before doing
livenesschecks (HeapTupleSatisfiesVacuum) on the remaining tuples. But concurrent transaction commit/abort may turn
DEADsome of the HOT tuples that survived the prune, before HeapTupleSatisfiesVacuum tests them. This happens to
activatethe code that decides to freeze the tuple ... which resuscitates it, duplicating data. (This is especially
badif there's any unique constraints, because those are now internally violated due to the duplicate entries, though
youwon't know until you try to REINDEX or dump/restore the table.) One possible fix would be to simply skip doing
anythingto the tuple, and hope that the next HOT prune would remove it. But there is a problem: if the tuple is older
thanfreeze horizon, this would leave an unfrozen XID behind, and if no HOT prune happens to clean it up before the
containingpg_clog segment is truncated away, it'd later cause an error when the XID is looked up. Fix the problem by
havingthe tuple freezing routines cope with the situation: don't freeze the tuple (and keep it dead). In the cases
thatthe XID is older than the freeze age, set the HEAP_XMAX_COMMITTED flag so that there is no need to look up the XID
inpg_clog later on. An isolation test is included, authored by Michael Paquier, loosely based on Daniel Wood's
originalreproducer. It only tests one particular scenario, though, not all the possible ways for this problem to
surface;it be good to have a more reliable way to test this more fully, but it'd require more work. In message
https://postgr.es/m/20170911140103.5akxptyrwgpc25bw@alvherre.pgsqlI outlined another test case (more closely matching
DanWood's) that exposed a few more ways for the problem to occur. Backpatch all the way back to 9.3, where this
problemwas introduced by multixact juggling. In branches 9.3 and 9.4, this includes a backpatch of commit e5ff9fefcd50
(of9.5 era), since the original is not correctable without matching the coding pattern in 9.5 up. Reported-by: Daniel
WoodDiagnosed-by: Daniel Wood Reviewed-by: Yi Wen Wong, Michaël Paquier Discussion:
https://postgr.es/m/E5711E62-8FDF-4DCA-A888-C200BF6B5742@amazon.com
https://git.postgresql.org/pg/commitdiff/20b655224249e6d2daf7ef0595995228baddb381
Andres Freund pushed:
- Fix typo. Reported-By: Thomas Munro and Jesper Pedersen
https://git.postgresql.org/pg/commitdiff/f14241236ea2e306dc665635665c7f88669b6ca4
- Fix copy & pasto in 510b8cbff15f. Reported-By: Peter Geoghegan
https://git.postgresql.org/pg/commitdiff/248e33756b425335d94a32ffc8e9aace04f82c31
- Extend & revamp pg_bswap.h infrastructure. Upcoming patches are going to address performance issues that involve
slowsystem provided ntohs/htons etc. To address that expand pg_bswap.h to provide pg_ntoh{16,32,64}, pg_hton{16,32,64}
andoptimize their respective implementations by using compiler intrinsics for gcc compatible compilers and msvc. Fall
backto manual implementations using shifts etc otherwise. Additionally remove multiple evaluation hazards from the
existingBSWAP32/64 macros, by replacing them with inline functions when necessary. In the course of that the naming
schemeis changed to pg_bswap16/32/64. Author: Andres Freund Discussion:
https://postgr.es/m/20170927172019.gheidqy6xvlxb325@alap3.anarazel.de
https://git.postgresql.org/pg/commitdiff/510b8cbff15fcece246f66f2273ccf830a6c7e98
Heikki Linnakangas pushed:
- Fix busy-wait in pgbench, with --rate. If --rate was used to throttle pgbench, it failed to sleep when it had
nothingto do, leading to a busy-wait with 100% CPU usage. This bug was introduced in the refactoring in v10. Before
that,sleep() was called with a timeout, even when there were no file descriptors to wait for. Reported by Jeff Janes,
patchby Fabien COELHO. Backpatch to v10. Discussion:
https://www.postgresql.org/message-id/CAMkU%3D1x5hoX0pLLKPRnXCy0T8uHoDvXdq%2B7kAM9eoC9_z72ucw%40mail.gmail.com
https://git.postgresql.org/pg/commitdiff/396ef1561878a5d42ea9191f60098b7fbbec6e41
== Pending Patches ==
Rafia Sabih sent in another revision of a patch to speed up gather.
Alexander Kuzmenkov sent in another revision of a patch to implement CSN-based
snapshots.
Doug Rady sent in a patch to pgbench to break out timing data for the
initialization phases.
Doug Rady sent in a patch to enable building pgbench to use ppoll() instead of
select() to allow for more than (FD_SETSIZE - 10) connections.
Thomas Munro sent in another revision of a patch to get parallel queries working
with SERIALIZABLE isolation mode.
Amit Langote sent in two more revisions of a patch to set pd_lower correctly in
the GIN, BRIN, and SP-GiST metapages.
Peter Geoghegan sent in a patch to consistently canonicalize ICU collations'
collcollate as BCP 47.
Maksim Milyutin sent in two revisions of a patch to fix a cache invalidation
bug which manifests in queries that contain constants of a temporary composite
type.
Shubham Barai sent in another revision of a patch to implement predicate locking
for hash indexes.
Michaël Paquier sent in a patch to shore up some shaky coding for vacuuming
partitioned relations.
Chen Huajun sent in six more revisions of a patch to make pg_rewind to not copy
useless WAL files.
Michaël Paquier sent in another revision of a patch to remove
ALLOW_DANGEROUS_LO_FUNCTIONS for LO-related superuser checks, replace superuser
checks of large object import/export with ACL checks, and move ACL checks for
large objects when opening them.
Beena Emerson sent in a PoC patch to implement runtime partition pruning.
Amit Langote sent in a patch to move certain partitioning code to the executor.
Amit Langote sent in a patch to teach ValidatePartitionConstraints to skip
validation in more cases and skip scanning default partition's child tables if
possible.
Haribabu Kommi sent in another revision of a patch to add a pg_stat_walwrites
statistics view.
Amul Sul sent in a patch to restrict concurrent update/delete with UPDATE of
a partition key.
Emre Hasegeli sent in another revision of a patch to refactor the geometric
functions and operators code, provide a header file for the built-in float
datatypes, use the built-in float datatype to implement geometric types, and fix
some obvious problems around the line datatype.
Pavel Stěhule sent in two more revisions of a patch to add default namespaces
for XPath expressions.
Michaël Paquier sent in a patch to fix an infelicity in the use of RangeVar for
partitioned tables in autovacuum.
Jeevan Chalke sent in another revision of a patch to implement partition-wise
aggregation/grouping.
Kyotaro HORIGUCHI and Yura Sokolov traded patches to add a failing test:
wal_sender_timeout+logical decoding of a big transaction, and fix walsender
timeouts when decoding large transaction.
Nathan Bossart sent in three more revisions of a patch to enable specifying
multiple tables in VACUUM.
Stas Kelvich sent in another revision of a patch to fix an issue with
transactions involving multiple postgres foreign servers by adding a contrib
extension called fdw_transaction_resovler.
Amit Langote sent in another revision of a patch to make planner-side changes
for partition-pruning, interface changes for partition_bound_{cmp/bsearch},
implement get_partitions_for_keys(), and add more tests for the new
partitioning-related planning code.
Amul Sul and Amit Langote traded patches to implement hash partitioning.
Etsuro Fujita sent in a patch to change postgresPlanForeignModify so that it
handles "with check option" the same way as for the RETURNING case.
Alexander Kuzmenkov sent in another revision of a patch to implement full merge
join on comparison clause.
Tom Lane sent in a patch to modifies eqjoinsel_semi by replacing the previous
number-of-distinct-values estimate for the inner rel inner_rel->rows,
effectively assuming that the inside of the IN or EXISTS is unique, and dropping
the fallback to selectivity 0.5 altogether, instead applying the nd1 vs nd2
heuristic all the time.
Jesper Pedersen sent in a patch to change the message for restarting a server
from a directory without a PID file to account for the case where a restart
happens after an initdb.
Andres Freund sent in a patch to speed up fmgr_isbuiltin() by keeping an oid ->
builtin mapping.
Fabien COELHO sent in a patch to fix an issue where pgbench would fail but get
stuck with 100% CPU usage.
Konstantin Knizhnik sent in two revisions of a patch to add
pg_prepared_xact_status().
Amit Khandekar sent in another revision of a patch to implement parallel append.
Martin Marques sent in two revisions of a patch to add an option to
pg_basebackup to output messages as if it were running in batch-mode, as opossed
to running in a tty.
Nikolay Shaplov sent in a patch to add a series of tests that triggers
reloptions related code in all access methods.
Shubham Barai sent in another revision of a patch to implement predicate locking
for GIN indexes.
Nikita Glukhov sent in another revision of a patch to implement SQL/JSON.
--
Sent via pgsql-announce mailing list (pgsql-announce@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-announce