== PostgreSQL Weekly News - February 23, 2020 == - Mailing list pgsql-announce
| From | David Fetter |
|---|---|
| Subject | == PostgreSQL Weekly News - February 23, 2020 == |
| Date | |
| Msg-id | 20200223231050.GA6334@fetter.org Whole thread Raw |
| List | pgsql-announce |
== PostgreSQL Weekly News - February 23, 2020 == PGConfNepal 2020 will be held April 17-18, 2020 at Kathmandu University, Dhulikhel, Nepal. The CfP is open until February 28, 2020 at https://pgconf.org.np/call-for-paper/ https://pgconf.org.np/ == PostgreSQL Product News == Pgpool-II 4.1.1, 4.0.8, 3.7.13, 3.6.20, 3.5.24 and pgpoolAdmin 4.1.0 http://www.pgpool.net/docs/latest/en/html/release.html http://pgpool.net/mediawiki/index.php/Downloads PostGIS 3.1.0, the industry standard geographic information system package for PostgreSQL, released. https://git.osgeo.org/gitea/postgis/postgis/src/tag/3.0.1/NEWS pglogical 2.3, a logical-WAL-based replication system for PostgreSQL, released. https://www.2ndquadrant.com/en/resources/pglogical/release-notes/ Joe bot 0.5.0, a Slack chatbot that helps backend developers and DBAs troubleshoot and optimize PostgreSQL queries, releaesd. https://gitlab.com/postgres-ai/joe/-/releases#0.5.0 pg_partman 4.3.0, a management system for partitioned tables, released. https://github.com/pgpartman/pg_partman postgres-checkup 1.4.0, a tool that automates detailed health checks of PostgreSQL clusters, released. https://gitlab.com/postgres-ai/postgres-checkup/-/releases wal2json 2.2, a JSON output plugin for changeset extraction, released. As this release fixes important bugs in PostgreSQL 9.4 and 9.5, it would be good to upgrade ASAP. https://github.com/eulerto/wal2json/releases == PostgreSQL Jobs for February == http://archives.postgresql.org/pgsql-jobs/2020-02/ == PostgreSQL Local == pgDay Israel 2020 will take place on March 19, 2020 in Tel Aviv. http://pgday.org.il/ pgDay Paris 2020 will be held in Paris, France on March 26, 2020 at Espace Saint-Martin. https://2020.pgday.paris/ Nordic PGDay 2020 will be held in Helsinki, Finland at the Hilton Helsinki Strand Hotel on March 24, 2020. PGConf India 2020 will be on February 26-28, 2020 in Bengaluru, Karnataka. http://pgconf.in/ PostgreSQL@SCaLE is a two day, two track event which takes place on March 5-6, 2020, at Pasadena Convention Center, as part of SCaLE 18X. https://www.socallinuxexpo.org/scale/18x/postgresscale The German-speaking PostgreSQL Conference 2020 will take place on May 15, 2019 in Stuttgart. PGCon 2020 will take place in Ottawa on May 26-29, 2020. https://www.pgcon.org/2020/ PGDay.IT 2020 will take place June 11-12 in Bergamo, Italy. https://2020.pgday.it/en/ Swiss PGDay 2020 will take place in Rapperswil (near Zurich) on June 18-19, 2020. The Call for Speakers is open through March 17, 2020. https://www.pgday.ch/2020/ PostgresLondon 2020 will be July 7-8, 2020 with an optional training day on July 6. The CfP is open at https://forms.gle/5m8ybUt9YDZG4gVU7 through March 27, 2020. http://postgreslondon.org PG Day Russia will be in Saint Petersburg on July 10, 2020. The CfP is open at https://pgday.ru/en/2020/for-speakers through April 6, 2020. https://pgday.ru/en/2020/ FOSS4G 2020, will take place in Calgary, Alberta, Canada August 24-29 2020. the Call for Papers is currently open at https://2020.foss4g.org/speakers/ https://2020.foss4g.org/ Austrian pgDay will take place September 18, 2020 at Schloss Schoenbrunn (Apothekertrakt) in Vienna. The CfP is open until April 19, 2020 at https://pgday.at/en/talk-commitee/ https://pgday.at/en/ == 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@fetter.org. == Applied Patches == Fujii Masao pushed: - Add description about LogicalRewriteTruncate wait event into document. Back-patch to v10 where commit 249cf070e3 introduced LogicalRewriteTruncate wait event. Author: Fujii Masao Reviewed-by: Michael Paquier Discussion: https://postgr.es/m/949931aa-4ed4-d867-a7b5-de9c02b2292b@oss.nttdata.com https://git.postgresql.org/pg/commitdiff/e593148d43adb8ef1721ccd987046c02f4532920 - Add description about GSSOpenServer wait event into document. This commit also updates wait event enum into alphabetical order. Previously the enum entry for GSSOpenServer was added out-of-order. Back-patch to v12 where commit b0b39f72b9 introduced GSSOpenServer wait event. In v12, the commit doesn't include the update of wait event enum, not to break ABI. Author: Fujii Masao Reviewed-by: Michael Paquier Discussion: https://postgr.es/m/949931aa-4ed4-d867-a7b5-de9c02b2292b@oss.nttdata.com https://git.postgresql.org/pg/commitdiff/f4ae7221413d5176e4d5a49883217ea87312104a - Make inherited LOCK TABLE perform access permission checks on parent table only. Previously, LOCK TABLE command through a parent table checked the permissions on not only the parent table but also the children tables inherited from it. This was a bug and inherited queries should perform access permission checks on the parent table only. This commit fixes LOCK TABLE so that it does not check the permission on children tables. This patch is applied only in the master branch. We decided not to back-patch because it's not hard to imagine that there are some applications expecting the old behavior and the change breaks their security. Author: Amit Langote Reviewed-by: Fujii Masao Discussion: https://postgr.es/m/CAHGQGwE+GauyG7POtRfRwwthAGwTjPQYdFHR97+LzA4RHGnJxA@mail.gmail.com https://git.postgresql.org/pg/commitdiff/b7e51b350c4e6b1cb3404588cf11801525e2336f - Fix mesurement of elapsed time during truncating heap in VACUUM. VACUUM may truncate heap in several batches. The activity report is logged for each batch, and contains the number of pages in the table before and after the truncation, and also the elapsed time during the truncation. Previously the elapsed time reported in each batch was the total elapsed time since starting the truncation until finishing each batch. For example, if the truncation was processed dividing into three batches, the second batch reported the accumulated time elapsed during both first and second batches. This is strange and confusing because the number of pages in the table reported together is not total. Instead, each batch should report the time elapsed during only that batch. The cause of this issue was that the resource usage snapshot was initialized only at the beginning of the truncation and was never reset later. This commit fixes the issue by changing VACUUM so that the resource usage snapshot is reset at each batch. Back-patch to all supported branches. Reported-by: Tatsuhito Kasahara Author: Tatsuhito Kasahara Reviewed-by: Masahiko Sawada, Fujii Masao Discussion: https://postgr.es/m/CAP0=ZVJsf=NvQuy+QXQZ7B=ZVLoDV_JzsVC1FRsF1G18i3zMGg@mail.gmail.com https://git.postgresql.org/pg/commitdiff/007491979461ff10d487e1da9bcc87f2fd834f26 Peter Eisentraut pushed: - Fill in extraUpdatedCols in logical replication. The extraUpdatedCols field of the target RTE records which generated columns are affected by an update. This is used in a variety of places, including per-column triggers and foreign data wrappers. When an update was initiated by a logical replication subscription, this field was not filled in, so such an update would not affect generated columns in a way that is consistent with normal updates. To fix, factor out some code from analyze.c to fill in extraUpdatedCols in the logical replication worker as well. Reviewed-by: Pavel Stehule <pavel.stehule@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/b05e781a-fa16-6b52-6738-761181204567@2ndquadrant.com https://git.postgresql.org/pg/commitdiff/ad3ae64770e5b4391ad07fb2a689d72aa45b4689 - Optimize update of tables with generated columns. When updating a table row with generated columns, only recompute those generated columns whose base columns have changed in this update and keep the rest unchanged. This can result in a significant performance benefit. The required information was already kept in RangeTblEntry.extraUpdatedCols; we just have to make use of it. Reviewed-by: Pavel Stehule <pavel.stehule@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/b05e781a-fa16-6b52-6738-761181204567@2ndquadrant.com https://git.postgresql.org/pg/commitdiff/c6679e4fca21d253ced84c51ac1a31c1b2aec72f - Set gen_random_uuid() to volatile. It was set to immutable. This was a mistake in the initial commit (5925e5549890416bcf588334d9d0bc99f8ad6c7f). Reported-by: hubert depesz lubaczewski <depesz@depesz.com> Discussion: https://www.postgresql.org/message-id/flat/20200218185452.GA8710%40depesz.com https://git.postgresql.org/pg/commitdiff/2ed19a488edb3980e054e6064e7028ccb652e5df - Fix typo. Reported-by: Daniel Verite <daniel@manitou-mail.org> https://git.postgresql.org/pg/commitdiff/2f9c46a32b43d72c9384378827ee51fde896807c - Require stdint.h. stdint.h belongs to the compiler (as opposed to inttypes.h), so by requiring a C99 compiler we can also require stdint.h unconditionally. Remove configure checks and other workarounds for it. This also removes a few steps in the required portability adjustments to the imported time zone code, which can be applied on the next import. When using GCC on a platform that is otherwise pre-C99, this will now require at least GCC 4.5, which is the first release that supplied a standard-conforming stdint.h if the native platform didn't have it. Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us> Discussion: https://www.postgresql.org/message-id/flat/5d398bbb-262a-5fed-d839-d0e5cff3c0d7%402ndquadrant.com https://git.postgresql.org/pg/commitdiff/957338418b69e9774ccc1bab59f765a62f0aa6f9 - Fix compiler warnings on 64-bit Windows. GCC reports various instances of warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] warning: cast from pointer to integer of different size [-Wpointer-to-int-cast] and MSVC equivalently warning C4312: 'type cast': conversion from 'int' to 'void *' of greater size warning C4311: 'type cast': pointer truncation from 'void *' to 'long' in ECPG test files. This is because void* and long are cast back and forth, but on 64-bit Windows, these have different sizes. Fix by using intptr_t instead. The code actually worked fine because the integer values in use are all small. So this is just to get the test code to compile warning-free. This change is simplified by having made stdint.h required (commit 957338418b69e9774ccc1bab59f765a62f0aa6f9). Before this it would have been more complicated because the ecpg test source files don't use the full pg_config.h. Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us> Discussion: https://www.postgresql.org/message-id/flat/5d398bbb-262a-5fed-d839-d0e5cff3c0d7%402ndquadrant.com https://git.postgresql.org/pg/commitdiff/3f9c1697dca0b4964f1f5ba624d361d4e0e53051 - Allow running src/tools/msvc/mkvcbuild.pl under not Windows. This to allow verifying the MSVC build file generation without having to have Windows. To do this, we avoid Windows-specific Perl modules and don't run the "cl" compiler or "nmake". The resulting build files won't actually be completely correct, but it's useful enough. Reviewed-by: Michael Paquier <michael@paquier.xyz> Reviewed-by: Julien Rouhaud <rjuju123@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/d73b2c7b-f081-8357-8422-7564d55f1aac%402ndquadrant.com https://git.postgresql.org/pg/commitdiff/73c8596488fd5fd619991f56dae5d22f551b06d9 - Fix perlcritic warnings. https://git.postgresql.org/pg/commitdiff/b24e125696a94b6ebdd2262c0c41bf45cbeba23f - Factor out InitControlFile() from BootStrapXLOG(). Right now this only makes BootStrapXLOG() a bit more manageable, but in the future there may be external callers. Discussion: https://www.postgresql.org/message-id/e8f86ba5-48f1-a80a-7f1d-b76bcb9c5c47@2ndquadrant.com https://git.postgresql.org/pg/commitdiff/79c2385915dd4aa43127e766c3dce323ec562ba0 - Reformat code comment. Discussion: https://www.postgresql.org/message-id/e8f86ba5-48f1-a80a-7f1d-b76bcb9c5c47@2ndquadrant.com https://git.postgresql.org/pg/commitdiff/9745f93afc56829f9cf10ca3e43a29f0b4409fe3 - pg_resetwal: Rename function to avoid potential conflict. ReadControlFile() here conflicts with a function of the same name in xlog.c. There is no actual conflict right now, but since pg_resetwal.c reaches deep inside backend headers, it's possible in the future. Discussion: https://www.postgresql.org/message-id/e8f86ba5-48f1-a80a-7f1d-b76bcb9c5c47@2ndquadrant.com https://git.postgresql.org/pg/commitdiff/5f1b8260af961ddf05968c9c3e7f178a3ad5d2ea Tom Lane pushed: - Teach pg_dump to dump comments on RLS policy objects. This was unaccountably omitted in the original RLS patch. The SQL syntax is basically the same as for comments on triggers, so crib code from dumpTrigger(). Per report from Marc Munro. Back-patch to all supported branches. Discussion: https://postgr.es/m/1581889298.18009.15.camel@bloodnok.com https://git.postgresql.org/pg/commitdiff/f31364676dfd9e7a2b712ea0da7b4ab6b7d28f5e - Fix confusion about event trigger vs. plain function in plpgsql. The function hash table keys made by compute_function_hashkey() failed to distinguish event-trigger call context from regular call context. This meant that once we'd successfully made a hash entry for an event trigger (either by validation, or by normal use as an event trigger), an attempt to call the trigger function as a plain function would find this hash entry and thereby bypass the you-can't-do-that check in do_compile(). Thus we'd attempt to execute the function, leading to strange errors or even crashes, depending on function contents and server version. To fix, add an isEventTrigger field to PLpgSQL_func_hashkey, paralleling the longstanding infrastructure for regular triggers. This fits into what had been pad space, so there's no risk of an ABI break, even assuming that any third-party code is looking at these hash keys. (I considered replacing isTrigger with a PLpgSQL_trigtype enum field, but felt that that carried some API/ABI risk. Maybe we should change it in HEAD though.) Per bug #16266 from Alexander Lakhin. This has been broken since event triggers were invented, so back-patch to all supported branches. Discussion: https://postgr.es/m/16266-fcd7f838e97ba5d4@postgresql.org https://git.postgresql.org/pg/commitdiff/761a5688b179d46091e7314458571c5fc537084a - Remove support for upgrading extensions from "unpackaged" state. Andres Freund pointed out that allowing non-superusers to run "CREATE EXTENSION ... FROM unpackaged" has security risks, since the unpackaged-to-1.0 scripts don't try to verify that the existing objects they're modifying are what they expect. Just attaching such objects to an extension doesn't seem too dangerous, but some of them do more than that. We could have resolved this, perhaps, by still requiring superuser privilege to use the FROM option. However, it's fair to ask just what we're accomplishing by continuing to lug the unpackaged-to-1.0 scripts forward. None of them have received any real testing since 9.1 days, so they may not even work anymore (even assuming that one could still load the previous "loose" object definitions into a v13 database). And an installation that's trying to go from pre-9.1 to v13 or later in one jump is going to have worse compatibility problems than whether there's a trivial way to convert their contrib modules into extension style. Hence, let's just drop both those scripts and the core-code support for "CREATE EXTENSION ... FROM". Discussion: https://postgr.es/m/20200213233015.r6rnubcvl4egdh5r@alap3.anarazel.de https://git.postgresql.org/pg/commitdiff/70a7732007bc4689f4c7a44e738eb2d892dac1e3 - Doc: discourage use of partial indexes for poor-man's-partitioning. Creating a bunch of non-overlapping partial indexes is generally a bad idea, so add an example saying not to do that. Back-patch to v10. Before that, the alternative of using (real) partitioning wasn't available, so that the tradeoff isn't quite so clear cut. Discussion: https://postgr.es/m/CAKVFrvFY-f7kgwMRMiPLbPYMmgjc8Y2jjUGK_Y0HVcYAmU6ymg@mail.gmail.com https://git.postgresql.org/pg/commitdiff/6a8e5605b53561464c53006337596699f41b1de2 - Assume that we have utime() and <utime.h>. These are required by POSIX since SUSv2, and no live platforms fail to provide them. On Windows, utime() exists and we bring our own <utime.h>, so we're good there too. So remove the configure probes and ad-hoc substitute code. We don't need to check for utimes() anymore either, since that was only used as a substitute. In passing, make the Windows build include <sys/utime.h> only where we need it, not everywhere. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/481c8e9232386e289fbd3e8f75893ae8d9814298 - Assume that we have <wchar.h>. Windows has this, and so do all other live platforms according to the buildfarm; it's been required by POSIX since SUSv2. So remove the configure probe and tests of HAVE_WCHAR_H. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/97cf1fa4ed57cc1ed556ac4f827e7df8538f2d29 - Assume that we have isinf(). Windows has this, and so do all other live platforms according to the buildfarm, so remove the configure probe and src/port/ substitution. This also lets us get rid of some configure probes that existed only to support src/port/isinf.c. I kept the port.h hack to force using __builtin_isinf() on clang, though. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/7fde892bc191e4df9fcd52ce11d1502673498d97 - Assume that we have memmove(). Windows has this, and so do all other live platforms according to the buildfarm, so remove the configure probe and c.h's substitute code. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/1200d71a09177b7e64167df440d0c9bd16111ebe - Assume that we have cbrt(). Windows has this, and so do all other live platforms according to the buildfarm, so remove the configure probe and float.c's substitute code. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/abe41f453a5c42129e21825d75450aced0053eb8 - Assume that we have rint(). Windows has this since _MSC_VER >= 1200, and so do all other live platforms according to the buildfarm, so remove the configure probe and src/port/ substitution. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/f88a058200a40032e88a8bfc1aea453c22d2dcb0 - Assume that we have functional, 64-bit fseeko()/ftello(). Windows has this, and so do all other live platforms according to the buildfarm, so remove the configure probe and src/port/ substitution. Keep the probe that detects whether _LARGEFILE_SOURCE has to be defined to get that, though ... that seems to be still relevant in some places. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/799d22461a932aace890d61a82186e0d64de0ee8 - Assume that we have signed integral types and flexible array members. These compiler features are required by C99, so remove the configure probes for them. This is part of a series of commits to get rid of no-longer-relevant configure checks and dead src/port/ code. I'm committing them separately to make it easier to back out individual changes if they prove less portable than I expect. Discussion: https://postgr.es/m/15379.1582221614@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/f4d59369d2ddf0ad7850112752ec42fd115825d4 - Adjust Solution.pm to set HAVE_STDINT_H. We're not testing that anywhere anymore, but for consistency, it should get defined. https://git.postgresql.org/pg/commitdiff/ec4a7851d5914bbdc5b65175e2489ec19020217e Michaël Paquier pushed: - Fix grammar in monitoring.sgml. This is related to progress reporting for ANALYZE and partitioned tables. Author: Amit Langote Reviewed-by: Daniel Gustafsson, Julien Rouhaud Discussion: https://postgr.es/m/CA+HiwqGx6C=-bFTX=ryMThyvM7CcSC3b1x8_6zh4Uo41Kvu-zw@mail.gmail.com https://git.postgresql.org/pg/commitdiff/ddfc3c1499b05956094a7f219fe2b6ac0c9ce5eb - Remove duplicated words in comments. Author: Daniel Gustafsson Reviewed-by: Vik Fearing Discussion: https://postgr.es/m/EBC3BFEB-664C-4063-81ED-29F1227DB012@yesql.se https://git.postgresql.org/pg/commitdiff/958f9fb98da97a1d781a3766ea86bcdba3b216c5 - Clean up some code, comments and docs referring to Windows 2000 and older. This fixes and updates a couple of comments related to outdated Windows versions. Particularly, src/common/exec.c had a fallback implementation to read a file's line from a pipe because stdin/stdout/stderr does not exist in Windows 2000 that is removed to simplify src/common/ as there are unlikely versions of Postgres running on such platforms. Author: Michael Paquier Reviewed-by: Kyotaro Horiguchi, Juan José Santamaría Flecha Discussion: https://postgr.es/m/20191219021526.GC4202@paquier.xyz https://git.postgresql.org/pg/commitdiff/e2e02191e23379502a38a6b0436ab7f41b2efc08 - Cleanup more code related to ws2_32.dll loading in src/port/getaddrinfo.c. e2e0219 has removed a code path for Windows 2000 that attempts to load wship6.dll as fallback if ws2_32.dll is found but not getaddrinfo(), leaving behind a dangling pointer as the library is freed. However, there is no point in this check as ws2_32.dll exists since Windows XP, so just remove the duplicated check. Reported-by: Tom Lane Discussion: https://postgr.es/m/9781.1582146114@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/d55e9ae2633ee462d554dddc08d640d725dd6494 - Doc: Fix instructions to control build environment with MSVC. The documentation included some outdated instructions to change the architecture, build type or target OS of a build done with MSVC. This commit updates the documentation to include the modern options available, down to Visual Studio 2013. Reported-by: Kyotaro Horiguchi Author: Juan José Santamaría Flecha Discussion: https://postgr.es/m/CAC+AXB0J7tAqW_2F1fCE4Dh2=Ccz96TcLpsGXOCvka7VvWG9Qw@mail.gmail.com Backpatch-through: 12 https://git.postgresql.org/pg/commitdiff/dca3911a81f0ba823b56c3d4462419c83f385e55 Peter Geoghegan pushed: - Remove obsolete _bt_compare() comment. btbuild() has nothing to say about how NULL values compare in nbtree. Besides, there are _bt_compare() header comments that describe how NULL values are handled. https://git.postgresql.org/pg/commitdiff/fe9b92854e7d5f66d3abb565c5b60a879b6c9442 Amit Kapila pushed: - Stop demanding that top xact must be seen before subxact in decoding. Manifested as ERROR: subtransaction logged without previous top-level txn record this check forbids legit behaviours like - First xl_xact_assignment record is beyond reading, i.e. earlier restart_lsn. - After restart_lsn there is some change of a subxact. - After that, there is second xl_xact_assignment (for another subxact) revealing the relationship between top and first subxact. Such a transaction won't be streamed anyway because we hadn't seen it in full. Saying for sure whether xact of some record encountered after the snapshot was deserialized can be streamed or not requires to know whether it wrote something before deserialization point --if yes, it hasn't been seen in full and can't be decoded. Snapshot doesn't have such info, so there is no easy way to relax the check. Reported-by: Hsu, John Diagnosed-by: Arseny Sher Author: Arseny Sher, Amit Kapila Reviewed-by: Amit Kapila, Dilip Kumar Backpatch-through: 9.5 Discussion: https://postgr.es/m/AB5978B2-1772-4FEE-A245-74C91704ECB0@amazon.com https://git.postgresql.org/pg/commitdiff/e3ff789acfb2754cd7b5e87f6f4463fd08e35996 Jeff Davis pushed: - logtape.c: allocate read buffer even for an empty tape. Prior to this commit, the read buffer was allocated at the time the tape was rewound; but as an optimization, would not be allocated at all if the tape was empty. That optimization meant that it was valid to have a rewound tape with the buffer set to NULL, but only if a number of conditions were met and only if the API was used properly. After 7fdd919a refactored the code to support lazily-allocating the buffer, Coverity started complaining. The optimization for empty tapes doesn't seem important, so just allocate the buffer whether the tape has any data or not. Discussion: https://postgr.es/m/20351.1581868306%40sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/8021985d791902a9eeda51acdede759fbc67ae01 - Minor refactor of nodeAgg.c. * Separate calculation of hash value from the lookup. * Split build_hash_table() into two functions. * Change lookup_hash_entry() to return AggStatePerGroup. That's all the caller needed, anyway. These changes are to support the upcoming Disk-based Hash Aggregation work. Discussion: https://postgr.es/m/31f5ab871a3ad5a1a91a7a797651f20e77ac7ce3.camel%40j-davis.com https://git.postgresql.org/pg/commitdiff/5b618e1f48aecc66e3a9f60289491da520faae19 - Fixup for nodeAgg.c refactor. Commit 5b618e1f made an unintended behavior change. https://git.postgresql.org/pg/commitdiff/b7fabe80df9a65010bfe5e5d0a979bacebfec382 Etsuro Fujita pushed: - Remove extra word from comment. https://git.postgresql.org/pg/commitdiff/53b01acd463d64c385db0ab728b8cc336549a315 - Avoid redundant checks in partition_bounds_copy(). Previously, partition_bounds_copy() checked whether the strategy for the given partition bounds was hash or not, and then determined the number of elements in the datums in the datums array for the partition bounds, on each iteration of the loop for copying the datums array, but there is no need to do that. Perform the checks only once before the loop iteration. Author: Etsuro Fujita Reported-by: Amit Langote and Julien Rouhaud Discussion: https://postgr.es/m/CAPmGK14Rvxrm8DHWvCjdoks6nwZuHBPvMnWZ6rkEx2KhFeEoPQ@mail.gmail.com https://git.postgresql.org/pg/commitdiff/032f9ae012d87c730be0e261a6bae9323c2b2d6c Álvaro Herrera pushed: - Simplify FK-to-partitioned regression test query. Avoid a join between relations having the FK to detect FK violation. The planner might optimize this considering the PK must exist on the referenced side at some point, effectively masking a bug this test tries to detect. Tom Lane and Jehan-Guillaume de Rorthais Discussion: https://postgr.es/m/467.1581270529@sss.pgh.pa.us https://git.postgresql.org/pg/commitdiff/b2304a71748a3e664a1b8354d746f28c38aaa816 == Pending Patches == Surafel Temesgen sent in two more revisions of a patch to enable conflict handling in COPY ... FROM ... Asif Rehman sent in another revision of a patch to document parallel backup. Dmitry Dolgov sent in another revision of a patch to implement index skip scans. Álvaro Herrera sent in two more revisions of a patch to fix ALTER ... DEPENDS ON EXTENSION ... Sergei Kornilov and Michaël Paquier traded patches to make wal_receiver create a temporary replication slot. Amit Langote sent in a patch to change an "on" to an "in" in the documentation of progress reporting. Fujii Masao sent in another revision of a patch to ensure that wait events are reported while waiting for WAL archiving to finish. Álvaro Herrera sent in a patch to record parents of triggers in pg_trigger. Kasahara Tatsuhito sent in two more revisions of a patch to make a small improvement of the elapsed time for truncating heap in vacuum. Fujii Masao sent in two more revisions of a patch to add pg_stat_progress_basebackup, which reports progress for pg_basebackup, on the server side. Justin Pryzby sent in three more revisions of a patch to make vacuum errcontext show the block being processed and adds a callback for truncation. David Zhang sent in another revision of a patch to fix psql to report IO errors on writing query results. Alexander Korotkov sent in two more revisions of a patch to improve checking for missing parent links by traversing from one downlink to subsequent using rightlinks instead of collecting a lossy bitmap. Andrew Dunstan sent in another revision of a patch to add an SSL passphrase callback. Hubert Zhang sent in another revision of a patch to print the physical file path when verifying a checksum fails. Amit Langote sent in five revisions of a patch to reduce the plan cache overhead on plpgsql expressions. Kyotaro HORIGUCHI sent in a patch to fix some misbehaviors for repeated builds on Windows. Jeff Davis sent in three more revisions of a patch to implement memory-bounded hash aggregation. David Zhang sent in another revision of a patch to add a Fastpath while arranging the changes in LSN order in logical decoding. Masahiko Sawada sent in another revision of a patch to fix some issues in copy_slot. Fujii Masao sent in another revision of a patch to fix the descriptions of RecoveryWalAll and RecoveryWalStream wait events. Hamid Akhtar sent in a patch to throw warnings when an autovacuum worker encounters an orphaned prepared transaction and when a vacuum command is issued. It also introduces two new GUCs: max_age_prepared_xacts, and prepared_xacts_vacuum_warn_timeout. Fujii Masao sent in a patch to regularize the size of password client utilities allow. Ildar Musin and Amit Langote traded patches to fix an issue where NO INHERIT was ignored in CREATE TABLE LIKE. Kyotaro HORIGUCHI sent in a patch to fix an issue where pg_regress cleans up a tablespace twice. Mark Dilger sent in two more revisions of a patch to turn Portal->commandTag into an enum. Juan José Santamaría Flecha sent in another revision of a patch to fix the output colors on Windows. Justin Pryzby sent in another revision of a patch to make explain HashAggregate report bucket and memory stats. Noah Misch and Kyotaro HORIGUCHI traded patches to fix the WAL-skipping feature. Michaël Paquier and Bernd Helmle traded patches to make scanning pg_tblspc more robust. Soumyadeep Chakraborty sent in another revision of a patch to Resolve PL handler names for JITed code instead of using const pointers. Thomas Munro sent in a WIP patch to add infrastruction for hardware transactional memory and use same to implement serializable snapshot isolation. Kirill Bychik sent in another revision of a patch to calculate WAL usage. Fabrízio de Royes Mello sent in another revision of a patch to fix an infelicity between parallel pg_restore and event triggers. Andres Freund sent in a PoC patch to improve heavyweight locks so they can be used more generally. Peter Eisentraut sent in another revision of a patch to change client-side fsync_fname() to report errors fatally. Álvaro Herrera sent in another revision of a patch to add uri percent-encoding for binary data. Yuzuko Hosoya and Amit Kapila traded patches to fix an infelicity between autovacuum and partitioned tables. Etsuro Fujita sent in a patch to partition_bounds_copy() to move the setting of the hash_part and natts variables outside a loop, making it more efficient. Michaël Paquier and Bernd Helmle traded patches to make pg_checksums skip foreign tablespace directories. Peter Eisentraut sent in another revision of a patch to add the optiono to include the backend type in log_line_prefix. Zeng Wenjing sent in another revision of a patch to implement global temporary tables. Ants Aasma sent in a PoC patch to use SIMD instructions to make COPY ... FROM ... CSV more efficient. Masahiko Sawada sent in another revision of a patch to support atomic commit among multiple foreign servers. John Dent sent in another revision of a patch to implement UNNEST for REFCURSORs. Julien Rouhaud sent in two more revisions of a patch to avoid reindexing invalid indexes on TOAST tables. Alexander Korotkov sent in another revision of a patch to add a pg_shmem_init_time() function returning the shmem initialization time. Daniel Gustafsson sent in a patch to fix a bug in a previous patch to use heap_multi_insert() for pg_attribute/depend insertions. Tom Lane sent in a patch to account honestly for external FD usage.
pgsql-announce by date: