Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Historically, Yugabyte merged the following upstream PG commits - YB 984667e: PG ab5e9caa4a3ec4765348a0482e88edcf3f6aab4a (tag: REL_10_4) - YB cb4bd4c: PG 6cd404b344f7e27f4d64555bb133f18a758fe851 (tag: REL_11_2) This commit is the start of the third merge: PG 78ec02d612a9b69039ec2610740f738968fe144d (tag: REL_15_2). The initial merge was a series of commits primarily authored by neil. I, jason, squash those commits together into this one because his individual commits were authored in a way that makes the history hard to read. I also use YugaBot as the official author/committer as that is how it's historically done. I leave the original commit messages below as reference. commit a6974e5 Author: neil <[email protected]> AuthorDate: Wed Oct 26 15:54:10 2022 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:22 2023 -0700 Merge with some of Postgres 15 Summary: Merge with PG13 Test Plan: None Suppressed compiling Postgres's extension. Removed obsolete Postgres's types from docdb commit f073ef1 Author: Emily Na <[email protected]> AuthorDate: Tue Jan 24 12:18:49 2023 -0500 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:22 2023 -0700 [MERGE PG13][#14509]YSQL: Fix merge conflict occuring while checking the condition for creating a primary key index Summary: The attribute `yb_is_add_primary_key` was created as part of the `AlterTableCmd` struct to indicate this alter command is for adding primary key. In the past, there were 2 subcommands involved when creating a primary key: 1) `AddIndex` 2) `SetNotNull` (only if any one of the primary key attributes are not set as NOT NULL). In PG13, setting the primary key columns to NOT NULL is triggered in a separate API called `transformIndexConstraint`. The boolean flag has been set in the corresponding section where `AlterTableCmd` is prepared to alter the primary key column to be not null. Here is the excerpt from the PG13 comment mentioning the need to satisfy NOT NULL requirement prior to creating primary key index: ```Now we expect that the parser inserted any required ALTER TABLE SET NOT NULL operations before trying to create a primary-key index.``` Test Plan: Jenkins Reviewers: neil, myang Reviewed By: myang Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D21331 commit 85d6668 Author: Jason Kim <[email protected]> AuthorDate: Fri Jan 27 12:11:07 2023 -0800 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:22 2023 -0700 [MERGE PG13] ysql: fix some compilation errors Summary: Fix two compilation errors: - release build got -Wsometimes-uninitialized compilation error because of Assert(false) that is treated differently in release/debug builds - gcc build got -Werror=comment because of bad comment Test Plan: Jenkins: compile only, rebase: pg13 ./yb_build.sh fastdebug --gcc11 --no-initdb ./yb_build.sh release --clang15 --no-initdb Reviewers: smishra, neil Reviewed By: neil Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D22577 commit b15ae1e Author: Fizaa Luthra <[email protected]> AuthorDate: Thu Feb 2 11:58:29 2023 -0500 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:23 2023 -0700 [MERGE PG13][#14506] Fix YbClearCurrentTransactionId() Summary: YbClearCurrentTransactionId() was introduced to handle temp tables created during concurrent materialized view refreshes correctly. PG13 no longer uses `CurrentTransactionState->transactionId` and `MyPgXact`. This diff changes the code to use the PG13 equivalent of these fields/structures - `CurrentTransactionState->fullTransactionId` and `MyProc->xid`. Test Plan: Jenkins: skip Reviewers: neil, smishra Reviewed By: neil, smishra Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D21721 commit f675a5f Author: Jason Kim <[email protected]> AuthorDate: Wed Feb 22 18:23:38 2023 -0800 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:23 2023 -0700 [MERGE PG15] ysql: add yb_make_greater_string Summary: When migrating from PG 11.2 to PG 15, we come across commit 49fa99e54ec0ded180b52a4a14e543702d53e8c9 that makes make_greater_string private: Move pattern selectivity code from selfuncs.c to like_support.c. ... This change localizes the use of pattern_fixed_prefix() and make_greater_string() so that they no longer need be exported. (We might get pushback from extensions about that, perhaps, in which case I'd be inclined to re-export them in a new header file like_support.h.) ... ybgin uses make_greater_string for tsvector prefix logic. Since it is not trivial to indirectly get the benefit of make_greater_string through the publicly exposed support functions (they take complex Node argument), lean towards re-exposing the function through new header yb_like_support.h and new function yb_make_greater_string. If, in the future, upstream postgres decides to expose make_greater_string, these changes can be scrapped in favor of using make_greater_string again. Test Plan: Jenkins: compile only, rebase: pg15 Reviewers: neil, amartsinchyk Reviewed By: amartsinchyk Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D23110 commit 68f9934 Author: Jason Kim <[email protected]> AuthorDate: Wed Mar 1 18:34:27 2023 -0800 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:23 2023 -0700 [MERGE PG15] test: fix python doctest failure Summary: build-support/jenkins/build-and-test.sh involves running doctest on select python files. This is what jenkins runs, whereas developers typically run ./yb_build.sh. doctest appears to have a strange bug where it does not like it when sys.stdout is set to something else by the time of the end of file. This doesn't seem to matter when executing the python script itself: it only matters for doctest. PG 15 merge involves updating generate_unaccent_rules.py, and this update hits the above bug. Appease doctest by restoring sys.stdout to the original object at the end of the file. It seems wasteful to run doctest on files that don't have docstrings, but that can be fixed in the future. Test Plan: jenkins: compile only, rebase: pg15 Reviewers: mbautin, neil, steve.varnau Reviewed By: steve.varnau Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D23266 commit 56401ef Author: Steve Song <[email protected]> AuthorDate: Wed Nov 16 09:39:49 2022 +0000 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:23 2023 -0700 [MERGE 15] YSQL: Fix name conflict in aset realloc Summary: In the master branch, for memory tracking, we added `oldsize` for the block. It overwrites the variable with the same name in the outer scope, defined for the chunk. However, in PG upstream, the `oldsize` in the outer scope is referenced in the inner scope. Because of the naming overwritting, after merging, the reference used becomes incorrect. **Solution** I basically "backport" how the upstream creates the variable for "block old size" and use it is for YB memory tracking. Along with this, it also fixes two potential overflow issues in the existing code. The fixes are needed in the master as well. So a backport to the PG13 branch will be committed later in a separate diff. Original commit: 83c6b6a/D21128 Test Plan: Jenkins: rebase: pg15 Reviewers: neil, mbautin, telgersma Reviewed By: telgersma Subscribers: smishra, yql Differential Revision: https://phabricator.dev.yugabyte.com/D24000 commit 7e541a1 Author: neil <[email protected]> AuthorDate: Mon Apr 10 09:31:46 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:26 2023 -0700 Complete merging with Postgres 15.2 Fixed compilation errors in /access, /bootstrap, /catalog, and /commands Fixed compilation errors in /parser Fixed compilation errors in /commands Fixed compilation errors in /executor Fixed compilation errors in /libpq Fixed compilation errors in /optimizer Fixed compilation errors in /utils Completed fixing errors for directories in /backend/* Commented out Yugabyte's code for tools in src/bin - Revisit later Add missing files in contrib Add missing files in postgres/doc commit 06d9200 Author: neil <[email protected]> AuthorDate: Wed May 3 21:38:01 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:28 2023 -0700 Rebase with yb::master up to commit 61302b0 Fixed compilation errors in src/backend Fixed compilation errors in pg_dump Used Pg15 structure for FunctionCallInfo Remove Yugabyte version of pull_varattnos_walker() to avoid duplicating code with Pg15. Activate Yugabyte cluster-connecting code in InitPostgres commit 065c1a2 Author: Arpan Agrawal <[email protected]> AuthorDate: Mon May 1 23:59:08 2023 +0530 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:28 2023 -0700 [MERGE PG15] Adds missing oid columns and indexes to YB system tables Summary: Adds missing oid columns and indexes to YB system tables Test Plan: N/A Reviewers: mihnea, neil Reviewed By: neil Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D24959 commit 69fa310 Author: neil <[email protected]> AuthorDate: Sun May 7 12:37:33 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:28 2023 -0700 Rebase with yb::master up to commit 3d3480b commit 1cc244c Author: Arpan Agrawal <[email protected]> AuthorDate: Fri May 5 16:26:21 2023 +0530 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:28 2023 -0700 [MERGE PG15] Initdb: Bootstrap step working Summary: Initdb bootstrap step working Test Plan: Jenkins: skip Reviewers: mihnea, neil Reviewed By: neil Subscribers: yql Differential Revision: https://phabricator.dev.yugabyte.com/D25090 commit 57688f4 Author: neil <[email protected]> AuthorDate: Sat May 13 15:27:45 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:29 2023 -0700 Rebase with yb::master up to commit 9662afb commit 963f869 Author: neil <[email protected]> AuthorDate: Sat May 20 15:58:54 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:29 2023 -0700 Rebase with yb::master up to commit d5d857e commit b2ca34c Author: Arpan Agrawal <[email protected]> AuthorDate: Mon May 15 13:46:56 2023 +0530 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:29 2023 -0700 [MERGE PG15] Initdb fixes till setup_collation() Summary: Initdb fixes till setup_collation() Test Plan: Jenkins: skip Reviewers: neil, mihnea Reviewed By: neil Subscribers: yql Differential Revision: https://phorge.dev.yugabyte.com/D25341 commit fb32975 Author: Arpan Agrawal <[email protected]> AuthorDate: Thu Jun 1 22:08:41 2023 +0530 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:29 2023 -0700 [MERGE PG15] Initdb working end to end Summary: With this diff, initdb is working end to end except for pg_stat_statements. Test Plan: Jenkins: rebase: no Reviewers: neil, mihnea Reviewed By: neil Subscribers: yql Differential Revision: https://phorge.dev.yugabyte.com/D25787 commit eae7ba6 Author: neil <[email protected]> AuthorDate: Sun Jun 4 11:53:12 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:29 2023 -0700 Merge with yb::master up to commit 6d92aa6 commit 59f8df7 Author: neil <[email protected]> AuthorDate: Mon Jun 19 15:04:10 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 02:08:30 2023 -0700 Rebase with master upto commit dc4781a commit 50a962e Author: neil <[email protected]> AuthorDate: Thu Jul 20 10:24:57 2023 -0700 Commit: neil <[email protected]> CommitDate: Thu Jul 20 10:24:57 2023 -0700 Rebase with master upto commit 1c781b7
- Loading branch information
