DocumentationVersion comparison
POSTGRESQL · VERSION COMPARE
All features, fixes, and compatibility notes in this release, with related records from other versions.
Includes changes after the source version through the target. A major version name means its initial release.
From PostgreSQL 9.0 onward: 17 major branches and 352 release notes. Updated 2026-09-26.
Complete release changes
2019-11-14
12.1 HistoricalSupport ends 2024-11-21
A dump/restore is not required for those running 12.X.
Changes
Fix lock handling in REINDEX CONCURRENTLY (Michael Paquier) § §
REINDEX CONCURRENTLY neglected to take a session-level lock on the new index version, potentially allowing other sessions to manipulate it too soon. Also, a query-cancel or session-termination interrupt arriving at the wrong time could result in failure to release the session-level locks that REINDEX CONCURRENTLY does hold.
Changes
Avoid crash due to race condition when reporting the progress of a CREATE INDEX CONCURRENTLY or REINDEX CONCURRENTLY command (Álvaro Herrera) §
Changes
Avoid creating duplicate dependency entries during REINDEX CONCURRENTLY (Michael Paquier) §
This bug resulted in bloat in pg_depend, but no worse consequences than that.
Changes
Prevent VACUUM from trying to freeze an old multixact ID involving a still-running transaction (Nathan Bossart, Jeremy Schneider) §
This case would lead to VACUUM failing until the old transaction terminates.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix “wrong type of slot” error when trying to CLUSTER on an expression index (Andres Freund) §
Changes
SET CONSTRAINTS ... DEFERRED failed on partitioned tables, incorrectly complaining about lack of triggers (Álvaro Herrera) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix failure when creating indexes for a partition, if the parent partitioned table contains any dropped columns (Michael Paquier) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Ensure that a partition index can be dropped after a failure to reindex it concurrently (Michael Paquier) §
The index's pg_class.relispartition flag was left in the wrong state in such a case, causing DROP INDEX to fail.
Changes
Fix handling of equivalence class members for partition-wise joins (Amit Langote) §
This oversight could lead either to failure to use a feasible partition-wise join plan, or to a “could not find pathkey item to sort” planner failure.
Changes
Ensure that offset expressions in WINDOW clauses are processed when a query's expressions are manipulated (Andrew Gierth) §
This oversight could result in assorted failures when the offsets are nontrivial expressions. One example is that a function parameter reference in such an expression would fail if the function was inlined.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid postmaster failure if a parallel query requests a background worker when no postmaster child process array slots remain free (Tom Lane) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix crash triggered by an EvalPlanQual recheck on a table with a BEFORE UPDATE trigger (Andres Freund) §
Changes
Fix “unexpected relkind” error when a query tries to access a TOAST table (John Hsu, Michael Paquier, Tom Lane) §
The error should say that permission is denied, but this case got broken during code refactoring.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Provide a relevant error context line when an error occurs while setting GUC parameters during parallel worker startup (Thomas Munro) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Ensure that fsync() is applied only to files that are opened read/write (Andres Freund, Michael Paquier) § §
Some code paths tried to do this after opening a file read-only, but on some platforms that causes “bad file descriptor” or similar errors.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Allow encoding conversion to succeed on longer strings than before (Álvaro Herrera, Tom Lane) §
Previously, there was a hard limit of 0.25GB on the input string, but now it will work as long as the converted output is not over 1GB.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid creating unnecessarily-bulky tuple stores for window functions (Andrew Gierth) §
In some cases the tuple storage would include all columns of the source table(s), not just the ones that are needed by the query.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Allow repalloc() to give back space when a large chunk is reduced in size (Tom Lane) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Ensure that temporary WAL and history files are removed at the end of archive recovery (Sawada Masahiko) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid failure in archive recovery if recovery_min_apply_delay is enabled (Fujii Masao) §
recovery_min_apply_delay is not typically used in this configuration, but it should work.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Ignore restore_command, recovery_end_command, and recovery_min_apply_delay settings during crash recovery (Fujii Masao) § §
Now that these settings can be specified in postgresql.conf, they could be turned on during crash recovery, but honoring them then is undesirable. Ignore these settings until crash recovery is complete.
Changes
Fix logical replication failure when publisher and subscriber have different ideas about a table's replica identity columns (Jehan-Guillaume de Rorthais, Peter Eisentraut) § §
Declaring a column as part of the replica identity on the subscriber, when it does not exist at all on the publisher, led to “negative bitmapset member not allowed” errors.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid unwanted delay during shutdown of a logical replication walsender (Craig Ringer, Álvaro Herrera) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix timeout handling in logical replication walreceiver processes (Julien Rouhaud) §
Erroneous logic prevented wal_receiver_timeout from working in logical replication deployments.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Correctly time-stamp replication messages for logical decoding (Jeff Janes) §
This oversight resulted, for example, in pg_stat_subscription.last_msg_send_time usually reading as NULL.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix race condition during backend exit, when the backend process has previously waited for synchronous replication to occur (Dongming Liu) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid logging complaints about abandoned connections when using PAM authentication (Tom Lane) §
libpq-based clients will typically make two connection attempts when a password is required, since they don't prompt their user for a password until their first connection attempt fails. Therefore the server is coded not to generate useless log spam when a client closes the connection upon being asked for a password. However, the PAM authentication code hadn't gotten that memo, and would generate several messages about a phantom authentication failure.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix misbehavior of bitshiftright() (Tom Lane) § §
The bitstring right shift operator failed to zero out padding space that exists in the last byte of the result when the bitstring length is not a multiple of 8. While invisible to most operations, any nonzero bits there would result in unexpected comparison behavior, since bitstring comparisons don't bother to ignore the extra bits, expecting them to always be zero.
If you have inconsistent data as a result of saving the output of bitshiftright() in a table, it's possible to fix it with something like
UPDATE mytab SET bitcol = ~(~bitcol) WHERE bitcol != ~(~bitcol);
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix result of text position() function (also known as strpos()) for an empty search string (Tom Lane) §
Historically, and per the SQL standard, the result should be one in such cases, but 12.0 returned zero.
Changes
Fix detection of edge-case integer overflow in interval multiplication (Yuya Watari) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Avoid memory leak while vacuuming a GiST index (Dilip Kumar) §
Changes
On Windows, recognize additional spellings of the “Norwegian (Bokmål)” locale name (Tom Lane) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
In libpq, correctly report CONNECTION_BAD connection status after a failure caused by a syntactically invalid connect_timeout parameter value (Lars Kanis) §
Changes
Avoid compile failure if an ECPG client includes ecpglib.h while having ENABLE_NLS defined (Tom Lane) §
This risk was created by a misplaced declaration: ecpg_gettext() should not be visible to client code.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix scheduling of parallel restore of a foreign key constraint on a partitioned table (Álvaro Herrera) §
pg_dump failed to emit full dependency information for partitioned tables' foreign keys. This could allow parallel pg_restore to try to recreate a foreign key constraint too soon.
Changes
In pg_dump, ensure stable output order for similarly-named triggers and row-level-security policy objects (Benjie Gillam) §
Previously, if two triggers on different tables had the same names, they would be sorted in OID-based order, which is less desirable than sorting them by table name. Likewise for RLS policies.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Improve pg_upgrade's checks for the use of a data type that has changed representation, such as line (Tomas Vondra) § §
The previous coding could be fooled by cases where the data type of interest underlies a stored column of a domain or composite type.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
In pg_rewind with the --dry-run option, avoid updating pg_control (Alexey Kondratov) §
This could lead to failures in subsequent pg_rewind attempts.
Changes
Fix failure in pg_waldump with the -s option, when a continuation WAL record ends exactly at a page boundary (Andrey Lepikhov) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
In pg_waldump with the --bkp-details option, avoid emitting extra newlines for WAL records involving full-page writes (Andres Freund) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix small memory leak in pg_waldump (Andres Freund) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Put back pqsignal() as an exported libpq symbol (Tom Lane) §
This function was removed on the grounds that no clients should be using it, but that turns out to break usage of current libpq with very old versions of psql, and perhaps other applications.
Changes
Fix configure's test for presence of libperl so that it works on recent Red Hat releases (Tom Lane) § §
Previously, it could fail if the user sets CFLAGS to -O0.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Ensure correct code generation for spinlocks on PowerPC (Noah Misch) §
The previous spinlock coding allowed the compiler to select register zero for use with an assembly instruction that does not accept that register, causing a build failure. We have seen only one long-ago report that matches this bug, but it could cause problems for people trying to build modified PostgreSQL code or use atypical compiler options.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
On AIX, don't use the compiler option -qsrcmsg (Noah Misch) §
This avoids an internal compiler error with xlc v16.1.0, with little consequence other than changing the format of compiler error messages.
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Changes
Fix MSVC build process to cope with spaces in the file path of OpenSSL (Andrew Dunstan) §
“Same change” requires complete matching evidence. “Related commits” can cover independent changes, a partial backport, or a follow-up correction; each release keeps its own explanation.
Try another search, or .
The comparison follows PostgreSQL release notes from just after the source through the target version. For a major upgrade, maintenance releases from each older branch are included only up to the next major release date, and never after the target date. A major version such as 18 means its initial release, 18.0. Previews and development snapshots are labeled separately.
Entries come from the original English manuals. Release coverage and commit evidence are verified against upstream sources. Every entry retains its complete explanation and source link. Categories aid browsing; read the full notes for impact, conditions, and migration steps.
Fixes can be backported to several branches. Confirmed duplicates are merged conservatively, with every branch explanation retained. A note describing several independent fixes is excluded only when all are already present in the source. Major-release features remain distinct from related maintenance patches unless their complete original descriptions match. Uncertain matches are retained. This is a release-note history, not an exhaustive comparison of compiled binaries.
CVE results are calculated independently from the PostgreSQL security registry and vulnerability records. A CVE counts as gained protection only when the source is affected and the target is fixed or unaffected. Remaining vulnerabilities are listed separately. Security entries and distinct CVEs are counted separately.
Interaction inspired by pgversions.com and pgversionreport. Content comes from PostgreSQL release notes. See the release notes archive.