On Mon, Sep 22, 2025 at 10:23:45AM +0200, Daniel Gustafsson wrote:
> I am so sorry, I realized my mistake the second I hit enter, I meant to just
> stage it and accidentally instead did all branches =( Reverted immediately. I
> am going to go stand in the corner.
Let's share a beer in the corner
for other purposes than only tests.
The coverage increases from 85.2% to 93.4%. It should be possible to
cover more code paths, but at least it's a beginning.
Author: Greg Burd
Reviewed-by: Nathan Bossart
Reviewed-by: Michael Paquier
Discussion:
https://postgr.es/m/7bd1abdb-b03a-464a
On Sun, Sep 21, 2025 at 07:57:05PM -0400, Tom Lane wrote:
> Michael Paquier writes:
>> Argh. Sorry about that! Would you prefer if I revert immediately on
>> ~18?
>
> Yeah, reverting on 18 and leaving the other branches alone
> would be fine with me. We can put
Revert "Fix meson build with -Duuid=ossp when using version older than 0.60"
This reverts commit 5f565b0aee90 temporarily on v18. This branch is in
a release freeze state until tagged. Let's re-add this commit once the
release is out. The other branches are left untouched.
Discussion: https://
Fix meson build with -Duuid=ossp when using version older than 0.60
The package for the UUID library may be named "uuid" or "ossp-uuid", and
meson.build has been using a single call of dependency() with multiple
names, something only supported since meson 0.60.0.
The minimum version of meson supp
On Sun, Sep 21, 2025 at 07:49:44PM -0400, Tom Lane wrote:
> Michael Paquier writes:
> > Fix meson build with -Duuid=ossp when using version older than 0.60
>
> Um ... we are in release freeze for 18.0. Not sure that this
> is something that justifies breaking freeze.
Argh.
Fix meson build with -Duuid=ossp when using version older than 0.60
The package for the UUID library may be named "uuid" or "ossp-uuid", and
meson.build has been using a single call of dependency() with multiple
names, something only supported since meson 0.60.0.
The minimum version of meson supp
Fix meson build with -Duuid=ossp when using version older than 0.60
The package for the UUID library may be named "uuid" or "ossp-uuid", and
meson.build has been using a single call of dependency() with multiple
names, something only supported since meson 0.60.0.
The minimum version of meson supp
Fix meson build with -Duuid=ossp when using version older than 0.60
The package for the UUID library may be named "uuid" or "ossp-uuid", and
meson.build has been using a single call of dependency() with multiple
names, something only supported since meson 0.60.0.
The minimum version of meson supp
Fix incorrect file reference in guc.h
GucSource_Names was documented as being in guc.c, but since 0a20ff54f5e6
it is located in guc_tables.c. The reference to the location of
GucSource_Names is important, as GucSource needs to be kept in sync with
GucSource_Names.
Author: David G. Johnston
Disc
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
: Chao Li
Reviewed-by: Michael Paquier
Discussion:
https://postgr.es/m/caakru_a1vzrzrwo3_jv_x13ryoqlrvipgo0237g5pkzpa2y...@mail.gmail.com
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/deb208df4559c922df217354f2c0cb689016a2ed
Modified Files
--
src
f the hash key is already 8 bytes, which should be
plenty enough already. A comment is added to discourage the addition of
new fields.
Author: Michael Paquier
Reviewed-by: Sami Imseih
Discussion:
https://postgr.es/m/caa5rz0t9omat+hvsakjxwtmwvhpyfcazb41rpwkwrkfugma...@mail.gmail.com
B
On Thu, Sep 18, 2025 at 09:56:44AM -0500, Nathan Bossart wrote:
> Thanks for reviewing. Committed.
Thanks. batta looks happy now.
--
Michael
signature.asc
Description: PGP signature
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
On Wed, Sep 17, 2025 at 09:34:24AM -0500, Nathan Bossart wrote:
> Yeah, this stuff is rather precarious, and I'm a bit surprised there
> haven't been more problems in this area.
LWLock handling is quite special, but I am also getting suspicious
about some more shmem areas. Perhaps these deserve a
injection_points: Fix incrementation of variable-numbered stats
The pending entry was not used when incrementing its data, directly
manipulating the shared memory pointer, without even locking it. This
could mean losing statistics under concurrent activity. The flush
callback was a no-op.
This
On Tue, Sep 16, 2025 at 11:10:34PM -0500, Nathan Bossart wrote:
> I was able to simplify the patch.
+ if (LocalNamedLWLockTrancheRequestArray)
+ NamedLWLockTrancheRequestArray = LocalNamedLWLockTrancheRequestArray;
It's not a common practice in the code to rely on a shmem state that
shoul
Ni Nathan,
On Thu, Sep 11, 2025 at 09:15:12PM +, Nathan Bossart wrote:
> Move named LWLock tranche requests to shared memory.
>
> In EXEC_BACKEND builds, GetNamedLWLockTranche() can segfault when
> called outside of the postmaster process, as it might access
> NamedLWLockTrancheRequestArray,
injection_points: Fix incrementation of variable-numbered stats
The pending entry was not used when incrementing its data, directly
manipulating the shared memory pointer, without even locking it. This
could mean losing statistics under concurrent activity. The flush
callback was a no-op.
This
Fix shared memory calculation size of PgAioCtl
The shared memory size was calculated based on an offset of io_handles,
which is itself a pointer included in the structure. We tend to
overestimate the shared memory size overall, so this was unlikely an
issue in practice, but let's be correct and u
Fix shared memory calculation size of PgAioCtl
The shared memory size was calculated based on an offset of io_handles,
which is itself a pointer included in the structure. We tend to
overestimate the shared memory size overall, so this was unlikely an
issue in practice, but let's be correct and u
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
Fix leak with SMgrRelations in startup process
The startup process does not process shared invalidation messages, only
sending them, and never calls AtEOXact_SMgr() which clean up any
unpinned SMgrRelations. Hence, it is never able to free SMgrRelations
on a periodic basis, bloating its hashtable
Fix corruption of pgstats shared hashtable due to OOM failures
A new pgstats entry is created as a two-step process:
- The entry is looked at in the shared hashtable of pgstats, and is
inserted if not found.
- When not found and inserted, its fields are then initialized. This
part include a DSA c
Remove whitespace in comment of pg_stat_statements.c
Introduced by 6b4d23feef6e, spotted while reading this area of the code.
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/306dd13079ed616c414c9411c5deadffea273266
Modified Files
--
contrib/pg_stat_stat
Add more information for WAL records of hash index AMs
hashdesc.c was missing a couple of fields in its record descriptions, as
of:
- is_prev_bucket_same_wrt for SQUEEZE_PAGE.
- procid for INIT_META_PAGE.
- old_bucket_flag and new_bucket_flag for SPLIT_ALLOCATE_PAGE.
The author has noted the firs
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
Fix description of WAL record blocks in hash_xlog.h
hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE
Fix incorrect file reference in guc.h
GucSource_Names was documented as being in guc.c, but since 0a20ff54f5e6
it is located in guc_tables.c. The reference to the location of
GucSource_Names is important, as GucSource needs to be kept in sync with
GucSource_Names.
Author: David G. Johnston
Disc
Fix incorrect file reference in guc.h
GucSource_Names was documented as being in guc.c, but since 0a20ff54f5e6
it is located in guc_tables.c. The reference to the location of
GucSource_Names is important, as GucSource needs to be kept in sync with
GucSource_Names.
Author: David G. Johnston
Disc
Fix incorrect file reference in guc.h
GucSource_Names was documented as being in guc.c, but since 0a20ff54f5e6
it is located in guc_tables.c. The reference to the location of
GucSource_Names is important, as GucSource needs to be kept in sync with
GucSource_Names.
Author: David G. Johnston
Disc
Remove dynahash.h
All the callers of my_log2() are now limited inside dynahash.c, so let's
remove this header. The same capability is provided by pg_bitutils.h
already.
Discussion:
https://postgr.es/m/CAEZATCUJPQD_7sC-wErak2CQGNa6bj2hY-mr8wsBki=kx7f...@mail.gmail.com
Branch
--
master
Deta
Replace callers of dynahash.h's my_log() by equivalent in pg_bitutils.h
All the calls replaced by this commit use 4-byte integers for their
variables used in input of my_log2(). Hence, the limit against
too-large inputs does not really apply. Thresholds are also applied, as
of:
- In nodeAgg.c, t
Fix leak with SMgrRelations in startup process
The startup process does not process shared invalidation messages, only
sending them, and never calls AtEOXact_SMgr() which clean up any
unpinned SMgrRelations. Hence, it is never able to free SMgrRelations
on a periodic basis, bloating its hashtable
Fix leak with SMgrRelations in startup process
The startup process does not process shared invalidation messages, only
sending them, and never calls AtEOXact_SMgr() which clean up any
unpinned SMgrRelations. Hence, it is never able to free SMgrRelations
on a periodic basis, bloating its hashtable
Fix corruption of pgstats shared hashtable due to OOM failures
A new pgstats entry is created as a two-step process:
- The entry is looked at in the shared hashtable of pgstats, and is
inserted if not found.
- When not found and inserted, its fields are then initialized. This
part include a DSA c
Fix corruption of pgstats shared hashtable due to OOM failures
A new pgstats entry is created as a two-step process:
- The entry is looked at in the shared hashtable of pgstats, and is
inserted if not found.
- When not found and inserted, its fields are then initialized. This
part include a DSA c
Fix corruption of pgstats shared hashtable due to OOM failures
A new pgstats entry is created as a two-step process:
- The entry is looked at in the shared hashtable of pgstats, and is
inserted if not found.
- When not found and inserted, its fields are then initialized. This
part include a DSA c
Fix corruption of pgstats shared hashtable due to OOM failures
A new pgstats entry is created as a two-step process:
- The entry is looked at in the shared hashtable of pgstats, and is
inserted if not found.
- When not found and inserted, its fields are then initialized. This
part include a DSA c
Update parser README to include parse_jsontable.c
The README was missing parse_jsontable.c which handles JSON_TABLE.
Oversight in de3600452b61.
Author: Karthik S
Discussion:
https://postgr.es/m/CAK4gQD9gdcj+vq_FZGp=Rv-W+41v8_C7cmCUmDeu=cfrodf...@mail.gmail.com
Backpatch-through: 17
Branch
Update parser README to include parse_jsontable.c
The README was missing parse_jsontable.c which handles JSON_TABLE.
Oversight in de3600452b61.
Author: Karthik S
Discussion:
https://postgr.es/m/CAK4gQD9gdcj+vq_FZGp=Rv-W+41v8_C7cmCUmDeu=cfrodf...@mail.gmail.com
Backpatch-through: 17
Branch
Update parser README to include parse_jsontable.c
The README was missing parse_jsontable.c which handles JSON_TABLE.
Oversight in de3600452b61.
Author: Karthik S
Discussion:
https://postgr.es/m/CAK4gQD9gdcj+vq_FZGp=Rv-W+41v8_C7cmCUmDeu=cfrodf...@mail.gmail.com
Backpatch-through: 17
Branch
Fix outdated comments in slru.c
SlruRecentlyUsed() is an inline function since 53c2a97a9266, not a
macro. The description of long_segment_names was missing at the top of
SimpleLruInit(), part forgotten in 4ed8f0913bfd.
Author: Julien Rouhaud
Discussion: https://postgr.es/m/aLpBLMOYwEQkaleF@jrou
On Fri, Sep 05, 2025 at 10:52:55PM +, Tatsuo Ishii wrote:
> Allow to log raw parse tree.
>
> This commit allows to log the raw parse tree in the same way we
> currently log the parse tree, rewritten tree, and plan tree.
>
> To avoid unnecessary log noise for users not interested in this
> det
Fix outdated comments in slru.c
SlruRecentlyUsed() is an inline function since 53c2a97a9266, not a
macro. The description of long_segment_names was missing at the top of
SimpleLruInit(), part forgotten in 4ed8f0913bfd.
Author: Julien Rouhaud
Discussion: https://postgr.es/m/aLpBLMOYwEQkaleF@jrou
Fix outdated comments in slru.c
SlruRecentlyUsed() is an inline function since 53c2a97a9266, not a
macro. The description of long_segment_names was missing at the top of
SimpleLruInit(), part forgotten in 4ed8f0913bfd.
Author: Julien Rouhaud
Discussion: https://postgr.es/m/aLpBLMOYwEQkaleF@jrou
Switch some numeric-related functions to use soft error reporting
This commit changes some functions related to the data type numeric to
use the soft error reporting rather than a custom boolean flag (called
"have_error") that callers of these functions could rely on to bypass
the generation of ER
Change pg_lsn_in_internal() to use soft error reporting
pg_lsn includes pg_lsn_in_internal() for the purpose of parsing a LSN
position for the GUC recovery_target_lsn (21f428ebde39). It relies on a
boolean called "have_error" that would be set when the LSN parsing
fails, then let its callers hand
Fix incorrect comment in pgstat_backend.c
The counters saved from pgWalUsage, used for the difference calculations
when flushing the backend WAL stats, are updated when calling
pgstat_flush_backend() under PGSTAT_BACKEND_FLUSH_WAL, and not
pgstat_report_wal(). The comment updated in this commit r
Fix incorrect comment in pgstat_backend.c
The counters saved from pgWalUsage, used for the difference calculations
when flushing the backend WAL stats, are updated when calling
pgstat_flush_backend() under PGSTAT_BACKEND_FLUSH_WAL, and not
pgstat_report_wal(). The comment updated in this commit r
Update outdated references to the SLRU ControlLock
SLRU bank locks are referred as "bank locks" or "SLRU bank locks" in the
code comments. The comments updated in this commit use the latter term.
Oversight in 53c2a97a9266, that has replaced the single ControlLock by
the bank control locks.
Auth
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
libpq: Fix PQtrace() format for non-printable characters
PQtrace() was generating its output for non-printable characters without
casting the characters printed with unsigned char, leading to some extra
"\xff" generated in the output due to the fact that char may be
signed.
Oversights introdu
Update outdated references to the SLRU ControlLock
SLRU bank locks are referred as "bank locks" or "SLRU bank locks" in the
code comments. The comments updated in this commit use the latter term.
Oversight in 53c2a97a9266, that has replaced the single ControlLock by
the bank control locks.
Auth
Update outdated references to the SLRU ControlLock
SLRU bank locks are referred as "bank locks" or "SLRU bank locks" in the
code comments. The comments updated in this commit use the latter term.
Oversight in 53c2a97a9266, that has replaced the single ControlLock by
the bank control locks.
Auth
Generate pgstat_count_slru*() functions for slru using macros
This change replaces seven functions definitions by macros, reducing a
bit some repetitive patterns in the code. An interesting side effect is
that this removes an inconsistency in the naming of SLRU increment
functions with the field
postgres_fdw: Use psql variables for connection parameters
Several statements need to reference the current connection's current
database name and current port value. Until now, this has been
accomplished by creating dynamic SQL statements inside of a DO block,
which is not as easy to parse. It
Change dynahash.c and hsearch.h to use int64 instead of long
This code was relying on "long", which is signed 8 bytes everywhere
except on Windows where it is 4 bytes, that could potentially expose it
to overflows, even if the current uses in the code are fine as far as I
know. This code is now a
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Reviewed-By: Ashutosh Bapat
Reviewed-By: Robert Haas
Reviewed-By: Kyotaro Horiguchi
Reviewed-By: Takamichi Osumi
Reviewed-By: Michael Paquier
Reviewed-By: Masahiko Sawada
Reported-By: Shenhao Wang
Discussion: https://postgr.es/m/bbaaf9f9-ebb2-645f-54bb-34d6efc7a...@fujitsu.com
Backpatch
Apply some fat commas to commands of TAP tests
This is similar to 19c6e92b13b2, in order to keep the style used in the
scripts consistent for the option names and values used in commands.
The places updated in this commit have been added recently in
71ea0d679543.
These changes are cosmetic; there
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
doc: Improve description of wal_compression
The description of this GUC provides a list of the situations where
full-page writes are generated. However, it is not completely exact,
mentioning only the cases where full_page_writes=on or base backups. It
is possible to generate full-page writes in
On Tue, Aug 19, 2025 at 12:31:14PM +0200, Alvaro Herrera wrote:
> Hmm, don't you find strange that dbcommands.h is still included there?
> I wondered about that and found out that we have get_database_name()
> declared there. It sort-of makes sense, because it was originally done
> by scanning pg_
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
Fix assertion failure with replication slot release in single-user mode
Some replication slot manipulations (logical decoding via SQL,
advancing) were failing an assertion when releasing a slot in
single-user mode, because active_pid was not set in a ReplicationSlot
when its slot is acquired.
Rep
added to multixact.h, becoming available for
out-of-core code.
Extracted from a larger patch by the same author.
Author: Naga Appani
Reviewed-by: Ashutosh Bapat
Reviewed-by: Michael Paquier
Discussion:
https://postgr.es/m/ca+qey+aasyk6wvbw4qyzhz4bahhycday_q5ecmhkev_eb9c...@mail.gmail.com
Remove useless pointer update in StatsShmemInit()
This pointer was not used after its last update. This variable
assignment was most likely a vestige artifact of the earlier versions of
the patch set that have led to 5891c7a8ed8f.
This pointer update is useless, so let's remove it. It removes o
test_ddl_deparse: Rename test create_sequence_1 to create_sequence
This test was the only one named following the convention used for
alternate output files. This was a little bit confusing when looking at
the diffs of the test, because one would think that the diffs are based
on an uncommon case
Remove unneeded header declarations in multixact.c
Two header declarations were related to SQL-callable functions, that
should have been cleaned up in df9133fa6384. Some more includes can be
removed on closer inspection, so let's clean up these as well, while on
it.
Reported-by: Tom Lane
Discus
On Mon, Aug 18, 2025 at 09:47:14AM -0400, Tom Lane wrote:
> Michael Paquier writes:
> > Move SQL-callable code related to multixacts into its own file
>
> Couldn't this have removed some #include-s from multixact.c?
Right. funcapi.h and fmgrprotos.h are direct dependencies
mxstatus_to_string() to multixact.h.
Extracted from a larger patch by the same author, tweaked by me.
Author: Naga Appani
Reviewed-by: Michael Paquier
Reviewed-by: Ashutosh Bapat
Discussion:
https://postgr.es/m/ca+qey+aasyk6wvbw4qyzhz4bahhycday_q5ecmhkev_eb9c...@mail.gmail.com
Branch
s relates to WAL logging. We still need to
depend on FormData_pg_sequence.
Author: Michael Paquier
Reviewed-by: Kirill Reshke
Reviewed-by: Tomas Vondra
Discussion: https://postgr.es/m/zwlohtkas0uvv...@paquier.xyz
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/
Remove md5() call from isolation test for CLUSTER and TOAST
This test was failing because MD5 computations are not supported in
these environments. This switches the test to rely on sha256() instead,
providing the same coverage while avoiding the failure.
Oversight in f57e214d1cbb. Per buildfar
On Sun, Aug 17, 2025 at 06:27:57AM +, Michael Paquier wrote:
> Add isolation test for TOAST value reuse during CLUSTER
A couple of buildfarm members have been complaining about this test
using md5(). Will fix in a bit.
--
Michael
signature.asc
Description: PGP signature
Add isolation test for TOAST value reuse during CLUSTER
This test exercises the corner case in toast_save_datum() where CLUSTER
operations encounter duplicated TOAST references, reusing the existing
TOAST data instead of creating redundant copies.
During table rewrites like CLUSTER, both live and
.
Author: Nikhil Kumar Veldanda
Co-authored-by: Michael Paquier
Discussion:
https://postgr.es/m/cafafj_e+kw5p713s8_jzyvgqagvffzfitujprgo-tttjjoa...@mail.gmail.com
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/69f75d6714753e594fb383d2e53a8003a8b54dfd
Modified
Fix compilation warning with SerializeClientConnectionInfo()
This function uses an argument named "maxsize" that is only used in
assertions, being set once outside the assertion area. Recent gcc
versions with -Wunused-but-set-parameter complain about a warning when
building without assertions ena
1 - 100 of 3609 matches
Mail list logo
