Fix inconsistent comments for some function declarations in headers
Some of the headers list a couple of function prototypes located in a
different file than what is referred to. This fixes a couple of
places where this inconsistency exists.
Author: Richard Guo
Discussion:
https://postgr.es/m/c
Clarify DROP EXTENSION docs regarding explicitly dependent routines
Per suggestion from Robert Haas
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/CA%2BTgmoZ1QvHquYHLkMy1oHKqz4-E7QQctj6e0ocq_GP1B5%2B9bA%40mail.gmail.com
Branch
--
master
Details
---
https://git.post
Clarify DROP EXTENSION docs regarding explicitly dependent routines
Per suggestion from Robert Haas
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/CA%2BTgmoZ1QvHquYHLkMy1oHKqz4-E7QQctj6e0ocq_GP1B5%2B9bA%40mail.gmail.com
Branch
--
REL_15_STABLE
Details
---
https://g
Clarify DROP EXTENSION docs regarding explicitly dependent routines
Per suggestion from Robert Haas
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/CA%2BTgmoZ1QvHquYHLkMy1oHKqz4-E7QQctj6e0ocq_GP1B5%2B9bA%40mail.gmail.com
Branch
--
REL_14_STABLE
Details
---
https://g
Fix assorted doc typos
Erik Rijkers and Justin Pryzby
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/b79bfeff-d0e3-29a3-2576-0e325848dede%40xs4all.nl
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/9f08803828d7dc9691e773f5bec4eed21f5ff520
Mod
Fix assorted doc typos
Erik Rijkers and Justin Pryzby
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/b79bfeff-d0e3-29a3-2576-0e325848dede%40xs4all.nl
Branch
--
REL_15_STABLE
Details
---
https://git.postgresql.org/pg/commitdiff/b25566ab3c6536040df4df3ad99dd69c812215
Fix assorted doc typos
Erik Rijkers and Justin Pryzby
Backpatch to v14
Discussion:
https://www.postgresql.org/message-id/b79bfeff-d0e3-29a3-2576-0e325848dede%40xs4all.nl
Branch
--
REL_14_STABLE
Details
---
https://git.postgresql.org/pg/commitdiff/120e159b7d93fa36b94ded8a51fe4e619e1261
Fix formatting and comment typos
Justin Pryzby
Discussion:
https://www.postgresql.org/message-id/20220801181136.GJ15006%40telsasoft.com
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/bcabbfc6a94bb1d59e380f1bcb7cd5da8338cd11
Modified Files
--
src/back
Rephrase comments to make them clearer
The use of "we" when referring to the active backend might be
misunderstood, so rephrase to make it clearer who is performing
the actions discussed in the comment.
Author: Junwang Zhao
Reviewed-by: Erikjan Rijkers
Reviewed-by: Robert Treat
Reviewed-by: To
Add proper regression test for the recent SRFs-in-pathkeys problem.
Remove the test case added by commit fac1b470a, which never actually
worked to expose the problem it claimed to test. Replace it with
a case that does expose the problem, and also covers the SRF-not-
at-the-top deficiency repaire
Add proper regression test for the recent SRFs-in-pathkeys problem.
Remove the test case added by commit fac1b470a, which never actually
worked to expose the problem it claimed to test. Replace it with
a case that does expose the problem, and also covers the SRF-not-
at-the-top deficiency repaire
Add proper regression test for the recent SRFs-in-pathkeys problem.
Remove the test case added by commit fac1b470a, which never actually
worked to expose the problem it claimed to test. Replace it with
a case that does expose the problem, and also covers the SRF-not-
at-the-top deficiency repaire
Add proper regression test for the recent SRFs-in-pathkeys problem.
Remove the test case added by commit fac1b470a, which never actually
worked to expose the problem it claimed to test. Replace it with
a case that does expose the problem, and also covers the SRF-not-
at-the-top deficiency repaire
Fix check_exclusion_or_unique_constraint for UNIQUE NULLS NOT DISTINCT.
Adjusting this function was overlooked in commit 94aa7cc5f. The only
visible symptom (so far) is that INSERT ... ON CONFLICT could go into
an endless loop when inserting a null that has a conflict.
Richard Guo and Tom Lane,
Fix check_exclusion_or_unique_constraint for UNIQUE NULLS NOT DISTINCT.
Adjusting this function was overlooked in commit 94aa7cc5f. The only
visible symptom (so far) is that INSERT ... ON CONFLICT could go into
an endless loop when inserting a null that has a conflict.
Richard Guo and Tom Lane,
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Add CHECK_FOR_INTERRUPTS in ExecInsert's speculative insertion loop.
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it
Revert recent changes to 002_pg_upgrade.pl.
The test is proving to be unreliable in the buildfarm, and we neither
agree on how best to fix it nor have time to do so before the upcoming
release. So for now, put things back to the way they were before commit
d498e052b4b84ae21b3b68d5b3fda6ead65d1d4d.
Revert recent changes to 002_pg_upgrade.pl.
The test is proving to be unreliable in the buildfarm, and we neither
agree on how best to fix it nor have time to do so before the upcoming
release. So for now, put things back to the way they were before commit
d498e052b4b84ae21b3b68d5b3fda6ead65d1d4d.
Remove configure probe for link.
link() is in SUSv2 and all targeted Unix systems have it. We have
replacement code for Windows that doesn't require a configure probe.
Since only Windows needs it, rename src/port/link.c to win32link.c like
other similar things.
There is no need for a vestigial H
Remove dead setenv, unsetenv replacement code.
setenv() and unsetenv() are in SUSv3 and targeted Unix systems have
them. We still need special code for these on Windows, but that doesn't
require a configure probe.
This marks the first time we require a SUSv3 (POSIX.1-2001) facility
(rather than
Remove configure probes for poll and poll.h.
poll() and are in SUSv2 and all targeted Unix systems have
them.
Retain HAVE_POLL and HAVE_POLL_H macros for readability. There's an
error in latch.c that is now unreachable (since we always have one of
WIN32 or HAVE_POLL defined), but that falls out
Remove configure probe for dlopen, and refactor.
dlopen() is in SUSv2 and all targeted Unix systems have it. We still
need replacement functions for Windows, but we don't need a configure
probe for that.
Since it's no longer needed by other operating systems, rename dlopen.c
to win32dlopen.c and
Remove configure probe and related tests for getrlimit.
getrlimit() is in SUSv2 and all targeted systems have it.
Windows doesn't have it. We could just use #ifndef WIN32, but for a
little more explanation about why we're making things conditional, let's
retain the HAVE_GETRLIMIT macro. It's de
Remove configure probe for setsid.
setsid() is in SUSv2 and all targeted Unix systems have it. Retain a
HAVE_SETSID macro, defined on Unix only. That's easier to understand
than !defined(WIN32), for the optional code it governs.
Reviewed-by: Tom Lane
Reviewed-by: Andres Freund
Discussion:
ht
Remove dead pread and pwrite replacement code.
pread() and pwrite() are in SUSv2, and all targeted Unix systems have
them.
Previously, we defined pg_pread and pg_pwrite to emulate these function
with lseek() on old Unixen. The names with a pg_ prefix were a reminder
of a portability hazard: they
Remove configure probe for shm_open.
shm_open() is in SUSv2 and all targeted Unix systems have it.
We retain a HAVE_SHM_OPEN macro, because it's clearer to readers than
something like !defined(WIN32).
Reviewed-by: Tom Lane
Reviewed-by: Andres Freund
Discussion:
https://postgr.es/m/CA+hUKGJ3LH
Remove configure probes for symlink/readlink, and dead code.
symlink() and readlink() are in SUSv2 and all targeted Unix systems have
them. We have partial emulation on Windows. Code that raised runtime
errors on systems without it has been dead for years, so we can remove
that and also referenc
Fix failure to set correct operator in window run condition
This was a simple omission in 9d9c02ccd where the code didn't correctly
set the operator to use in the run condition OpExpr when the window
function was both monotonically increasing and decreasing.
Bug discovered by Julien Roze, althoug
Fix failure to set correct operator in window run condition
This was a simple omission in 9d9c02ccd where the code didn't correctly
set the operator to use in the run condition OpExpr when the window
function was both monotonically increasing and decreasing.
Bug discovered by Julien Roze, althoug
Use hba_file/ident_file GUCs rather than pg_hba.conf/pg_ident.conf in logs
This is particularly useful when log_min_messages is set to FATAL, so as
one can know which file was not getting loaded whether hba_file or
ident_file are set to some non-default values. If using the default
values of thes
Make consistent a couple of log messages when parsing HBA files
This commit adjusts two log messages:
- When a field in pg_ident.conf is not populated, report the line of the
configuration file in an error context message instead of the main
entry.
- When parsing pg_ident.conf and finding an inval
Simplify replacement code for preadv and pwritev.
preadv() and pwritev() are not standardized by POSIX, but appeared in
NetBSD in 1999 and were adopted by at least OpenBSD, FreeBSD,
DragonFlyBSD, Linux, AIX, illumos and macOS. We don't use them much
yet, but an active proposal uses them heavily.
Fix nbtree maximum item size macro.
Commit dd299df8189, which made heap TID a tiebreaker nbtree index
column, introduced new rules on page space management to make suffix
truncation safe for v4+ indexes. New pivot tuples (generated by suffix
truncation during leaf page splits) sometimes require d
Remove configure probe for clock_gettime.
clock_gettime() is in SUSv2 and all targeted Unix systems have it.
Remove a chunk of fallback code for old Unix is no longer reachable on
modern systems, and untested as of the retirement of build farm animal
prairiedog.
There is no need to retain a HAVE_
Remove configure probe for fdatasync.
fdatasync() is in SUSv2, and all targeted Unix systems have it. We have
a replacement function for Windows.
We retain the probe for the function declaration, which allows us to
supply the mysteriously missing declaration for macOS, and also for
Windows. No
41 matches
Mail list logo
