Tom Lane [Fri, 28 Sep 2018 18:28:19 +0000 (14:28 -0400)]
Build src/common files as a library with -fPIC.
Build a third version of libpgcommon.a, with -fPIC and -DFRONTEND,
as commit ea53100d5 did for src/port. Use that in libpq to avoid
symlinking+rebuilding source files retail.
Also adjust ecpg to use the new src/port and src/common libraries.
Arrange to install these libraries, too, to simplify out-of-tree
builds of shared libraries that need any of these modules.
Tom Lane [Fri, 28 Sep 2018 16:38:10 +0000 (12:38 -0400)]
Remove pqsignal() from libpq's official exports list.
Client applications should get this function, if they need it, from
libpgport.
The fact that it's exported from libpq is a hack left over from before
we set up libpgport. It's never been documented, and there's no good
reason for non-PG code to be calling it anyway, so hopefully this won't
cause any problems. Moreover, with the previous setup it was not real
clear whether our clients that use the function were getting it from
libpgport or libpq, so this might actually prevent problems.
The reason for changing it now is that in the wake of commit ea53100d5,
some linkers won't export the symbol, apparently because it's coming from
a .a library instead of a .o file. We could get around that by continuing
to symlink pqsignal.c into libpq as before; but unless somebody complains
very hard, I don't want to adopt such a kluge.
Amit Kapila [Fri, 28 Sep 2018 06:20:17 +0000 (11:50 +0530)]
Fix assertion failure when updating full_page_writes for checkpointer.
When the checkpointer receives a SIGHUP signal to update its configuration,
it may need to update the shared memory for full_page_writes and need to
write a WAL record for it. Now, it is quite possible that the XLOG
machinery has not been initialized by that time and it will lead to
assertion failure while doing that. Fix is to allow the initialization of
the XLOG machinery outside critical section.
This bug has been introduced by the commit 2c03216d83 which added the XLOG
machinery initialization in RecoveryInProgress code path.
Reported-by: Dilip Kumar
Author: Dilip Kumar Reviewed-by: Michael Paquier and Amit Kapila
Backpatch-through: 9.5
Discussion: https://postgr.es/m/CAFiTN-u4BA8KXcQUWDPNgaKAjDXC=C2whnzBM8TAcv=stckYUw@mail.gmail.com
Michael Paquier [Fri, 28 Sep 2018 02:54:38 +0000 (11:54 +0900)]
Fix WAL recycling on standbys depending on archive_mode
A restart point or a checkpoint recycling WAL segments treats segments
marked with neither ".done" (archiving is done) or ".ready" (segment is
ready to be archived) in archive_status the same way for archive_mode
being "on" or "always". While for a primary this is fine, a standby
running a restart point with archive_mode = on would try to mark such a
segment as ready for archiving, which is something that will never
happen except after the standby is promoted.
Note that this problem applies only to WAL segments coming from the
local pg_wal the first time archive recovery is run. Segments part of a
self-contained base backup are the most common case where this could
happen, however even in this case normally the .done markers would be
most likely part of the backup. Segments recovered from an archive are
marked as .ready or .done by the startup process, and segments finished
streaming are marked as such by the WAL receiver, so they are handled
already.
Reported-by: Haruka Takatsuka
Author: Michael Paquier
Discussion: https://postgr.es/m/15402-a453c90ed4cf88b2@postgresql.org
Backpatch-through: 9.5, where archive_mode = always has been added.
Tom Lane [Thu, 27 Sep 2018 22:15:06 +0000 (18:15 -0400)]
Fix assorted bugs in pg_get_partition_constraintdef().
It failed if passed a nonexistent relation OID, or one that was a non-heap
relation, because of blindly applying heap_open to a user-supplied OID.
This is not OK behavior for a SQL-exposed function; we have a project
policy that we should return NULL in such cases. Moreover, since
pg_get_partition_constraintdef ought now to work on indexes, restricting
it to heaps is flat wrong anyway.
The underlying function generate_partition_qual() wasn't on board with
indexes having partition quals either, nor for that matter with rels
having relispartition set but yet null relpartbound. (One wonders
whether the person who wrote the function comment blocks claiming that
these functions allow a missing relpartbound had ever tested it.)
Fix by testing relispartition before opening the rel, and by using
relation_open not heap_open. (If any other relkinds ever grow the
ability to have relispartition set, the code will work with them
automatically.) Also, don't reject null relpartbound in
generate_partition_qual.
Back-patch to v11, and all but the null-relpartbound change to v10.
(It's not really necessary to change generate_partition_qual at all
in v10, but I thought s/heap_open/relation_open/ would be a good
idea anyway just to keep the code in sync with later branches.)
Andres Freund [Thu, 27 Sep 2018 18:24:49 +0000 (11:24 -0700)]
Clean up in the wake of TupleDescGetSlot() removal / 10763358c3f.
The previous commit wasn't careful enough to remove all traces of
TupleDescGetSlot().
Besides fixing the oversight of not removing TupleDescGetSlot()'s
declaration, this also removes FuncCallContext->slot. That was
documented to be for use in combination with TupleDescGetSlot(), a
cursory search over extensions finds no users, and there doesn't seem
to be convincing reasons to keep it around. If we later in the v12
release cycle find users, we can re-consider this part of the commit.
Reported-By: Michael Paquier
Discussion: https://postgr.es/m/20180926000413.GC1659@paquier.xyz
Tom Lane [Thu, 27 Sep 2018 15:23:43 +0000 (11:23 -0400)]
Build src/port files as a library with -fPIC, and use that in libpq.
libpq and ecpg need shared-library-friendly versions of assorted src/port/
and src/common/ modules. Up to now, they got those by symlinking the
individual source files and compiling them locally. That's baroque, and a
pain to maintain, and it results in some amount of duplicated compile work.
It might've made sense when only a couple of files were needed, but the
list has grown and grown and grown :-(
It makes more sense to have the originating directory build a third variant
of libpgport.a/libpgcommon.a containing modules built with $(CFLAGS_SL),
and just link that into the shared library. Unused files won't get linked,
so the end result should be the same.
This patch makes a down payment on that idea by having src/port/ build
such a library and making libpq use it. If the buildfarm doesn't expose
fatal problems with the approach, I'll extend it to the other cases.
Tom Lane [Wed, 26 Sep 2018 23:03:33 +0000 (19:03 -0400)]
Fix another portability issue from commit 758ce9b77.
strerror.c now requires strlcpy() in some cases, and a couple of the
ecpg libraries did not have that at hand. Pull it in from src/port/
following the usual recipe. Per buildfarm.
Michael Paquier [Wed, 26 Sep 2018 22:47:20 +0000 (07:47 +0900)]
Switch flags tracking pending interrupts to sig_atomic_t
Those previously used bool, which should be safe on any modern
platforms, however the C standard is clear that it is better to use
sig_atomic_t for variables manipulated in signal handlers. This commit
adds at the same time PGDLLIMPORT to ClientConnectionLost.
Author: Michael Paquier Reviewed-by: Tom Lane, Chris Travers, Andres Freund
Discussion: https://postgr.es/m/20180925011311.GD1354@paquier.xyz
Tom Lane [Wed, 26 Sep 2018 22:23:13 +0000 (18:23 -0400)]
Try another way to detect the result type of strerror_r().
The method we've traditionally used, of redeclaring strerror_r() to
see if the compiler complains of inconsistent declarations, turns out
not to work reliably because some compilers only report a warning,
not an error. Amazingly, this has gone undetected for years, even
though it certainly breaks our detection of whether strerror_r
succeeded.
Let's instead test whether the compiler will take the result of
strerror_r() as a switch() argument. It's possible this won't
work universally either, but it's the best idea I could come up with
on the spur of the moment.
We should probably back-patch this once the dust settles, but
first let's see what the buildfarm thinks of it.
Tom Lane [Wed, 26 Sep 2018 21:35:01 +0000 (17:35 -0400)]
Clean up *printf macros to avoid conflict with format archetypes.
We must define the macro "printf" with arguments, else it can mess
up format archetype attributes in builds where PG_PRINTF_ATTRIBUTE
is just "printf". Fortunately, that's easy to do now that we're
requiring C99; we can use __VA_ARGS__.
On the other hand, it's better not to use __VA_ARGS__ for the rest
of the *printf crew, so that one can take the addresses of those
functions without surprises.
I'd proposed doing this some time ago, but forgot to make it happen;
buildfarm failures subsequent to 96bf88d52 reminded me.
Tom Lane [Wed, 26 Sep 2018 20:47:44 +0000 (16:47 -0400)]
Fix link failures due to snprintf/strerror changes.
snprintf.c requires isnan(), which requires -lm on some platforms.
libpq never bothered with -lm before, but now it needs it.
strerror.c tries to translate a string or two, which requires -lintl.
We'd managed never to need that anywhere in ecpg/pgtypeslib/ before,
but now we do.
Peter Eisentraut [Fri, 15 Jun 2018 03:22:14 +0000 (23:22 -0400)]
Recurse to sequences on ownership change for all relkinds
When a table ownership is changed, we must apply that also to any owned
sequences. (Otherwise, it would result in a situation that cannot be
restored, because linked sequences must have the same owner as the
table.) But this was previously only applied to regular tables and
materialized views. But it should also apply to at least foreign
tables. This patch removes the relkind check altogether, because it
doesn't save very much and just introduces the possibility of similar
omissions.
Bug: #15238 Reported-by: Christoph Berg <christoph.berg@credativ.de>
Tom Lane [Wed, 26 Sep 2018 17:31:56 +0000 (13:31 -0400)]
Implement %m in src/port/snprintf.c, and teach elog.c to rely on that.
I started out with the idea that we needed to detect use of %m format specs
in contexts other than elog/ereport calls, because we couldn't rely on that
working in *printf calls. But a better answer is to fix things so that it
does work. Now that we're using snprintf.c all the time, we can implement
%m in that and we've fixed the problem.
This requires also adjusting our various printf-wrapping functions so that
they ensure "errno" is preserved when they call snprintf.c.
Remove elog.c's handmade implementation of %m, and let it rely on
snprintf to support the feature. That should provide some performance
gain, though I've not attempted to measure it.
There are a lot of places where we could now simplify 'printf("%s",
strerror(errno))' into 'printf("%m")', but I'm not in any big hurry
to make that happen.
Tom Lane [Wed, 26 Sep 2018 17:13:57 +0000 (13:13 -0400)]
Always use our own versions of *printf().
We've spent an awful lot of effort over the years in coping with
platform-specific vagaries of the *printf family of functions. Let's just
forget all that mess and standardize on always using src/port/snprintf.c.
This gets rid of a lot of configure logic, and it will allow a saner
approach to dealing with %m (though actually changing that is left for
a follow-on patch).
Preliminary performance testing suggests that as it stands, snprintf.c is
faster than the native printf functions for some tasks on some platforms,
and slower for other cases. A pending patch will improve that, though
cases with floating-point conversions will doubtless remain slower unless
we want to put a *lot* of effort into that. Still, we've not observed
that *printf is really a performance bottleneck for most workloads, so
I doubt this matters much.
Tom Lane [Wed, 26 Sep 2018 16:35:57 +0000 (12:35 -0400)]
Incorporate strerror_r() into src/port/snprintf.c, too.
This provides the features that used to exist in useful_strerror()
for users of strerror_r(), too. Also, standardize on the GNU convention
that strerror_r returns a char pointer that may not be NULL.
I notice that libpq's win32.c contains a variant version of strerror_r
that probably ought to be folded into strerror.c. But lacking a
Windows environment, I should leave that to somebody else.
Tom Lane [Wed, 26 Sep 2018 15:06:42 +0000 (11:06 -0400)]
Convert elog.c's useful_strerror() into a globally-used strerror wrapper.
elog.c has long had a private strerror wrapper that handles assorted
possible failures or deficiencies of the platform's strerror. On Windows,
it also knows how to translate Winsock error codes, which the native
strerror does not. Move all this code into src/port/strerror.c and
define strerror() as a macro that invokes it, so that both our frontend
and backend code will have all of this behavior.
I believe this constitutes an actual bug fix on Windows, since AFAICS
our frontend code did not report Winsock error codes properly before this.
However, the main point is to lay the groundwork for implementing %m
in src/port/snprintf.c: the behavior we want %m to have is this one,
not the native strerror's.
Note that this throws away the prior use of src/port/strerror.c,
which was to implement strerror() on platforms lacking it. That's
been dead code for nigh twenty years now, since strerror() was
already required by C89.
We should likewise cause strerror_r to use this behavior, but
I'll tackle that separately.
Peter Eisentraut [Wed, 26 Sep 2018 11:26:24 +0000 (13:26 +0200)]
Update dummy CREATE ASSERTION grammar
While we are probably still far away from fully implementing
assertions, all patch proposals appear to take issue with the existing
dummy grammar CREATE/DROP ASSERTION productions, so update those a
little bit. Rename the rule, use any_name instead of name, and remove
some unused code. Also remove the production for DROP ASSERTION,
since that would most likely be handled via the generic DROP support.
Tomas Vondra [Wed, 26 Sep 2018 08:45:21 +0000 (10:45 +0200)]
Improve test coverage of geometric types
This commit significantly increases test coverage of geo_ops.c, adding
tests for various issues addressed by 2e2a392de3 (which went undetected
for a long time, at least partially due to not being covered).
This also removes alternative results expecting -0 on some platforms.
Instead the functions are should return the same results everywhere,
transforming -0 to 0 if needed.
The tests are added to geometric.sql file, sorted by the left hand side
of the operators. There are many cross datatype operators, so this seems
like the best solution.
Author: Emre Hasegeli Reviewed-by: Tomas Vondra
Discussion: https://www.postgresql.org/message-id/CAE2gYzxF7-5djV6-cEvqQu-fNsnt%3DEqbOURx7ZDg%2BVv6ZMTWbg%40mail.gmail.com
Tomas Vondra [Wed, 26 Sep 2018 08:25:24 +0000 (10:25 +0200)]
Fix problems in handling the line data type
According to the source history, the internal format of line data type
has changed, but various functions working with it did were not updated
and thus were producing wrong results.
This patch addresses various such issues, in particular:
* Reject invalid specification A=B=0 on receive
* Reject same points on line_construct_pp()
* Fix perpendicular operator when negative values are involved
* Avoid division by zero on perpendicular operator
* Fix intersection and distance operators when neither A nor B are 1
* Return NULL for closest point when objects are parallel
* Check whether closest point of line segments is the intersection point
* Fix closest point of line segments being on the wrong segment
Aside from handling those issues, the patch also aims to make operators
more symmetric and less sen to precision loss. The EPSILON interferes
with even minor changes, but the least we can do is applying it to both
sides of the operators equally.
Author: Emre Hasegeli Reviewed-by: Tomas Vondra
Discussion: https://www.postgresql.org/message-id/CAE2gYzxF7-5djV6-cEvqQu-fNsnt%3DEqbOURx7ZDg%2BVv6ZMTWbg%40mail.gmail.com
Michael Paquier [Wed, 26 Sep 2018 01:25:54 +0000 (10:25 +0900)]
Rework activation of commit timestamps during recovery
The activation and deactivation of commit timestamp tracking has not
been handled consistently for a primary or standbys at recovery. The
facility can be activated at three different moments of recovery:
- The beginning, where a primary would use the GUC value for the
decision-making, and where a standby relies on the contents of the
control file.
- When replaying a XLOG_PARAMETER_CHANGE record at redo.
- The end, where both primary and standby rely on the GUC value.
Using the GUC value for a primary at the beginning of recovery causes
problems with commit timestamp access when doing crash recovery.
Particularly, when replaying transaction commits, it could be possible
that an attempt to read commit timestamps is done for a transaction
which committed at a moment when track_commit_timestamp was disabled.
A test case is added to reproduce the failure. The test works down to
v11 as it takes advantage of transaction commits within procedures.
Reported-by: Hailong Li
Author: Masahiko Sawasa, Michael Paquier Reviewed-by: Kyotaro Horiguchi
Discussion: https://postgr.es/m/11224478-a782-203b-1f17-e4797b39bdf0@qunar.com
Backpatch-through: 9.5, where commit timestamps have been introduced.
Andres Freund [Tue, 25 Sep 2018 23:28:57 +0000 (16:28 -0700)]
Remove absolete function TupleDescGetSlot().
TupleDescGetSlot() was kept around for backward compatibility for
user-written SRFs. With the TupleTableSlot abstraction work, that code
will need to be version specific anyway, so there's no point in
keeping the function around any longer.
Author: Ashutosh Bapat Reviewed-By: Andres Freund
Discussion: https://postgr.es/m/20180220224318.gw4oe5jadhpmcdnm@alap3.anarazel.de
Andres Freund [Tue, 25 Sep 2018 23:27:48 +0000 (16:27 -0700)]
Split ExecStoreTuple into ExecStoreHeapTuple and ExecStoreBufferHeapTuple.
Upcoming changes introduce further types of tuple table slots, in
preparation of making table storage pluggable. New storage methods
will have different representation of tuples, therefore the slot
accessor should refer explicitly to heap tuples.
Instead of just renaming the functions, split it into one function
that accepts heap tuples not residing in buffers, and one accepting
ones in buffers. Previously one function was used for both, but that
was a bit awkward already, and splitting will allow us to represent
slot types for tuples in buffers and normal memory separately.
This is split out from the patch introducing abstract slots, as this
largely consists out of mechanical changes.
Author: Ashutosh Bapat Reviewed-By: Andres Freund
Discussion: https://postgr.es/m/20180220224318.gw4oe5jadhpmcdnm@alap3.anarazel.de
Andres Freund [Tue, 25 Sep 2018 22:59:46 +0000 (15:59 -0700)]
Change TupleTableSlot->tts_nvalid to type AttrNumber.
Previously it was an int / 4 bytes. The maximum number of attributes
in a tuple is restricted by the maximum value Var->varattno, which is
an AttrNumber/int16. Hence use the same data type for
TupleTableSlot->tts_nvalid.
Andres Freund [Tue, 25 Sep 2018 19:54:29 +0000 (12:54 -0700)]
Collect JIT instrumentation from workers.
Previously, when using parallel query, EXPLAIN (ANALYZE)'s JIT
compilation timings did not include the overhead from doing so on the
workers. Fix that.
We do so by simply aggregating the cost of doing JIT compilation on
workers and the leader together. Arguably that's not quite accurate,
because the total time spend doing so is spent in parallel - but it's
hard to do much better. For additional detail, when VERBOSE is
specified, the stats for workers are displayed separately.
Author: Amit Khandekar and Andres Freund
Discussion: https://postgr.es/m/CAJ3gD9eLrz51RK_gTkod+71iDcjpB_N8eC6vU2AW-VicsAERpQ@mail.gmail.com
Backpatch: 11-
Tom Lane [Tue, 25 Sep 2018 17:23:29 +0000 (13:23 -0400)]
Make some fixes to allow building Postgres on macOS 10.14 ("Mojave").
Apple's latest rearrangements of the system-supplied headers have broken
building of PL/Perl and PL/Tcl. The only practical way to fix PL/Tcl is to
start using the "-isysroot" compiler flag to point to SDK-supplied headers,
as Apple expects. We must also start distinguishing where to find Perl's
headers from where to find its shared library; but that seems like good
cleanup anyway.
Extensions that formerly did something like -I$(perl_archlibexp)/CORE
should now do -I$(perl_includedir)/CORE instead. perl_archlibexp
is still the place to look for libperl.so, though.
If for some reason you don't like the default -isysroot setting, you can
override that by setting PG_SYSROOT in configure's arguments. I don't
currently think people would need to do so, unless maybe for cross-version
build purposes.
In addition, teach configure where to find tclConfig.sh. Our traditional
method of searching $auto_path hasn't worked for the last couple of macOS
releases, and it now seems clear that Apple's not going to change that.
The workaround of manually specifying --with-tclconfig was annoying
already, but Mojave's made it a lot more so because the sysroot path now
has to be included as well. Let's just wire the knowledge into configure
instead. To avoid breaking builds against non-default Tcl installations
(e.g. MacPorts) wherein the $auto_path method probably still works,
arrange to try the additional case only after all else has failed.
Back-patch to all supported versions, since at least the buildfarm
cares about that. The changes are set up to not do anything on macOS
releases that are old enough to not have functional sysroot trees.
Tom Lane [Tue, 25 Sep 2018 15:09:18 +0000 (11:09 -0400)]
Avoid unnecessary precision loss for pgbench's --rate target.
It's fairly silly to truncate the throttle_delay to integer when the only
math we ever do with it requires converting back to double. Furthermore,
given that people are starting to complain about restrictions like only
supporting 1K client connections, I don't think we're very far away from
situations where the precision loss matters. As the code stood, for
example, there's no difference between --rate 100001 and --rate 111111;
both get converted to throttle_delay = 9. Somebody trying to run 100
threads and have each one dispatch around 1K TPS would find this lack of
precision rather surprising, especially since the required per-thread
delays are around 1ms, well within the timing precision of modern systems.
Michael Paquier [Tue, 25 Sep 2018 02:03:56 +0000 (11:03 +0900)]
Ignore publication tables when --no-publications is used
96e1cb4 has added support for --no-publications in pg_dump, pg_dumpall
and pg_restore, but forgot the fact that publication tables also need to
be ignored when this option is used.
Author: Gilles Darold Reviewed-by: Michael Paquier
Discussion: https://postgr.es/m/3f48e812-b0fa-388e-2043-9a176bdee27e@dalibo.com
Backpatch-through: 10, where publications have been added.
Michael Paquier [Tue, 25 Sep 2018 00:55:44 +0000 (09:55 +0900)]
Revoke pg_stat_statements_reset() permissions
Commit 25fff40 has granted execute permission of the function
pg_stat_statements_reset() to default role "pg_read_all_stats", but this
role is meant to read statistics, and not to reset them. The
permissions on this function are revoked from "pg_read_all_stats". The
version of pg_stat_statements is bumped up in consequence.
Author: Haribabu Kommi Reviewed-by: Michael Paquier, Amit Kapila
Discussion: https://postgr.es/m/CAJrrPGf5fCnKqXObpwGN9nMyD--tzOf-7LFCJiz59Z1wJ5qj9A@mail.gmail.com
Tom Lane [Mon, 24 Sep 2018 21:29:08 +0000 (17:29 -0400)]
Sync our Snowball stemmer dictionaries with current upstream.
We haven't touched these since text search functionality landed in core
in 2007 :-(. While the upstream project isn't a beehive of activity,
they do make additions and bug fixes from time to time. Update our
copies of these files.
Also update our documentation about how to keep things in sync, since
they're not making distribution tarballs these days. Fortunately,
their source code turns out to be a breeze to build.
Notable changes:
* The non-UTF8 version of the hungarian stemmer now works in LATIN2
not LATIN1.
* New stemmers have appeared for arabic, indonesian, irish, lithuanian,
nepali, and tamil. These all work in UTF8, and the indonesian and
irish ones also work in LATIN1.
(There are some new stemmers that I did not incorporate, mainly because
their names don't match the underlying languages, suggesting that they're
not to be considered mainstream.)
Worth noting: the upstream Nepali dictionary was contributed by
Arthur Zakirov.
initdb forced because the contents of snowball_create.sql have
changed.
Still TODO: see about updating the stopword lists.
Andres Freund [Mon, 24 Sep 2018 20:27:08 +0000 (13:27 -0700)]
Make EXPLAIN output for JIT compilation more dense.
A discussion about also reporting JIT compilation overhead on workers
brought unhappiness with the verbosity of the current explain format
to light. Make the text format more dense, and restructure the
structured output to mirror that more closely.
As we're re-jiggering the output format anyway: The denser format
allows us to report all flags for JIT compilation (now also reporting
PGJIT_EXPR and PGJIT_DEFORM), and report the total time in addition to
the individual times.
Per complaint from Tom Lane.
Author: Andres Freund
Discussion: https://postgr.es/m/27812.1537221015@sss.pgh.pa.us
Backpatch: 11-, where JIT compilation was introduced
Andrew Dunstan [Mon, 24 Sep 2018 20:11:24 +0000 (16:11 -0400)]
Fast default trigger and expand_tuple fixes
Ensure that triggers get properly filled in tuples for the OLD value.
Also fix the logic of detecting missing null values. The previous logic
failed to detect a missing null column before the first missing column
with a default. Fixing this has simplified the logic a bit.
Regression tests are added to test changes. This should ensure better
coverage of expand_tuple().
Original bug reports, and some code and test scripts from Tomas Vondra
Tom Lane [Mon, 24 Sep 2018 18:40:58 +0000 (14:40 -0400)]
Use ppoll(2), if available, to wait for input in pgbench.
Previously, pgbench always used select(2) for this purpose, but that's
problematic for very high client counts, because select() can't deal
with file descriptor numbers larger than FD_SETSIZE. It's pretty common
for that to be only 1024 or so, whereas modern OSes can allow many more
open files than that. Using poll(2) would surmount that problem, but it
creates another one: poll()'s timeout resolution is only 1ms, which is
poor enough to cause problems with --rate specifications approaching or
exceeding 1K TPS.
On platforms that have ppoll(2), which includes Linux and recent
FreeBSD, we can use that to avoid the FD_SETSIZE problem without any
loss of timeout resolution. Hence, add configure logic to test for
ppoll(), and use it if available.
This patch introduces an abstraction layer into pgbench that could
be extended to support other kernel event-wait APIs such as kevents.
But actually adding such support is a matter for some future patch.
Doug Rady, reviewed by Robert Haas and Fabien Coelho, and whacked around
a good bit more by me
Tom Lane [Mon, 24 Sep 2018 15:30:51 +0000 (11:30 -0400)]
Fix over-allocation of space for array_out()'s result string.
array_out overestimated the space needed for its output, possibly by
a very substantial amount if the array is multi-dimensional, because
of wrong order of operations in the loop that counts the number of
curly-brace pairs needed. While the output string is normally
short-lived, this could still cause problems in extreme cases.
An additional minor error was that it counted one more delimiter than
is actually needed.
Repair those errors, add an Assert that the space is now correctly
calculated, and make some minor improvements in the comments.
I also failed to resist the temptation to get rid of an integer
modulus operation per array element; a simple comparison is sufficient.
This bug dates clear back to Berkeley days, so back-patch to all
supported versions.
Joe Conway [Mon, 24 Sep 2018 14:14:57 +0000 (10:14 -0400)]
Document aclitem functions and operators
aclitem functions and operators have been heretofore undocumented.
Fix that. While at it, ensure the non-operator aclitem functions have
pg_description strings.
Does not seem worthwhile to back-patch.
Author: Fabien Coelho, with pg_description from John Naylor, and significant
refactoring and editorialization by me.
Reviewed by: Tom Lane
Discussion: https://postgr.es/m/flat/alpine.DEB.2.21.1808010825490.18204%40lancre
Initialize random() in bootstrap/stand-alone postgres and in initdb.
This removes a difference between the standard IsUnderPostmaster
execution environment and that of --boot and --single. In a stand-alone
backend, "SELECT random()" always started at the same seed.
On a system capable of using posix shared memory, initdb could still
conclude "selecting dynamic shared memory implementation ... sysv".
Crashed --boot or --single postgres processes orphaned shared memory
objects having names that collided with the not-actually-random names
that initdb probed. The sysv fallback appeared after ten crashes of
--boot or --single postgres. Since --boot and --single are rare in
production use, systems used for PostgreSQL development are the
principal candidate to notice this symptom.
Back-patch to 9.3 (all supported versions). PostgreSQL 9.4 introduced
dynamic shared memory, but 9.3 does share the "SELECT random()" problem.
Tom Lane [Sun, 23 Sep 2018 20:05:45 +0000 (16:05 -0400)]
Fix failure in WHERE CURRENT OF after rewinding the referenced cursor.
In a case where we have multiple relation-scan nodes in a cursor plan,
such as a scan of an inheritance tree, it's possible to fetch from a
given scan node, then rewind the cursor and fetch some row from an
earlier scan node. In such a case, execCurrent.c mistakenly thought
that the later scan node was still active, because ExecReScan hadn't
done anything to make it look not-active. We'd get some sort of
failure in the case of a SeqScan node, because the node's scan tuple
slot would be pointing at a HeapTuple whose t_self gets reset to
invalid by heapam.c. But it seems possible that for other relation
scan node types we'd actually return a valid tuple TID to the caller,
resulting in updating or deleting a tuple that shouldn't have been
considered current. To fix, forcibly clear the ScanTupleSlot in
ExecScanReScan.
Another issue here, which seems only latent at the moment but could
easily become a live bug in future, is that rewinding a cursor does
not necessarily lead to *immediately* applying ExecReScan to every
scan-level node in the plan tree. Upper-level nodes will think that
they can postpone that call if their child node is already marked
with chgParam flags. I don't see a way for that to happen today in
a plan tree that's simple enough for execCurrent.c's search_plan_tree
to understand, but that's one heck of a fragile assumption. So, add
some logic in search_plan_tree to detect chgParam flags being set on
nodes that it descended to/through, and assume that that means we
should consider lower scan nodes to be logically reset even if their
ReScan call hasn't actually happened yet.
Per bug #15395 from Matvey Arye. This has been broken for a long time,
so back-patch to all supported branches.
Replace CAS loop with single TAS in ProcArrayGroupClearXid()
Single pg_atomic_exchange_u32() is expected to be faster than loop of
pg_atomic_compare_exchange_u32(). Also, it would be consistent with
clog group update code.
Discussion: https://postgr.es/m/CAPpHfdtxLsC-bqfxFcHswZ91OxXcZVNDBBVfg9tAWU0jvn1tQA%40mail.gmail.com Reviewed-by: Amit Kapila
Michael Paquier [Sat, 22 Sep 2018 06:23:59 +0000 (15:23 +0900)]
Make GUC wal_sender_timeout user-settable
Being able to use a value that can be changed on a connection basis is
useful with clusters distributed geographically, and makes failure
detection more flexible. A note is added in the documentation about the
use of "options" in primary_conninfo, which can be hard to grasp for
newcomers with the need of two single quotes when listing a set of
parameters.
Tom Lane [Sat, 22 Sep 2018 00:50:18 +0000 (20:50 -0400)]
Get rid of explicit argument-count markings in tab-complete.c.
This replaces the "TailMatchesN" macros with just "TailMatches",
and likewise "HeadMatchesN" becomes "HeadMatches" and "MatchesN"
becomes "Matches". The various COMPLETE_WITH_LISTn macros are
reduced to COMPLETE_WITH, and the single-item COMPLETE_WITH_CONST
also gets folded into that. This eliminates a lot of minor
annoyance in writing tab-completion rules. Usefully, the compiled
code also gets a bit smaller (10% or so, on my machine).
The implementation depends on variadic macros, so we couldn't have
done this before we required C99.
Andres Freund and Thomas Munro; some cosmetic cleanup by me.
Bruce Momjian [Sat, 22 Sep 2018 00:28:55 +0000 (20:28 -0400)]
doc: JIT is enabled by default in PG 12
JIT was disabled by default in a PG 11 in a separate commit that will
normally not appear in the PG 12 git logs. Therefore, create a PG 12
document and mention the fact that JIT is enabled by default in this
release. (A similar change in parallelism was missed in a prior
release.)
Reported-by: Andres Freund
Discussion: https://postgr.es/m/20180922000554.qukbhhlagpnopvko@alap3.anarazel.de
Tom Lane [Fri, 21 Sep 2018 19:58:37 +0000 (15:58 -0400)]
Fix bogus tab-completion rule for CREATE PUBLICATION.
You can't use "FOR TABLE" as a single Matches argument, because readline
will consider that input to be two words not one. It's necessary to make
the pattern contain two arguments.
The case accidentally worked anyway because the words_after_create
test fired ... but only for the first such table name.
Noted by Edmund Horner, though this isn't exactly his proposed fix.
Backpatch to v10 where the faulty code came in.
Tom Lane [Fri, 21 Sep 2018 19:22:26 +0000 (15:22 -0400)]
Improve tab completion for ANALYZE, EXPLAIN, and VACUUM.
Previously, we made no attempt to provide tab completion in these
statements' optional parenthesized options lists. This patch teaches
psql to do so.
To prevent the option completions from being offered after we've already
seen a complete parenthesized option list, it's necessary to improve
word_matches() so that it allows a wildcard '*' in the middle of an
alternative, not only at the end as formerly. That requires only a
little more code than before, and it allows us to test for "incomplete
parenthesized options" with a test like
else if (HeadMatches2("EXPLAIN", "(*") &&
!HeadMatches2("EXPLAIN", "(*)"))
In addition, add some logic to offer column names in the context of
"ANALYZE tablename ( ...", and likewise for VACUUM. This isn't real
complete; it won't offer column names again after a comma. But it's
better than before, and it doesn't take much code.
Justin Pryzby, reviewed at various times by Álvaro Herrera, Arthur
Zakirov, and Edmund Horner; some additional fixups by me
Tom Lane [Fri, 21 Sep 2018 16:41:00 +0000 (12:41 -0400)]
Rationalize Query_for_list_of_[relations] query names in tab-complete.c.
The previous convention was to use names based on the set of relkinds being
selected for, which was not at all helpful for maintenance, especially
since people had been quite inconsistent about whether to change the names
when they changed the relkinds being selected for. Instead, use names
based on the functionality we need the relation to have, following the
model established by Query_for_list_of_updatables.
While at it, sort the list of Query constants a bit better; it had the
distinct air of code-assembled-by-dartboard before.
Thomas Munro [Fri, 21 Sep 2018 12:40:13 +0000 (00:40 +1200)]
Use size_t consistently in dsa.{ch}.
Takeshi Ideriha complained that there is a mixture of Size and size_t
in dsa.c and corresponding header. Let's use size_t. Back-patch to 10
where dsa.c landed, to make future back-patching easy.
Andres Freund [Thu, 13 Sep 2018 21:18:43 +0000 (14:18 -0700)]
Error out for clang on x86-32 without SSE2 support, no -fexcess-precision.
As clang currently doesn't support -fexcess-precision=standard,
compiling x86-32 code with SSE2 disabled, can lead to problems with
floating point overflow checks and the like.
This issue was noticed because clang, on at least some BSDs, defaults
to i386 compatibility, whereas it defaults to pentium4 on Linux. Our
forced usage of __builtin_isinf() lead to some overflow checks not
triggering when compiling for i386, e.g. when the result of the
calculation didn't overflow in 80bit registers, but did so in 64bit.
While we could just fall back to a non-builtin isinf, it seems likely
that the use of 80bit registers leads to other problems (which is why
we force the flag for GCC already). Therefore error out when
detecting clang in that situation.
Reported-By: Victor Wagner Analyzed-By: Andrew Gierth and Andres Freund
Author: Andres Freund
Discussion: https://postgr.es/m/20180905005130.ewk4xcs5dgyzcy45@alap3.anarazel.de
Backpatch: 9.3-, all supported versions are affected
Tom Lane [Thu, 20 Sep 2018 21:21:14 +0000 (17:21 -0400)]
Fix psql's tab completion for TABLE.
This should offer the same relation types that SELECT ... FROM would.
You can't select from an index for instance, so offering it here is
unhelpful. Noted while testing ilmari's recent patch.
Tom Lane [Thu, 20 Sep 2018 21:16:06 +0000 (17:16 -0400)]
Fix psql's tab completion for ALTER DATABASE ... SET TABLESPACE.
We have the infrastructure to offer a list of tablespace names, but
it wasn't being used here; instead you got "FROM", "CURRENT", and "TO"
which aren't actually legal in this syntax.
Dagfinn Ilmari Mannsåker, reviewed by Arthur Zakirov
Tom Lane [Thu, 20 Sep 2018 20:06:18 +0000 (16:06 -0400)]
Add missing pg_description strings for pg_type entries.
I noticed that all non-composite, non-array entries in pg_type.dat
had descr strings, except for "json" and the pseudo-types. The
lack for json seems certainly an oversight, and there's surely
little reason to not have entries for the pseudo-types either.
So add some.
"make reformat-dat-files" turned up some formatting issues in
pg_amop.dat, too, so fix those in passing.
No catversion bump since the backend doesn't care too much what is
in pg_description.
Tom Lane [Thu, 20 Sep 2018 19:14:46 +0000 (15:14 -0400)]
Teach genbki.pl to auto-generate pg_type entries for array types.
This eliminates some more tedium in adding new catalog entries,
specifically the need to set up an array type when adding a new
built-in data type. Now it's sufficient to assign an OID for the
array type and write it in an "array_type_oid" metadata field.
You don't have to fill the base type's typarray link explicitly, either.
No catversion bump since the contents of pg_type aren't changed.
(Well, their order might be different, but that doesn't matter.)
John Naylor, reviewed and whacked around a bit by
Dagfinn Ilmari Mannsåker, and some more by me.
Fix handling of format string text characters in to_timestamp()/to_date()
cf984672 introduced improvement of handling of spaces and separators in
to_timestamp()/to_date() functions. In particular, now we're skipping spaces
both before and after fields. That may cause format string text character to
consume part of field in the situations, when it didn't happen before cf984672.
This commit cause format string text character consume input string characters
only when since previous field (or string beginning) number of skipped input
string characters is not greater than number of corresponding format string
characters (that is we didn't skip any extra characters in input string).
Thomas Munro [Thu, 20 Sep 2018 03:52:39 +0000 (15:52 +1200)]
Fix segment_bins corruption in dsa.c.
If a segment has been freed by dsa.c because it is entirely empty, other
backends must make sure to unmap it before following links to new
segments that might happen to have the same index number, or they could
finish up looking at a defunct segment and then corrupt the segment_bins
lists. The correct protocol requires checking freed_segment_counter
after acquiring the area lock and before resolving any index number to a
segment. Add the missing checks and an assertion.
Back-patch to 10, where dsa.c first arrived.
Author: Thomas Munro Reported-by: Tomas Vondra
Discussion: https://postgr.es/m/CAEepm%3D0thg%2Bja5zGVa7jBy-uqyHrTqTm8HGhEOtMmigGrAqTbw%40mail.gmail.com
Thomas Munro [Thu, 20 Sep 2018 02:02:40 +0000 (14:02 +1200)]
Defer restoration of libraries in parallel workers.
Several users of extensions complained of crashes in parallel workers
that turned out to be due to syscache access from their _PG_init()
functions. Reorder the initialization of parallel workers so that
libraries are restored after the caches are initialized, and inside a
transaction.
This was reported in bug #15350 and elsewhere. We don't consider it
to be a bug: extensions shouldn't do that, because then they can't be
used in shared_preload_libraries. However, it's a fairly obscure
hazard and these extensions worked in practice before parallel query
came along. So let's make it work. Later commits might add a warning
message and eventually an error.
Back-patch to 9.6, where parallel query landed.
Author: Thomas Munro Reviewed-by: Amit Kapila Reported-by: Kieran McCusker, Jimmy
Discussion: https://postgr.es/m/153512195228.1489.8545997741965926448%40wrigleys.postgresql.org
Michael Paquier [Wed, 19 Sep 2018 23:54:37 +0000 (08:54 +0900)]
Enforce translation mode for Windows frontends to text with open/fopen
Allowing frontends to use concurrent-safe open() and fopen() via 0ba06e0
has the side-effect of switching the default translation mode from text
to binary, so the switch can cause breakages for frontend tools when the
caller of those new versions specifies neither binary and text. This
commit makes sure to maintain strict compatibility with past versions,
so as no frontends should see a difference when upgrading.
Author: Laurenz Albe Reviewed-by: Michael Paquier, Tom Lane
Discussion: https://postgr.es/m/20180917140202.GF31460@paquier.xyz
Tom Lane [Wed, 19 Sep 2018 16:43:51 +0000 (12:43 -0400)]
Don't ignore locktable-full failures in StandbyAcquireAccessExclusiveLock.
Commit 37c54863c removed the code in StandbyAcquireAccessExclusiveLock
that checked the return value of LockAcquireExtended. That created a
bug, because it's still passing reportMemoryError = false to
LockAcquireExtended, meaning that LOCKACQUIRE_NOT_AVAIL will be returned
if we're out of shared memory for the lock table.
In such a situation, the startup process would believe it had acquired an
exclusive lock even though it hadn't, with potentially dire consequences.
To fix, just drop the use of reportMemoryError = false, which allows us
to simplify the call into a plain LockAcquire(). It's unclear that the
locktable-full situation arises often enough that it's worth having a
better recovery method than crash-and-restart. (I strongly suspect that
the only reason the code path existed at all was that it was relatively
simple to do in the pre-37c54863c implementation. But now it's not.)
LockAcquireExtended's reportMemoryError parameter is now dead code and
could be removed. I refrained from doing so, however, because there
was some interest in resurrecting the behavior if we do get reports of
locktable-full failures in the field. Also, it seems unwise to remove
the parameter concurrently with shipping commit f868a8143, which added a
parameter; if there are any third-party callers of LockAcquireExtended,
we want them to get a wrong-number-of-parameters compile error rather
than a possibly-silent misinterpretation of its last parameter.
Add support for nearest-neighbor (KNN) searches to SP-GiST
Currently, KNN searches were supported only by GiST. SP-GiST also capable to
support them. This commit implements that support. SP-GiST scan stack is
replaced with queue, which serves as stack if no ordering is specified. KNN
support is provided for three SP-GIST opclasses: quad_point_ops, kd_point_ops
and poly_ops (catversion is bumped). Some common parts between GiST and SP-GiST
KNNs are extracted into separate functions.
Discussion: https://postgr.es/m/570825e8-47d0-4732-2bf6-88d67d2d51c8%40postgrespro.ru
Author: Nikita Glukhov, Alexander Korotkov based on GSoC work by Vlad Sterzhanov
Review: Andrey Borodin, Alexander Korotkov
Tom Lane [Tue, 18 Sep 2018 21:11:54 +0000 (17:11 -0400)]
Add a debugging option to stress-test outfuncs.c and readfuncs.c.
In the normal course of operation, query trees will be serialized only if
they are stored as views or rules; and plan trees will be serialized only
if they get passed to parallel-query workers. This leaves an awful lot of
opportunity for bugs/oversights to not get detected, as indeed we've just
been reminded of the hard way.
To improve matters, this patch adds a new compile option
WRITE_READ_PARSE_PLAN_TREES, which is modeled on the longstanding option
COPY_PARSE_PLAN_TREES; but instead of passing all parse and plan trees
through copyObject, it passes them through nodeToString + stringToNode.
Enabling this option in a buildfarm animal or two will catch problems
at least for cases that are exercised by the regression tests.
A small problem with this idea is that readfuncs.c historically has
discarded location fields, on the reasonable grounds that parse
locations in a retrieved view are not relevant to the current query.
But doing that in WRITE_READ_PARSE_PLAN_TREES breaks pg_stat_statements,
and it could cause problems for future improvements that might try to
report error locations at runtime. To fix that, provide a variant
behavior in readfuncs.c that makes it restore location fields when
told to.
In passing, const-ify the string arguments of stringToNode and its
subsidiary functions, just because it annoyed me that they weren't
const already.
Tom Lane [Tue, 18 Sep 2018 19:08:28 +0000 (15:08 -0400)]
Fix some minor issues exposed by outfuncs/readfuncs testing.
A test patch to pass parse and plan trees through outfuncs + readfuncs
exposed several issues that need to be fixed to get clean matches:
Query.withCheckOptions failed to get copied; it's intentionally ignored
by outfuncs/readfuncs on the grounds that it'd always be NIL anyway in
stored rules. This seems less than future-proof, and it's not even
saving very much, so just undo the decision and treat the field like
all others.
Several places that convert a view RTE into a subquery RTE, or similar
manipulations, failed to clear out fields that were specific to the
original RTE type and should be zero in a subquery RTE. Since readfuncs.c
will leave such fields as zero, equalfuncs.c thinks the nodes are different
leading to a reported mismatch. It seems like a good idea to clear out the
no-longer-needed fields, even though in principle nothing should look at
them; the node ought to be indistinguishable from how it would look if
we'd built a new node instead of scribbling on the old one.
BuildOnConflictExcludedTargetlist randomly set the resname of some
TargetEntries to "" not NULL. outfuncs/readfuncs don't distinguish those
cases, and so the string will read back in as NULL ... but equalfuncs.c
does distinguish. Perhaps we ought to try to make things more consistent
in this area --- but it's just useless extra code space for
BuildOnConflictExcludedTargetlist to not use NULL here, so I fixed it for
now by making it do that.
catversion bumped because the change in handling of Query.withCheckOptions
affects stored rules.
Tom Lane [Tue, 18 Sep 2018 17:02:27 +0000 (13:02 -0400)]
Fix some probably-minor oversights in readfuncs.c.
The system expects TABLEFUNC RTEs to have coltypes, coltypmods, and
colcollations lists, but outfuncs doesn't dump them and readfuncs doesn't
restore them. This doesn't cause obvious failures, because the only things
that look at those fields are expandRTE() and get_rte_attribute_type(),
which are mostly used during parse analysis, before anything would've
passed the parsetree through outfuncs/readfuncs. But expandRTE() is used
in build_physical_tlist(), which means that that function will return a
wrong answer for a TABLEFUNC RTE that came from a view. Very accidentally,
this doesn't cause serious problems, because what it will return is NIL
which callers will interpret as "couldn't build a physical tlist because
of dropped columns". So you still get a plan that works, though it's
marginally less efficient than it could be. There are also some other
expandRTE() calls associated with transformation of whole-row Vars in
the planner. I have been unable to exhibit misbehavior from that, and
it may be unreachable in any case that anyone would care about ... but
I'm not entirely convinced, so this seems like something we should back-
patch a fix for. Fortunately, we can fix it without forcing a change
of stored rules and a catversion bump, because we can just copy these
lists from the subsidiary TableFunc object.
readfuncs.c was also missing support for NamedTuplestoreScan plan nodes.
This accidentally fails to break parallel query because a query using
a named tuplestore would never be considered parallel-safe anyway.
However, project policy since parallel query came in is that all plan
node types should have outfuncs/readfuncs support, so this is clearly
an oversight that should be repaired.
Noted while fooling around with a patch to test outfuncs/readfuncs more
thoroughly. That exposed some other issues too, but these are the only
ones that seem worth back-patching.
Back-patch to v10 where both of these features came in.
Thomas Munro [Tue, 18 Sep 2018 10:56:36 +0000 (22:56 +1200)]
Allow DSM allocation to be interrupted.
Chris Travers reported that the startup process can repeatedly try to
cancel a backend that is in a posix_fallocate()/EINTR loop and cause it
to loop forever. Teach the retry loop to give up if an interrupt is
pending. Don't actually check for interrupts in that loop though,
because a non-local exit would skip some clean-up code in the caller.
Back-patch to 9.4 where DSM was added (and posix_fallocate() was later
back-patched).
Author: Chris Travers Reviewed-by: Ildar Musin, Murat Kabilov, Oleksii Kliukin Tested-by: Oleksii Kliukin
Discussion: https://postgr.es/m/CAN-RpxB-oeZve_J3SM_6%3DHXPmvEG%3DHX%2B9V9pi8g2YR7YW0rBBg%40mail.gmail.com
Michael Paquier [Tue, 18 Sep 2018 03:00:18 +0000 (12:00 +0900)]
Refactor routines for subscription and publication lookups
Those routines gain a missing_ok argument, allowing a caller to get a
NULL result instead of an error if set to true. This is part of a
larger refactoring effort for objectaddress.c where trying to check for
non-existing objects does not result in cache lookup failures.
Author: Michael Paquier Reviewed-by: Aleksander Alekseev, Álvaro Herrera
Discussion: https://postgr.es/m/CAB7nPqSZxrSmdHK-rny7z8mi=EAFXJ5J-0RbzDw6aus=wB5azQ@mail.gmail.com
Tom Lane [Mon, 17 Sep 2018 17:16:32 +0000 (13:16 -0400)]
Fix parsetree representation of XMLTABLE(XMLNAMESPACES(DEFAULT ...)).
The original coding for XMLTABLE thought it could represent a default
namespace by a T_String Value node with a null string pointer. That's
not okay, though; in particular outfuncs.c/readfuncs.c are not on board
with such a representation, meaning you'll get a null pointer crash
if you try to store a view or rule containing this construct.
To fix, change the parsetree representation so that we have a NULL
list element, instead of a bogus Value node.
This isn't really a functional limitation since default XML namespaces
aren't yet implemented in the executor; you'd just get "DEFAULT
namespace is not supported" anyway. But crashes are not nice, so
back-patch to v10 where this syntax was added. Ordinarily we'd consider
a parsetree representation change to be un-backpatchable; but since
existing releases would crash on the way to storing such constructs,
there can't be any existing views/rules to be incompatible with.
Tom Lane [Mon, 17 Sep 2018 16:43:07 +0000 (12:43 -0400)]
Remove dead code from pop_next_work_item().
The pref_non_data heuristic has been dead code for nearly ten years,
and as far as I can tell was dead code even when it was first committed.
I'm tired of silencing Coverity complaints about it, so get rid of it.
If anyone is ever interested in pursuing the concept, they can get the
code out of our git history.
Tom Lane [Mon, 17 Sep 2018 16:11:43 +0000 (12:11 -0400)]
Fix pgbench lexer's "continuation" rule to cope with Windows newlines.
Our general practice in frontend code is to accept input with either
Unix-style newlines (\n) or DOS-style (\r\n). pgbench was mostly down
with that, but its rule for line continuations (backslash-newline) was
not. This had been masked on Windows buildfarm machines before commit 0ba06e0bf by use of Windows text mode to read files. We could have fixed
it by forcing text mode again, but it's better to fix the parsing code
so that Windows-style text files on Unix systems don't cause problems.
Back-patch to v10 where pgbench grew line continuations.
Andrew Gierth [Sun, 16 Sep 2018 17:46:45 +0000 (18:46 +0100)]
Fix out-of-tree build for transform modules.
Neither plperl nor plpython installed sufficient header files to
permit transform modules to be built out-of-tree using PGXS. Fix that
by installing all plperl and plpython header files (other than those
with special purposes such as generated data tables), and also install
plpython's special .mk file for mangling regression tests.
(This commit does not fix the windows install, which does not
currently install _any_ plperl or plpython headers.)
Also fix the existing transform modules for hstore and ltree so that
their cross-module #include directives work as anticipated by commit df163230b9 et seq. This allows them to serve as working examples of
how to reference other modules when doing separate out-of-tree builds.
Tom Lane [Sun, 16 Sep 2018 17:02:47 +0000 (13:02 -0400)]
Add outfuncs.c support for RawStmt nodes.
I noticed while poking at a report from Andrey Lepikhov that the
recent addition of RawStmt nodes at the top of raw parse trees
makes it impossible to print any raw parse trees whatsoever,
because outfuncs.c doesn't know RawStmt and hence fails to descend
into it.
While we generally lack outfuncs.c support for utility statements,
there is reasonably complete support for what you can find in a
raw SELECT statement. It was not my intention to make that all
dead code ... so let's add support for RawStmt.
Tom Lane [Sat, 15 Sep 2018 21:24:35 +0000 (17:24 -0400)]
In v11, disable JIT by default (it's still enabled by default in HEAD).
Per discussion, JIT isn't quite mature enough to ship enabled-by-default.
I failed to resist the temptation to do a bunch of copy-editing on the
related documentation. Also, clean up some inconsistencies in which
section of config.sgml the JIT GUCs are documented in vs. what guc.c
and postgresql.config.sample had.
Tom Lane [Sat, 15 Sep 2018 17:42:33 +0000 (13:42 -0400)]
Fix failure with initplans used conditionally during EvalPlanQual rechecks.
The EvalPlanQual machinery assumes that any initplans (that is,
uncorrelated sub-selects) used during an EPQ recheck would have already
been evaluated during the main query; this is implicit in the fact that
execPlan pointers are not copied into the EPQ estate's es_param_exec_vals.
But it's possible for that assumption to fail, if the initplan is only
reached conditionally. For example, a sub-select inside a CASE expression
could be reached during a recheck when it had not been previously, if the
CASE test depends on a column that was just updated.
This bug is old, appearing to date back to my rewrite of EvalPlanQual in
commit 9f2ee8f28, but was not detected until Kyle Samson reported a case.
To fix, force all not-yet-evaluated initplans used within the EPQ plan
subtree to be evaluated at the start of the recheck, before entering the
EPQ environment. This could be inefficient, if such an initplan is
expensive and goes unused again during the recheck --- but that's piling
one layer of improbability atop another. It doesn't seem worth adding
more complexity to prevent that, at least not in the back branches.
It was convenient to use the new-in-v11 ExecEvalParamExecParams function
to implement this, but I didn't like either its name or the specifics of
its API, so revise that.
Back-patch all the way. Rather than rewrite the patch to avoid depending
on bms_next_member() in the oldest branches, I chose to back-patch that
function into 9.4 and 9.3. (This isn't the first time back-patches have
needed that, and it exhausted my patience.) I also chose to back-patch
some test cases added by commits 71404af2a and 342a1ffa2 into 9.4 and 9.3,
so that the 9.x versions of eval-plan-qual.spec are all the same.
Andrew Gierth diagnosed the problem and contributed the added test cases,
though the actual code changes are by me.
Tom Lane [Fri, 14 Sep 2018 21:31:51 +0000 (17:31 -0400)]
Improve parallel scheduling logic in pg_dump/pg_restore.
Previously, the way this worked was that a parallel pg_dump would
re-order the TABLE_DATA items in the dump's TOC into decreasing size
order, and separately re-order (some of) the INDEX items into decreasing
size order. Then pg_dump would dump the items in that order. Later,
parallel pg_restore just followed the TOC order. This method had lots
of deficiencies:
* TOC ordering randomly differed between parallel and non-parallel
dumps, and was hard to predict in the former case, causing problems
for building stable pg_dump test cases.
* Parallel restore only followed a well-chosen order if the dump had
been done in parallel; in particular, this never happened for restore
from custom-format dumps.
* The best order for restore isn't necessarily the same as for dump,
and it's not really static either because of locking considerations.
* TABLE_DATA and INDEX items aren't the only things that might take a lot
of work during restore. Scheduling was particularly stupid for the BLOBS
item, which might require lots of work during dump as well as restore,
but was left to the end in either case.
This patch removes the logic that changed the TOC order, fixing the
test instability problem. Instead, we sort the parallelizable items
just before processing them during a parallel dump. Independently
of that, parallel restore prioritizes the ready-to-execute tasks
based on the size of the underlying table. In the case of dependent
tasks such as index, constraint, or foreign key creation, the largest
relevant table is used as the metric for estimating the task length.
(This is pretty crude, but it should be enough to avoid the case we
want to avoid, which is ending the run with just a few large tasks
such that we can't make use of all N workers.)
Patch by me, responding to a complaint from Peter Eisentraut,
who also reviewed the patch.
Fix ALTER/TYPE on columns referenced by FKs in partitioned tables
When ALTER TABLE ... SET DATA TYPE affects a column referenced by
constraints and indexes, it drop those constraints and indexes and
recreates them afterwards, so that the definitions match the new data
type. The original code did this by dropping one object at a time
(commit 077db40fa1f3 of May 2004), which worked fine because the
dependencies between the objects were pretty straightforward, and
ordering the objects in a specific way was enough to make this work.
However, when there are foreign key constraints in partitioned tables,
the dependencies are no longer so straightforward, and we were getting
errors when attempted:
ERROR: cache lookup failed for constraint 16398
This can be fixed by doing all the drops in one pass instead, using
performMultipleDeletions (introduced by df18c51f2955 of Aug 2006). With
this change we can also remove the code to carefully order the list of
objects to be deleted.
Reported-by: Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com> Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/CAKcux6nWS_m+s=1Udk_U9B+QY7pA-Ac58qR5BdUfOyrwnWHDew@mail.gmail.com
Andrew Gierth [Fri, 14 Sep 2018 16:35:42 +0000 (17:35 +0100)]
Order active window clauses for greater reuse of Sort nodes.
By sorting the active window list lexicographically by the sort clause
list but putting longer clauses before shorter prefixes, we generate
more chances to elide Sort nodes when building the path.
Author: Daniel Gustafsson (with some editorialization by me) Reviewed-by: Alexander Kuzmenkov, Masahiko Sawada, Tom Lane
Discussion: https://postgr.es/m/124A7F69-84CD-435B-BA0E-2695BE21E5C2%40yesql.se
Amit Kapila [Fri, 14 Sep 2018 04:06:30 +0000 (09:36 +0530)]
Don't allow LIMIT/OFFSET clause within sub-selects to be pushed to workers.
Allowing sub-select containing LIMIT/OFFSET in workers can lead to
inconsistent results at the top-level as there is no guarantee that the
row order will be fully deterministic. The fix is to prohibit pushing
LIMIT/OFFSET within sub-selects to workers.
Reported-by: Andrew Fletcher
Bug: 15324
Author: Amit Kapila Reviewed-by: Dilip Kumar
Backpatch-through: 9.6
Discussion: https://postgr.es/m/153417684333.10284.11356259990921828616@wrigleys.postgresql.org
Michael Paquier [Fri, 14 Sep 2018 01:04:14 +0000 (10:04 +0900)]
Allow concurrent-safe open() and fopen() in frontend code for Windows
PostgreSQL uses a custom wrapper for open() and fopen() which is
concurrent-safe, allowing multiple processes to open and work on the
same file. This has a couple of advantages:
- pg_test_fsync does not handle O_DSYNC correctly otherwise, leading to
false claims that disks are unsafe.
- TAP tests can run into race conditions when a postmaster and pg_ctl
open postmaster.pid, fixing some random failures in the buildfam.
pg_upgrade is one frontend tool using workarounds to bypass file locking
issues with the log files it generates, however the interactions with
pg_ctl are proving to be tedious to get rid of, so this is left for
later.
Michael Paquier [Thu, 13 Sep 2018 22:35:39 +0000 (07:35 +0900)]
Improve autovacuum logging for aggressive and anti-wraparound runs
A log message was being generated when log_min_duration is reached for
autovacuum on a given relation to indicate if it was an aggressive run,
and missed the point of mentioning if it is doing an anti-wrapround
run. The log message generated is improved so as one, both or no extra
details are added depending on the option set.
Author: Sergei Kornilov Reviewed-by: Masahiko Sawada, Michael Paquier
Discussion: https://postgr.es/m/11587951532155118@sas1-19a94364928d.qloud-c.yandex.net
Andres Freund [Thu, 13 Sep 2018 17:41:39 +0000 (10:41 -0700)]
Detect LLVM 7 without specifying binaries explicitly.
Before this commit LLVM 7 was supported, but only if one explicitly
provided LLVM_CONFIG= and CLANG= paths. As LLVM 7 is the first
version that includes our upstreamed debugging and profiling features,
and as debian is planning to default to 7 due to wider architecture
support, it seems good to support auto-detecting that version.
Author: Christoph Berg
Discussion: https://postgr.es/m/20180912124517.GD24584@msg.df7cb.de
Backpatch: 11, where LLVM was introduced
Tom Lane [Thu, 13 Sep 2018 16:36:21 +0000 (12:36 -0400)]
Attempt to identify system timezone by reading /etc/localtime symlink.
On many modern platforms, /etc/localtime is a symlink to a file within the
IANA database. Reading the symlink lets us find out the name of the system
timezone directly, without going through the brute-force search embodied in
scan_available_timezones(). This shortens the runtime of initdb by some
tens of ms, which is helpful for the buildfarm, and it also allows us to
reliably select the same zone name the system was actually configured for,
rather than possibly choosing one of IANA's many zone aliases. (For
example, in a system configured for "Asia/Tokyo", the brute-force search
would not choose that name but its alias "Japan", on the grounds of the
latter string being shorter. More surprisingly, "Navajo" is preferred
to either "America/Denver" or "US/Mountain", as seen in an old complaint
from Josh Berkus.)
If /etc/localtime doesn't exist, or isn't a symlink, or we can't make
sense of its contents, or the contents match a zone we know but that
zone doesn't match the observed behavior of localtime(), fall back to
the brute-force search.
Also, tweak initdb so that it prints the zone name it selected.
In passing, replace the last few references to the "Olson" database in
code comments with "IANA", as that's been our preferred term since
commit b2cbced9e.
Patch by me, per a suggestion from Robert Haas; review by Michael Paquier
Amit Kapila [Thu, 13 Sep 2018 10:00:26 +0000 (15:30 +0530)]
Attach FPI to the first record after full_page_writes is turned on.
XLogInsert fails to attach a required FPI to the first record after
full_page_writes is turned on by the last checkpoint. This bug got
introduced in 9.5 due to code rearrangement in commits 2c03216d83 and 2076db2aea. Fix it by ensuring that XLogInsertRecord performs a
recomputation when the given record is generated with FPW as off but
found that the flag has been turned on while actually inserting the
record.
Reported-by: Kyotaro Horiguchi
Author: Kyotaro Horiguchi Reviewed-by: Amit Kapila
Backpatch-through: 9.5 where this problem was introduced
Discussion: https://postgr.es/m/20180420.151043.74298611.horiguchi.kyotaro@lab.ntt.co.jp
Michael Paquier [Thu, 13 Sep 2018 07:56:57 +0000 (16:56 +0900)]
Simplify static function in extension.c
An extra argument for the filename defining the extension script
location was present, aimed at being used for error reporting, but has
never been used. This was around since extensions have been added in d9572c4.
Peter Eisentraut [Mon, 27 Aug 2018 13:50:50 +0000 (15:50 +0200)]
Simplify index tuple descriptor initialization
We have two code paths for initializing the tuple descriptor for a new
index: For a normal index, we copy the tuple descriptor from the table
and reset a number of fields that are not applicable to indexes. For an
expression index, we make a blank tuple descriptor and fill in the
needed fields based on the provided expressions. As pg_attribute has
grown over time, the number of fields that we need to reset in the first
case is now bigger than the number of fields we actually want to copy,
so it's sensible to do it the other way around: Make a blank descriptor
and copy just the fields we need. This also allows more code sharing
between the two branches, and it avoids having to touch this code for
almost every unrelated change to the pg_attribute structure.
Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru>
projects / postgresql / log