|From:||David Fetter <david(at)fetter(dot)org>|
|To:||PostgreSQL Announce <pgsql-announce(at)postgresql(dot)org>|
|Subject:||== PostgreSQL Weekly News - March 25 2018 ==|
|Views:||Raw Message | Whole Thread | Download mbox|
== PostgreSQL Weekly News - March 25 2018 ==
== PostgreSQL Product News ==
pg_chameleon 2.0.5, a tool for replicating from MySQL to PostgreSQL, released.
pglogical 2.2, a logical-WAL-based replication system for PostgreSQL, released.
== PostgreSQL Jobs for March ==
== PostgreSQL Local ==
PGConf APAC 2018 will be held in Singapore March 22-23, 2018.
The German-speaking PostgreSQL Conference 2018 will take place on April 13th,
2018 in Berlin.
PGConfNepal 2018 will be held May 4-5, 2018 at Kathmandu University, Dhulikhel,
PGCon 2018 will take place in Ottawa on May 29 - June 1, 2018.
Swiss PGDay 2018 will take place in Rapperswil (near Zurich) on June 29, 2018.
The CfP is open February 6, 2018 through April 14, 2018, and registration is
open February 6, 2018 through June 28, 2018.
PGConf.Brazil 2018 will take place in São Paulo, Brazil on August 3-4 2018.
== PostgreSQL in the News ==
Planet PostgreSQL: http://planet.postgresql.org/
PostgreSQL Weekly News is brought to you this week by David Fetter
Submit news and announcements by Sunday at 3:00pm EST5EDT. Please send English
language ones to david(at)fetter(dot)org, German language to pwn(at)pgug(dot)de, Italian
language to pwn(at)itpug(dot)org(dot)
== Applied Patches ==
Magnus Hagander pushed:
- Fix typo in comment. Author: Daniel Gustafsson <daniel(at)yesql(dot)se>
Robert Haas pushed:
- Rewrite recurse_union_children to iterate, rather than recurse. Also, rename
it to plan_union_chidren, so the old name wasn't very descriptive. This
results in a small net reduction in code, seems at least to me to be easier to
understand, and saves space on the process stack. Patch by me, reviewed and
tested by Ashutosh Bapat and Rajkumar Raghuwanshi. Discussion:
- Generate a separate upper relation for each stage of setop planning. Commit
3fc6e2d7f5b652b417fa6937c34de2438d60fa9f made setop planning stages return
paths rather than plans, but all such paths were loosely associated with a
single RelOptInfo, and only the final path was added to the RelOptInfo. Even
at the time, it was foreseen that this should be changed, because there is
otherwise no good way for a single stage of setop planning to return multiple
paths. With this patch, each stage of set operation planning now creates a
separate RelOptInfo; these are distinguished by using appropriate relid sets.
Note that this patch does nothing whatsoever about actually returning multiple
paths for the same set operation; it just makes it possible for a future patch
to do so. Along the way, adjust things so that create_upper_paths_hook is
called for each of these new RelOptInfos rather than just once, since that
might be useful to extensions using that hook. It might be a good to provide
an FDW API here as well, but I didn't try to do that for now. Patch by me,
reviewed and tested by Ashutosh Bapat and Rajkumar Raghuwanshi. Discussion:
- Defer creation of partially-grouped relation until it's needed. This avoids
unnecessarily creating a RelOptInfo for which we have no actual need. This
idea is from Ashutosh Bapat, who wrote a very different patch to accomplish a
similar goal. It will be more important if and when we get partition-wise
aggregate, since then there could be many partially grouped relations all of
which could potentially be unnecessary. In passing, this sets the grouping
relation's reltarget, which wasn't done previously but makes things simpler
for this refactoring. Along the way, adjust things so that
add_paths_to_partial_grouping_rel, now renamed create_partial_grouping_paths,
does not perform the Gather or Gather Merge steps to generate non-partial
paths from partial paths; have the caller do it instead. This is again for
the convenience of partition-wise aggregate, which wants to inject additional
partial paths are created and before we decide which ones to Gather/Gather
Merge. This might seem like a separate change, but it's actually pretty
closely entangled; I couldn't really see much value in separating it and
having to change some things twice. Patch by me, reviewed by Ashutosh Bapat.
- Determine grouping strategies in create_grouping_paths. Partition-wise
aggregate will call create_ordinary_grouping_paths multiple times and we don't
want to redo this work every time; have the caller do it instead and pass the
details down. Patch by me, reviewed by Ashutosh Bapat. Discussion:
- Don't pass the grouping target around unnecessarily. Since commit
4f15e5d09de276fb77326be5567dd9796008ca2e made grouped_rel set reltarget, a
variety of other functions can just get it from grouped_rel instead of having
to pass it around explicitly. Simplify accordingly. Patch by me, reviewed by
Ashutosh Bapat. Discussion:
- Call pgstat_report_activity() in parallel CREATE INDEX workers. Also set
debug_query_string. Oversight in commit
9da0cc35284bdbe8d442d732963303ff0e0a40bc Peter Geoghegan, per a report by Phil
- doc: Update parallel join documentation for Parallel Shared Hash. Thomas
- Fix typo in comment. Michael Paquier Discussion:
- Avoid creating a TOAST table for a partitioned table. It's useless. Amit
- Consider Parallel Append of partial paths for UNION [ALL]. Without this
patch, we can implement a UNION or UNION ALL as an Append where Gather appears
beneath one or more of the Append branches, but this lets us put the Gather
node on top, with a partial path for each relation underneath. There is
considerably more work that could be done to improve planning in this area,
but that will probably need to wait for a future release. Patch by me,
reviewed and tested by Ashutosh Bapat and Rajkumar Raghuwanshi. Discussion:
- Implement partition-wise grouping/aggregation. If the partition keys of input
relation are part of the GROUP BY clause, all the rows belonging to a given
group come from a single partition. This allows aggregation/grouping over a
partitioned relation to be broken down * into aggregation/grouping on each
partition. This should be no worse, and often better, than the normal
approach. If the GROUP BY clause does not contain all the partition keys, we
can still perform partial aggregation for each partition and then finalize
aggregation after appending the partial results. This is less certain to be a
win, but it's still useful. Jeevan Chalke, Ashutosh Bapat, Robert Haas. The
larger patch series of which this patch is a part was also reviewed and tested
by Antonin Houska, Rajkumar Raghuwanshi, David Rowley, Dilip Kumar, Konstantin
Knizhnik, Pascal Legrand, and Rafia Sabih. Discussion:
Álvaro Herrera pushed:
- Fix state reversal after partition tuple routing. We make some changes to
ModifyTableState and the EState it uses whenever we route tuples to
partitions; but we weren't restoring properly in all cases, possibly causing
crashes when partitions with different tuple descriptors are targeted by
tuples inserted in the same command. Refactor some code, creating
ExecPrepareTupleRouting, to encapsulate the needed state changing logic, and
have it invoked one level above its current place (ie. put it in
ExecModifyTable instead of ExecInsert); this makes it all more readable. Add
a test case to exercise this. We don't support having views as partitions;
and since only views can have INSTEAD OF triggers, there is no point in
testing for INSTEAD OF when processing insertions into a partitioned table.
Remove code that appears to support this (but which is actually never
relevant.) In passing, fix location of some very confusing comments in
ModifyTableState. Reported-by: Amit Langote Author: Etsuro Fujita, Amit
- Expand comment a little bit. The previous commit removed a comment that was a
bit more verbose than its replacement.
- Remove unnecessary members from ModifyTableState and ExecInsert. These values
can be obtained from the ModifyTable node which is already a part of both the
ModifyTableState and ExecInsert. Author: Álvaro Herrera, Amit Langote
Reviewed-by: Peter Geoghegan Discussion:
- Fix CommandCounterIncrement in partition-related DDL. It makes sense to do
the CCIs in the places that do catalog updates, rather than before the places
that error out because the former ones fail to do it. In particular, it looks
like StorePartitionBound() and IndexSetParentIndex() ought to make their own
CCIs. Per review comments from Peter Eisentraut for row-level triggers on
partitioned tables. Discussion:
- Fix relcache handling of the 'default' partition. My commit 4dba331cb3dc that
moved around CommandCounterIncrement calls in partitioning DDL code unearthed
a problem with the relcache handling for the 'default' partition: the
construction of a correct relcache entry for the partitioned table was at the
mercy of lack of CCI calls in non-trivial amounts of code. This was prone to
creating problems later on, as the code develops. This was visible as a test
failure in a compile with RELCACHE_FORCE_RELASE (buildfarm member prion). The
problem is that after the mentioned commit it was possible to create a
relcache entry that had incomplete information regarding the default partition
because I introduced a CCI between adding the catalog entries for the default
partition (StorePartitionBound) and the update of pg_partitioned_table entry
for its parent partitioned table (update_default_partition_oid). It seems the
best fix is to move the latter so that it occurs inside the former; the
purposeful lack of intervening CCI should be more obvious, and harder to
break. I also remove a check in RelationBuildPartitionDesc that returns NULL
if the key is not set. I couldn't find any place that needs this hack
anymore; probably it was required because of bugs that have since been fixed.
Fix a few typos I noticed while reviewing the code involved. Discussion:
- Allow FOR EACH ROW triggers on partitioned tables. Previously, FOR EACH ROW
triggers were not allowed in partitioned tables. Now we allow AFTER triggers
on them, and on trigger creation we cascade to create an identical trigger in
each partition. We also clone the triggers to each partition that is created
or attached later. This means that deferred unique keys are allowed on
partitioned tables, too. Author: Álvaro Herrera Reviewed-by: Peter
Eisentraut, Simon Riggs, Amit Langote, Robert Haas, Thomas Munro Discussion:
Tom Lane pushed:
- Fix performance hazard in REFRESH MATERIALIZED VIEW CONCURRENTLY. Jeff Janes
discovered that commit 7ca25b7de made one of the queries run by REFRESH
MATERIALIZED VIEW CONCURRENTLY perform badly. The root cause is bad
cardinality estimation for correlated quals, but a principled solution to that
problem is some way off, especially since the planner lacks any statistics
about whole-row variables. Moreover, in non-error cases this query produces
no rows, meaning it must be run to completion; but use of LIMIT 1 encourages
the planner to pick a fast-start, slow-completion plan, exactly not what we
want. Remove the LIMIT clause, and instead rely on the count parameter we
pass to SPI_execute() to prevent excess work if the query does return some
rows. While we've heard no field reports of planner misbehavior with this
query, it could be that people are having performance issues that haven't
reached the level of pain needed to cause a bug report. In any case, that
LIMIT clause can't possibly do anything helpful with any existing version of
the planner, and it demonstrably can cause bad choices in some cases, so
back-patch to 9.4 where the code was introduced. Thomas Munro Discussion:
- Fix some corner-case issues in REFRESH MATERIALIZED VIEW CONCURRENTLY.
refresh_by_match_merge() has some issues in the way it builds a SQL query to
construct the "diff" table: 1. It doesn't require the selected unique
index(es) to be indimmediate. 2. It doesn't pay attention to the particular
equality semantics enforced by a given index, but just assumes that they must
be those of the column datatype's default btree opclass. 3. It doesn't check
that the indexes are btrees. 4. It's insufficiently careful to ensure that
the parser will pick the intended operator when parsing the query. (This
would have been a security bug before CVE-2018-1058.) 5. It's not careful
about indexes on system columns. The way to fix #4 is to make use of the
existing code in ri_triggers.c for generating an arbitrary binary operator
clause. I chose to move that to ruleutils.c, since that seems a more
reasonable place to be exporting such functionality from than ri_triggers.c.
While #1, #3, and #5 are just latent given existing feature restrictions, and
#2 doesn't arise in the core system for lack of alternate opclasses with
different equality behaviors, #4 seems like an issue worth back-patching.
That's the bulk of the change anyway, so just back-patch the whole thing to
9.4 where this code was introduced. Discussion:
- Prevent query-lifespan memory leakage of SP-GiST traversal values. The
original coding of the SP-GiST scan traversalValue feature (commit ccd6eb49a)
arranged for traversal values to be stored in the query's main executor
context. That's fine if there's only one index scan per query, but if there
are many, we have a memory leak as successive scans create new traversal
values. Fix it by creating a separate memory context for traversal values,
which we can reset during spgrescan(). Back-patch to 9.6 where this code was
introduced. In principle, adding the traversalCxt field to
SpGistScanOpaqueData creates an ABI break in the back branches. But I (tgl)
have little sympathy for extensions including spgist_private.h, so I'm not
very worried about that. Alternatively we could stick the new field at the
end of the struct in back branches, but that has its own downsides. Anton
Dignös, reviewed by Alexander Kuzmenkov Discussion:
- Doc: typo fix, "PG_" should be "TG_" here. Too much PG on the brain in commit
769159fd3, evidently. Noted by marcelhuberfoo(at)gmail(dot)com(dot) Discussion:
- Make configure check for a couple more Perl modules for --enable-tap-tests.
Red Hat's notion of a basic Perl installation doesn't include Test::More or
Time::HiRes, and reportedly some Debian installs also omit Time::HiRes. Check
for those during configure to spare the user the pain of digging through
check-world output to find out what went wrong. While we're at it, we should
also check the version of Test::More, since TestLib.pm requires at least 0.87.
In principle this could be back-patched, but it's probably not necessary.
- Change oddly-chosen OID allocation. I noticed while fooling with John
Naylor's bootstrap-data patch that we had one high-numbered manually assigned
OID, 8888, which evidently came from a submission that the committer didn't
bother to bring into line with usual OID allocation practices before
committing. That's a bad idea, because it creates a hazard for other patches
that may be temporarily using high OID numbers. Change it to something more
in line with what we usually do. This evidently dates to commit abb173392.
It's too late to change it in released branches, but we can fix it in HEAD.
- Improve predtest.c's handling of cases with NULL-constant inputs. Currently,
if operator_predicate_proof() is given an operator clause like "something op
NULL", it just throws up its hands and reports it can't prove anything. But
we can often do better than that, if the operator is strict, because then we
know that the clause returns NULL overall. Depending on whether we're trying
to prove or refute something, and whether we need weak or strong semantics for
NULL, this may be enough to prove the implication, especially when we rely on
the standard rule that "false implies anything". In particular, this lets us
do something useful with questions like "does X IN (1,3,5,NULL) imply X <= 5?"
The null entry in the IN list can effectively be ignored for this purpose, but
the proof rules were not previously smart enough to deduce that. This patch
is by me, but it owes something to previous work by Amit Langote to try to
solve problems of the form mentioned. Thanks also to Emre Hasegeli and
Ashutosh Bapat for review. Discussion:
- Fix mishandling of quoted-list GUC values in pg_dump and ruleutils.c. Code
that prints out the contents of setconfig or proconfig arrays in SQL format
needs to handle GUC_LIST_QUOTE variables differently from other ones, because
for those variables, flatten_set_variable_args() already applied a layer of
quoting. The value can therefore safely be printed as-is, and indeed must be,
or flatten_set_variable_args() will muck it up completely on reload. For all
other GUC variables, it's necessary and sufficient to quote the value as a SQL
literal. We'd recognized the need for this long ago, but mis-analyzed the
need slightly, thinking that all GUC_LIST_INPUT variables needed the special
treatment. That's actually wrong, since a valid value of a LIST variable
might include characters that need quoting, although no existing variables
accept such values. More to the point, we hadn't made any particular effort
to keep the various places that deal with this up-to-date with the set of
variables that actually need special treatment, meaning that we'd do the wrong
thing with, for example, temp_tablespaces values. This affects dumping of SET
clauses attached to functions, as well as ALTER DATABASE/ROLE SET commands.
In ruleutils.c we can fix it reasonably honestly by exporting a guc.c function
that allows discovering the flags for a given GUC variable. But pg_dump
doesn't have easy access to that, so continue the old method of having a
hard-wired list of affected variable names. At least we can fix it to have
just one list not two, and update the list to match current reality. A
remaining problem with this is that it only works for built-in GUC variables.
pg_dump's list obvious knows nothing of third-party extensions, and even the
"ask guc.c" method isn't bulletproof since the relevant extension might not be
loaded. There's no obvious solution to that, so for now, we'll just have to
discourage extension authors from inventing custom GUCs that need
GUC_LIST_QUOTE. This has been busted for a long time, so back-patch to all
supported branches. Michael Paquier and Tom Lane, reviewed by Kyotaro
Horiguchi and Pavel Stehule Discussion:
- Prevent extensions from creating custom GUCs that are GUC_LIST_QUOTE. Pending
some solution for the problems noted in commit 742869946, disallow dynamic
creation of GUC_LIST_QUOTE variables. If there are any extensions out there
using this feature, they'd not be happy for us to start enforcing this rule in
minor releases, so this is a HEAD-only change. The previous commit didn't
make things any worse than they already were for such cases. Discussion:
- Fix errors in contrib/bloom index build. Count the number of tuples in the
index honestly, instead of assuming that it's the same as the number of tuples
in the heap. (It might be different if the index is partial.) Fix counting of
tuples in current index page, too. This error would have led to failing to
write out the final page of the index if it contained exactly one tuple, so
that the last tuple of the relation would not get indexed. Back-patch to 9.6
where contrib/bloom was added. Tomas Vondra and Tom Lane Discussion:
- Fix tuple counting in SP-GiST index build. Count the number of tuples in the
index honestly, instead of assuming that it's the same as the number of tuples
in the heap. (It might be different if the index is partial.) Back-patch to
all supported versions. Tomas Vondra Discussion:
- Sync up our various ways of estimating pg_class.reltuples. VACUUM thought
that reltuples represents the total number of tuples in the relation, while
ANALYZE counted only live tuples. This can cause "flapping" in the value when
background vacuums and analyzes happen separately. The planner's use of
reltuples essentially assumes that it's the count of live (visible) tuples, so
let's standardize on having it mean live tuples. Another issue is that the
definition of "live tuple" isn't totally clear; what should be done with
INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this
regard are made on the assumption that if the originating transaction commits
at all, it will happen after ANALYZE finishes, so we should ignore the effects
of the in-progress transaction --- unless it is our own transaction, and then
we should count it. Let's propagate this definition into VACUUM, too.
Likewise propagate this definition into CREATE INDEX, and into
contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed
by Haribabu Kommi, some corrections by me Discussion:
- Improve style guideline compliance of assorted error-report messages. Per the
project style guide, details and hints should have leading capitalization and
end with a period. On the other hand, errcontext should not be capitalized
and should not end with a period. To support well formatted error contexts in
dblink, extend dblink_res_error() to take a format+arguments rather than a
hardcoded string. Daniel Gustafsson Discussion:
- Fix make rules that generate multiple output files. For years, our makefiles
have correctly observed that "there is no correct way to write a rule that
generates two files". However, what we did is to provide empty rules that
"generate" the secondary output files from the primary one, and that's not
right either. Depending on the details of the creating process, the primary
file might end up timestamped later than one or more secondary files, causing
subsequent make runs to consider the secondary file(s) out of date. That's
harmless in a plain build, since make will just re-execute the empty rule and
nothing happens. But it's fatal in a VPATH build, since make will expect the
secondary file to be rebuilt in the build directory. This would manifest as
"file not found" failures during VPATH builds from tarballs, if we were ever
unlucky enough to ship a tarball with apparently out-of-date secondary files.
(It's not clear whether that has ever actually happened, but it definitely
could.) To ensure that secondary output files have timestamps >= their
primary's, change our makefile convention to be that we provide a "touch $@"
action not an empty rule. Also, make sure that this rule actually gets
invoked during a distprep run, else the hazard remains. It's been like this a
long time, so back-patch to all supported branches. In HEAD, I skipped the
changes in src/backend/catalog/Makefile, because those rules are due to get
replaced soon in the bootstrap data format patch, and there seems no need to
create a merge issue for that patch. If for some reason we fail to land that
patch in v11, we'll need to back-fill the changes in that one makefile from
v10. Discussion: https://firstname.lastname@example.org
- Mop-up for commit feb8254518752b2cb4a8964c374dd82d49ef0e0d. Missed these
occurrences of some of the adjusted error messages. Per buildfarm member
- Stabilize regression test result. If random() returns a result sufficiently
close to zero, float8out switches to scientific notation, breaking this test
case's expectation that the output should look like '0.xxxxxxxxx'. Casting to
numeric should fix that. Per buildfarm member pogona. Discussion:
- Add #includes missed in commit e22b27f0cb3ee03ee300d431997f5944ccf2d7b3.
Leaving out getopt_long.h works on some platforms, but not all. Per
- Doc: remove extra comma in syntax summary for array_fill(). Noted by Scott
Ure. Back-patch to all supported branches. Discussion:
- Remove useless if-test. Coverity complained that this check is pointless, and
it's right. There is no case where we'd call ExecutorStart with a null
plannedstmt, and if we did, it'd have crashed before here. Thinko in commit
Andrew Dunstan pushed:
- Don't use an Msys virtual path to create a tablespace. The new
unlogged_reinit recovery tests create a new tablespace using TestLib.pm's
tempdir. However, on msys that function returns a virtual path that isn't
understood by Postgres. Here we add a new function to TestLib.pm to turn such
a path into a real path on the underlying file system, and use it in the new
test to create the tablespace. The new function is essentially a NOOP
everywhere but msys.
Peter Eisentraut pushed:
- Add missing break.
- Attempt to fix build with unusual OpenSSL versions. Since
e3bdb2d92600ed45bd46aaf48309a436a9628218, libpq failed to build on some
platforms because they did not have SSL_clear_options(). Although mainline
OpenSSL introduced SSL_clear_options() after SSL_OP_NO_COMPRESSION, so the
code should have built fine, at least an old NetBSD version (build farm
"coypu" NetBSD 5.1 gcc 4.1.3 PR-20080704 powerpc) has SSL_OP_NO_COMPRESSION
but no SSL_clear_options(). So add a configure check for SSL_clear_options().
If we don't find it, skip the call. That means on such a platform one cannot
*enable* SSL compression if the built-in default is off, but that seems an
unlikely combination anyway and not very interesting in practice.
- doc: Small wording improvement.
- Add configure tests for stdbool.h and sizeof bool. This will allow us to
assess how many platforms have bool with a size other than 1, which will help
us decide how to go forward with using stdbool.h. Discussion:
- Handle heap rewrites even better in logical decoding. Logical decoding should
not publish anything about tables created as part of a heap rewrite during
DDL. Those tables don't exist externally, so consumers of logical decoding
cannot do anything sensible with that information. In
ab28feae2bd3d4629bd73ae3548e671c57d785f0, we worked around this for built-in
logical replication, but that was hack. This is a more proper fix: We mark
such transient heaps using the new field pg_class.relwrite, linking to the
original relation OID. By default, we ignore them in logical decoding before
they get to the output plugin. Optionally, a plugin can register their
interest in getting such changes, if they handle DDL specially, in which case
the new field will help them get information about the actual table.
Reviewed-by: Craig Ringer <craig(at)2ndquadrant(dot)com>
- pg_controldata: Prevent division-by-zero errors. If the control file is
corrupted and specifies the WAL segment size to be 0 bytes, calculating the
latest checkpoint's REDO WAL file will fail with a division-by-zero error.
Show it as "???" instead. Also reword the warning message a bit and send it
to stdout, like the other pre-existing warning messages. Add some tests for
dealing with a corrupted pg_control file. Author: Nathan Bossart
<bossartn(at)amazon(dot)com>, tests by me
- Remove stdbool workaround in sepgsql. Since we now use stdbool.h in c.h, this
workaround breaks the build and is no longer necessary, so remove it.
(Technically, there could be platforms with a 4-byte bool in stdbool.h, in
which case we would not include stdbool.h in c.h, and so the old problem that
caused this workaround would reappear. But this combination is not known to
happen on the range of platforms where sepgsql can be built.)
- Use stdbool.h if suitable. Using the standard bool type provided by C allows
some recent compilers and debuggers to give better diagnostics. Also, some
extension code and third-party headers are increasingly pulling in stdbool.h,
so it's probably saner if everyone uses the same definition. But PostgreSQL
code is not prepared to handle bool of a size other than 1, so we keep our own
old definition if we encounter a stdbool.h with a bool of a different size.
(Among current build farm members, this only applies to old macOS versions on
PowerPC.) To check that the used bool is of the right size, add a static
assertions about size of GinTernaryValue vs bool. This is currently the only
place that assumes that bool and char are of the same size. Discussion:
- pg_resetwal: Add simple test suite. Some subsequent patches will add to this,
but to avoid conflicts, set up the basics separately.
- pg_resetwal: Prevent division-by-zero errors. Handle the case where the
pg_control file specifies a WAL segment size of 0 bytes. This would
previously have led to a division by zero error. Change this to assume the
whole file is corrupt and go to guess everything. Discussion:
- Further fix interaction of Perl and stdbool.h. In the case that PostgreSQL
uses stdbool.h but Perl doesn't, we need to prevent Perl from defining bool,
to prevent compiler warnings about redefinition.
- Fix interaction of Perl and stdbool.h. Revert the PL/Perl-specific change in
9a95a77d9d5d3003d2d67121f2731b6e5fc37336. We must not prevent Perl from using
stdbool.h when it has been built to do so, even if it uses an incompatible
size. Otherwise, we would be imposing our bool on Perl, which will lead to
crashes because of the size mismatch. Instead, we undef bool after including
the Perl headers, as we did previously, but now only if we are not using
stdbool.h ourselves. Record that choice in c.h as USE_STDBOOL. This will
also make it easier to apply that coding pattern elsewhere if necessary.
- Small refactoring. Put the "atomic" argument of ExecuteDoStmt() and
ExecuteCallStmt() into a variable instead of repeating the formula.
- initdb: Improve --wal-segsize handling. Give separate error messages for when
the argument is not a number and when it is not the right kind of number. Fix
wording in the help message.
- Improve pg_resetwal documentation. Clarify that the -l option takes a file
name, not an "address", and that that might be different from the LSN if
nondefault WAL segment sizes are used.
- Add long options to pg_resetwal and pg_controldata. We were running out of
good single-letter options for some upcoming pg_resetwal functionality, so add
long options to create more possibilities. Add to pg_controldata as well for
symmetry. based on patch by Bossart, Nathan <bossartn(at)amazon(dot)com>
- pg_resetwal: Fix logical typo in code. introduced in
Andres Freund pushed:
- Add PGAC_PROG_VARCC_VARFLAGS_OPT autoconf macro. The new macro allows to test
flags for different compilers and to store them in different CFLAG like
variables. The existing PGAC_PROG_CC_CFLAGS_OPT and PGAC_PROG_CC_VAR_OPT are
changed to be just wrappers around the new function. This'll be used by the
upcoming LLVM support, to separately detect capabilities used by clang, when
generating bitcode. Author: Andres Freund Discussion:
- Add C++ support to configure. This is an optional dependency. It'll be used
for the upcoming LLVM based just in time compilation support, which needs to
wrap a few LLVM C++ APIs so they're accessible from C.. For now test for C++
compilers unconditionally, without failing if not present, to ensure wide
buildfarm coverage. If we're bothered by the additional test times (which are
quite short) or verbosity, we can later make the tests conditional on
--with-llvm. Author: Andres Freund Discussion:
- Add configure infrastructure (--with-llvm) to enable LLVM support. LLVM will
be used for *optional* Just-in-time compilation support. This commit just adds
the configure infrastructure that detects LLVM. No documentation has been
added for the --with-llvm flag, that'll be added after the actual supporting
code has been added. Author: Andres Freund Discussion:
- Handle EEOP_FUNCEXPR_[STRICT_]FUSAGE out of line. This isn't a very common
op, and it doesn't seem worth duplicating for JIT. Author: Andres Freund
- Basic JIT provider and error handling infrastructure. This commit introduces:
1) JIT provider abstraction, which allows JIT functionality to be implemented
in separate shared libraries. That's desirable because it allows to install
JIT support as a separate package, and because it allows experimentation with
different forms of JITing. 2) JITContexts which can be, using functions
introduced in follow up commits, used to emit JITed functions, and have them
be cleaned up on error. 3) The outline of a LLVM JIT provider, which will be
fleshed out in subsequent commits. Documentation for GUCs added, and for JIT
in general, will be added in later commits. Author: Andres Freund, with
architectural input from Jeff Davis Discussion:
- Fix typo in BITCODE_CXXFLAGS assignment. Typoed-In: 5b2526c83832e
Reported-By: Catalin Iacob
- Empty CXXFLAGS inherited from autoconf. We do the same for CFLAGS. This was
an omission in 6869b4f25. Reported-By: Catalin Iacob
- Add file containing extensions of the LLVM C API. Author: Andres Freund
- Support for optimizing and emitting code in LLVM JIT provider. This commit
introduces the ability to actually generate code using LLVM. In particular,
this adds: - Ability to emit code both in heavily optimized and largely
unoptimized fashion - Batching facility to allow functions to be defined in
small increments, but optimized and emitted in executable form in larger
batches (for performance and memory efficiency) - Type and function
declaration synchronization between runtime generated code and normal postgres
code. This is critical to be able to access struct fields etc. - Developer
oriented jit_dump_bitcode GUC, for inspecting / debugging the generated code.
- per JitContext statistics of number of functions, time spent generating
code, optimizing, and emitting it. This will later be employed for EXPLAIN
support. This commit doesn't yet contain any code actually generating
functions. That'll follow in later commits. Documentation for GUCs added, and
for JIT in general, will be added in later commits. Author: Andres Freund,
with contributions by Pierre Ducroquet Testing-By: Thomas Munro, Peter
- Add helpers for emitting LLVM IR. These basically just help to make code a
bit more concise and pgindent proof. Author: Andres Freund Discussion:
- Basic planner and executor integration for JIT. This adds simple cost based
plan time decision about whether JIT should be performed. jit_above_cost,
jit_optimize_above_cost are compared with the total cost of a plan, and if the
cost is above them JIT is performed / optimization is performed respectively.
For that PlannedStmt and EState have a jitFlags (es_jit_flags) field that
stores information about what JIT operations should be performed. EState now
also has a new es_jit field, which can store a JitContext. When there are no
errors the context is released in standard_ExecutorEnd(). It is likely that
the default values for jit_[optimize_]above_cost will need to be adapted
further, but in my test these values seem to work reasonably. Author: Andres
Freund, with feedback by Peter Eisentraut Discussion:
- Add FIELDNO_* macro designating offset into structs required for JIT. For any
interesting JIT target, fields inside structs need to be accessed. b96d550e
contains infrastructure for syncing the definition of types between postgres C
code and runtime code generation with LLVM. But that doesn't sync the number
or names of fields inside structs, just the types (including padding etc).
One option would be to hardcode the offset numbers in the JIT code, but that'd
be hard to keep in sync. Instead add macros indicating the field offset to the
fields that need to be accessed. Not pretty, but manageable. Author: Andres
- Expand list of synchronized types and functions in LLVM JIT provider. Author:
Andres Freund Discussion:
- Add expression compilation support to LLVM JIT provider. In addition to the
interpretation of expressions (which back evaluation of WHERE clauses, target
list projection, aggregates transition values etc) support compiling
expressions to native code, using the infrastructure added in earlier commits.
To avoid duplicating a lot of code, only support emitting code for cases that
are likely to be performance critical. For expression steps that aren't deemed
that, use the existing interpreter. The generated code isn't great - some
architectural changes are required to address that. But this already yields a
significant speedup for some analytics queries, particularly with WHERE
clauses filtering a lot, or computing multiple aggregates. Author: Andres
Freund Tested-By: Thomas Munro Discussion:
JITing for VALUES() nodes. VALUES() nodes are only ever executed once. This
is primarily helpful for debugging, when forcing JITing even for cheap
queries. Author: Andres Freund Discussion:
- Debugging and profiling support for LLVM JIT provider. This currently
requires patches to the LLVM codebase to be effective (submitted upstream),
the GUCs are available without those patches however. Author: Andres Freund
- Adapt expression JIT to stdbool.h introduction. The LLVM JIT provider uses
clang to synchronize types between normal C code and runtime generated code.
Clang represents stdbool.h style booleans in return values & parameters
differently from booleans stored in variables. Thus the expression
compilation code from 2a0faed9d needs to be adapted to 9a95a77d9. Instead of
hardcoding i8 as the type for booleans (which already was wrong on some edge
case platforms!), use postgres' notion of a boolean as used for storage and
for parameters. Per buildfarm animal xenodermus. Author: Andres Freund
Teodor Sigaev pushed:
- Rework word_similarity documentation, make it close to actual algorithm.
word_similarity before claimed as returning similarity of closest word in
string, but, actually it returns similarity of substring. Also fix mistyped
comments. Author: Alexander Korotkov Review by: David Steele, Liudmila
- Add strict_word_similarity to pg_trgm module. strict_word_similarity is
similar to existing word_similarity function but it takes into account word
boundaries to compute similarity. Author: Alexander Korotkov Review by: David
Steele, Liudmila Mantrova, me Discussion:
- UINT64CONST'fy long constants in pgbench. In commit
e51a04840a1c45db101686bef0b7025d5014c74b it was missed 64-bit constants, wrap
them with UINT64CONST(). Per buildfarm member dromedary and gripe from Tom
- Add conditional.c to libpgfeutils for MSVC build. conditional.c was moved in
f67b113ac62777d18cd20d3c4d05be964301b936 commit but forgotten to add to
Windows build system. I don't have a Windows box, so blind attempt.
- Add \if support to pgbench. Patch adds \if to pgbench as it done for psql.
Implementation shares condition stack code with psql, so, this code is moved
to fe_utils directory. Author: Fabien COELHO with minor editorization by me
Review by: Vik Fearing, Fedor Sigaev Discussion:
- Add general-purpose hashing functions to pgbench. Hashing function is useful
for simulating real-world workload in test like WEB workload, as an example -
YCSB benchmarks. Author: Ildar Musin with minor editorization by me Reviewed
by: Fabien Coelho, me Discussion:
- Exclude unlogged tables from base backups. Exclude unlogged tables from base
backup entirely except init fork which marks created unlogged table. The next
question is do not backup temp table but it's a story for separate patch.
Author: David Steele Review by: Adam Brightwell, Masahiko Sawada Discussion:
Andrew Gierth pushed:
- Repair crash with unsortable grouping sets. If there were multiple grouping
sets, none of them empty, all of which were unsortable, then an oversight in
consider_groupingsets_paths led to a null pointer dereference. Fix, and add a
regression test for this case. Per report from Dang Minh Huong, though I
didn't use their patch. Backpatch to 10.x where hashed grouping sets were
Tatsuo Ishii pushed:
- Fix typo. Patch by me.
Dean Rasheed pushed:
- Improve ANALYZE's strategy for finding MCVs. Previously, a value was included
in the MCV list if its frequency was 25% larger than the estimated average
frequency of all nonnull values in the table. For uniform distributions, that
can lead to values being included in the MCV list and significantly
overestimated on the basis of relatively few (sometimes just 2) instances
being seen in the sample. For non-uniform distributions, it can lead to too
few values being included in the MCV list, since the overall average frequency
may be dominated by a small number of very common values, while the remaining
values may still have a large spread of frequencies, causing both substantial
overestimation and underestimation of the remaining values. Furthermore,
increasing the statistics target may have little effect because the overall
average frequency will remain relatively unchanged. Instead, populate the MCV
list with the largest set of common values that are statistically
significantly more common than the average frequency of the remaining values.
This takes into account the variance of the sample counts, which depends on
the counts themselves and on the proportion of the table that was sampled. As
a result, it constrains the relative standard error of estimates based on the
frequencies of values in the list, reducing the chances of too many values
being included. At the same time, it allows more values to be included, since
the MCVs need only be more common than the remaining non-MCVs, rather than the
overall average. Thus it tends to produce fewer MCVs than the previous code
for uniform distributions, and more for non-uniform distributions, reducing
estimation errors in both cases. In addition, the algorithm responds better
to increasing the statistics target, allowing more values to be included in
the MCV list when more of the table is sampled. Jeff Janes, substantially
modified by me. Reviewed by John Naylor and Tomas Vondra. Discussion:
Noah Misch pushed:
- Don't qualify type pg_catalog.text in extend-extensions-example. Extension
scripts begin execution with pg_catalog at the front of the search path, so
type names reliably refer to pg_catalog. Remove these superfluous
qualifications. Earlier <programlisting> of this <sect1> already omitted
them. Back-patch to 9.3 (all supported versions).
== Pending Patches ==
Tom Lane sent in a patch to fix some issues in matview.c's refresh-query
Daniel Gustafsson and Magnus Hagander traded patches to make it possible to
enable checksums online.
Tomas Vondra sent in another revision of a patch to implement multivariate
histograms and MCV lists.
Chapman Flack sent in another revision of a patch to zero headers of unused
pages after WAL switch and add a test for ensuring WAL segment is zeroed out.
Amul Sul sent in another revision of a patch to restrict concurrent
update/delete with UPDATE of partition key.
Kyotaro HORIGUCHI sent in another revision of a patch to restrict maximum keep
segments by repslots.
Andrey Borodin sent in another revision of a patch to add SLRU checksums.
Masahiko Sawada sent in another revision of a patch to qualify datatype name in
log of data type conversion on subscriber and add a test module for same.
Masahiko Sawada sent in another revision of a patch to add a
Etsuro Fujita, Amit Langote, and Álvaro Herrera traded patches to make ON
CONFLICT .. DO UPDATE work on partitioned tables.
Amit Langote and David Rowley traded patches to lay infrastructure for speeding
up partition pruning.
Artur Zakirov and Tomas Vondra traded patches to implement shared ISpell
Pavel Stěhule sent in two more revisions of a patch to add extra checks to
Thomas Munro sent in a patch to add docs to the top-level Makefile for non-GNU
Pavel Stěhule sent in a patch to enable procedures to be called with default
arguments in PL/pgsql.
Alexander Korotkov sent in two more revisions of a patch to implement
Julian Markwort sent in two more revisions of a patch to add a plan option to
Pavel Stěhule sent in two more revisions of a patch to implement schema
Dilip Kumar sent in two more revisions of a patch to ensure that InitXLogInsert
is never called in a critical section.
Peter Eisentraut sent in another revision of a patch to use file cloning in
pg_upgrade and CREATE DATABASE.
Pavan Deolasee, Peter Geoghegan, and Amit Langote traded patches to implement
Daniel Gustafsson sent in another revision of a patch to support sending an
optional message in backend cancel/terminate.
Tomas Vondra sent in another revision of a patch to implement BRIN multi-range
indexes and BRIN Bloom indexes.
Michael Banck sent in a patch to allow setting replication slots in
recovery.conf even if wal-method is none.
Daniel Gustafsson sent in a patch to pg_basebackup to add missing newlines in
some error messages.
Konstantin Knizhnik sent in another revision of a patch to optimize secondary
Thomas Munro sent in a patch to tweak the JIT docs.
Doug Rady sent in another revision of a patch to enable building pgbench using
ppoll() for larger connection counts.
Tom Lane and John Naylor traded patches to rationalize the way bootstrap data is
Pavel Stěhule sent in a patch to enable CALL with named default arguments in PL/pgsql.
Nathan Bossart sent in a patch to combine options for RangeVarGetRelidExtended()
into a flags argument and add a skip-locked option to same.
Fabien COELHO sent in a patch to fix some constants in the new general-purpose
hashing functions for pgbench.
Amit Langote sent in another revision of a patch to refactor the partitioning
Dmitry Dolgov sent in another revision of a patch to implement generic type
subscripting and use same for arrays and JSONB.
Teodor Sigaev sent in another revision of a patch to add a prefix operator for
text with SP-GiST support.
David Rowley sent in two more revisions of a patch to remove useless DISTINCT
Fabien COELHO sent in two more revisions of a patch to implement --random-seed
David Steele and Michael Banck traded patches to verify checksums during
Pavan Deolasee sent in two more revisions of a patch to speed up inserts with
mostly-monotonically increasing values.
Daniel Vérité and Pavel Stěhule traded patches to add a csv format to psql.
Tomas Vondra sent in a patch to fix a a minor mistake in brin_inclusion.c
Ashutosh Bapat sent in a patch to fix a comment in BuildTupleFromCStrings().
Teodor Sigaev sent in another revision of a patch to implement predicate locking
on GIN indexes.
Ashutosh Bapat sent in a patch to fix an issue with procedure name resolution.
Julian Markwort sent in another revision of a patch to add a new auth option to
Simon Riggs sent in another revision of a patch to implement logical decoding of
David Steele sent in a patch to add more TAP tests to pgrewind.
Pavan Deolasee sent in another revision of a patch to change the WAL header to
reduce contention during ReserveXLogInsertLocation().
Haribabu Kommi sent in two more revisions of a patch to make PQhost return
connected host and hostaddr details.
Fabien COELHO sent in a patch to pgbench to test whether a variable exists.
Robert Haas sent in another revision of a patch to enable parallel seq scan for
Peter Eisentraut sent in another revision of a patch to enable nested CALL with
transactions in PL/pgSQL.
Fabien COELHO sent in another revision of a patch to pgbench to enable it to
store SELECT results into variables.
Tom Lane sent in a patch to help with backend memory dump analysis by adding
David Rowley sent in two more revisions of a patch to speed up execution of
ALTER TABLE ... ADD COLUMN ... DEFAULT [not null].
Michaël Paquier sent in a patch to simplify the final sync in pg_rewind's target
folder and add --no-sync.
Takayuki Tsunakawa sent in another revision of a patch to fix a problem in ECPG
where freeing memory for pgtypes crashes on Windows.
|Next Message||PostgreSQL - Hans-Jürgen Schönig||2018-03-27 08:51:57||ora_migrator: A new method to migration from Oracle to PostgreSQL|
|Previous Message||Monica Real Amores||2018-03-20 15:14:56||pglogical 2.2 Now Available|