This new DDL command splits a single partition into several parititions.
Just like ALTER TABLE ... MERGE PARTITIONS ... command, new patitions are
created using createPartitionTable() function with parent partition as the
template.
This commit comprises quite naive implementation which works in single process
and holds the ACCESS EXCLUSIVE LOCK on the parent table during all the
operations including the tuple routing. This is why this new DDL command
can't be recommended for large partitioned tables under a high load. However,
this implementation come in handy in certain cases even as is.
Also, it could be used as a foundation for future implementations with lesser
locking and possibly parallel.
Discussion: https://postgr.es/m/c73a1746-0cd0-6bdd-6b23-3ae0b7c0c582%40postgrespro.ru
Author: Dmitry Koval
Reviewed-by: Matthias van de Meent, Laurenz Albe, Zhihong Yu, Justin Pryzby
Reviewed-by: Alvaro Herrera, Robert Haas, Stephane Tachoires
This new DDL command merges several partitions into the one partition of the
target table. The target partition is created using new
createPartitionTable() function with parent partition as the template.
This commit comprises quite naive implementation which works in single process
and holds the ACCESS EXCLUSIVE LOCK on the parent table during all the
operations including the tuple routing. This is why this new DDL command
can't be recommended for large partitioned tables under a high load. However,
this implementation come in handy in certain cases even as is.
Also, it could be used as a foundation for future implementations with lesser
locking and possibly parallel.
Discussion: https://postgr.es/m/c73a1746-0cd0-6bdd-6b23-3ae0b7c0c582%40postgrespro.ru
Author: Dmitry Koval
Reviewed-by: Matthias van de Meent, Laurenz Albe, Zhihong Yu, Justin Pryzby
Reviewed-by: Alvaro Herrera, Robert Haas, Stephane Tachoires
as determined by include-what-you-use (IWYU)
While IWYU also suggests to *add* a bunch of #include's (which is its
main purpose), this patch does not do that. In some cases, a more
specific #include replaces another less specific one.
Some manual adjustments of the automatic result:
- IWYU currently doesn't know about includes that provide global
variable declarations (like -Wmissing-variable-declarations), so
those includes are being kept manually.
- All includes for port(ability) headers are being kept for now, to
play it safe.
- No changes of catalog/pg_foo.h to catalog/pg_foo_d.h, to keep the
patch from exploding in size.
Note that this patch touches just *.c files, so nothing declared in
header files changes in hidden ways.
As a small example, in src/backend/access/transam/rmgr.c, some IWYU
pragma annotations are added to handle a special case there.
Discussion: https://www.postgresql.org/message-id/flat/af837490-6b2f-46df-ba05-37ea6a6653fc%40eisentraut.org
Based on how postgres.h foes the Oid <-> Datum conversion, there is no
existing bugs but let's be consistent. 17 spots have been noticed as
incorrectly passing down Oids rather than Datums. Aleksander got one,
Zhang two and I the rest.
Author: Michael Paquier, Aleksander Alekseev, Zhang Mingli
Discussion: https://postgr.es/m/ZLUhqsqQN1MOaxdw@paquier.xyz
Run pgindent, pgperltidy, and reformat-dat-files.
This set of diffs is a bit larger than typical. We've updated to
pg_bsd_indent 2.1.2, which properly indents variable declarations that
have multi-line initialization expressions (the continuation lines are
now indented one tab stop). We've also updated to perltidy version
20230309 and changed some of its settings, which reduces its desire to
add whitespace to lines to make assignments etc. line up. Going
forward, that should make for fewer random-seeming changes to existing
code.
Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
When extracting an attr from a cached tuple in the syscache with
SysCacheGetAttr the isnull parameter must be checked in case the
attr cannot be NULL. For cases when this is known beforehand, a
wrapper is introduced which perform the errorhandling internally
on behalf of the caller, invoking an elog in case of a NULL attr.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Reviewed-by: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/AD76405E-DB45-46B6-941F-17B1EB3A9076@yesql.se
In a similar effort to f01592f91, here we mostly rename shadowed local
variables to remove the warnings produced when compiling with
-Wshadow=compatible-local.
This fixes 63 warnings and leaves just 5.
Author: Justin Pryzby, David Rowley
Reviewed-by: Justin Pryzby
Discussion https://postgr.es/m/20220817145434.GC26426%40telsasoft.com
Make sure that function declarations use names that exactly match the
corresponding names from function definitions in optimizer, parser,
utility, libpq, and "commands" code, as well as in remaining library
code. Do the same for all code related to frontend programs (with the
exception of pg_dump/pg_dumpall related code).
Like other recent commits that cleaned up function parameter names, this
commit was written with help from clang-tidy. Later commits will handle
ecpg and pg_dump/pg_dumpall.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/CAH2-WznJt9CMM9KJTMjJh_zbL5hD9oX44qdJ4aqZtjFi-zA3Tg@mail.gmail.com
guc.c has grown to be one of our largest .c files, making it
a bottleneck for compilation. It's also acquired a bunch of
knowledge that'd be better kept elsewhere, because of our not
very good habit of putting variable-specific check hooks here.
Hence, split it up along these lines:
* guc.c itself retains just the core GUC housekeeping mechanisms.
* New file guc_funcs.c contains the SET/SHOW interfaces and some
SQL-accessible functions for GUC manipulation.
* New file guc_tables.c contains the data arrays that define the
built-in GUC variables, along with some already-exported constant
tables.
* GUC check/assign/show hook functions are moved to the variable's
home module, whenever that's clearly identifiable. A few hard-
to-classify hooks ended up in commands/variable.c, which was
already a home for miscellaneous GUC hook functions.
To avoid cluttering a lot more header files with #include "guc.h",
I also invented a new header file utils/guc_hooks.h and put all
the GUC hook functions' declarations there, regardless of their
originating module. That allowed removal of #include "guc.h"
from some existing headers. The fallout from that (hopefully
all caught here) demonstrates clearly why such inclusions are
best minimized: there are a lot of files that, for example,
were getting array.h at two or more levels of remove, despite
not having any connection at all to GUCs in themselves.
There is some very minor code beautification here, such as
renaming a couple of inconsistently-named hook functions
and improving some comments. But mostly this just moves
code from point A to point B and deals with the ensuing
needs for #include adjustments and exporting a few functions
that previously weren't exported.
Patch by me, per a suggestion from Andres Freund; thanks also
to Michael Paquier for the idea to invent guc_funcs.c.
Discussion: https://postgr.es/m/587607.1662836699@sss.pgh.pa.us
When checking for interleaved partitions, we mark the partition as
interleaved when;
1. we find an earlier partition index when looping over the
sorted-by-Datum indexes[] array, or;
2. we find that the NULL partition allows some non-NULL Datum value.
In the code, as it was written in db632fbca we'll continue to check for
case 2 when we've already marked the partition as interleaved for case 1.
Here we make it so we don't bother marking the partition as interleaved
for case 2 when it's already been marked due to case 1.
Really all this saves is a useless call to bms_add_member(), but since
this code is new to PG15, it seems worth fixing it now to save anyone the
trouble of complaining at some time in the future. We have the
opportunity to improve this now before PG15 is out. This might ease some
future back-patching pain.
Per report and patch by Zhihong Yu. However, I slightly revised the
comments and altered the bms_add_member() code to match in both locations.
We already know that index is equal to boundinfo->null_index from the if
condition.
Author: Zhihong Yu
Discussion: https://postgr.es/m/CALNJ-vQbZR0pYxz9zQ5bqXVcwtGgNgVupeEpNT65HZ+yWZnc4g@mail.gmail.com
Backpatch-through: 15, same as db632fbca.
"const foo *" is quite different from "foo * const".
This code was evidently trying to avoid casting away
const from the arguments, but entirely failed to do so.
Per study of some buildfarm warnings from anole
(which unfortunately are mostly ignorable, since it
seems not to understand "restrict" very well).
I'm surprised though that nothing else has complained.
This field was recently added in db632fbca, however that commit missed one
place where it should have initialized the new field to NULL. The missed
location is where the PartitionBoundInfo is created for partition-wise
join relations. Technically there could be interleaved partitions in a
partition-wise join relation, but currently the only optimization we use
this field for only does so for base rels and other member rels. So just
document that we don't populate this field for join rels.
Reported-by: Amit Langote
Author: Amit Langote, David Rowley
Reviewed-by: Amit Langote, David Rowley
Discussion: https://postgr.es/m/CA+HiwqE76Rps24kwHsd2Cr82Ua07tJC9t9reG0c7ScX9n_xrEA@mail.gmail.com
959d00e9d added the ability to make use of an Append node instead of a
MergeAppend when we wanted to perform a scan of a partitioned table and
the required sort order was the same as the partitioned keys and the
partitioned table was defined in such a way that earlier partitions were
guaranteed to only contain lower-order values than later partitions.
However, previously we didn't allow these ordered partition scans for
LIST partitioned table when there were any partitions that allowed
multiple Datums. This was a very cheap check to make and we could likely
have done a little better by checking if there were interleaved
partitions, but at the time we didn't have visibility about which
partitions were pruned, so we still may have disallowed cases where all
interleaved partitions were pruned.
Since 475dbd0b7, we now have knowledge of pruned partitions, we can do a
much better job inside partitions_are_ordered().
Here we pass which partitions survived partition pruning into
partitions_are_ordered() and, for LIST partitioning, have it check to see
if any live partitions exist that are also in the new "interleaved_parts"
field defined in PartitionBoundInfo.
For RANGE partitioning we can relax the code which caused the partitions
to be unordered if a DEFAULT partition existed. Since we now know which
partitions were pruned, partitions_are_ordered() now returns true when the
DEFAULT partition was pruned.
Reviewed-by: Amit Langote, Zhihong Yu
Discussion: https://postgr.es/m/CAApHDvrdoN_sXU52i=QDXe2k3WAo=EVry29r2+Tq2WYcn2xhEA@mail.gmail.com
The error messages using the word "non-negative" are confusing
because it's ambiguous about whether it accepts zero or not.
This commit improves those error messages by replacing it with
less ambiguous word like "greater than zero" or
"greater than or equal to zero".
Also this commit added the note about the word "non-negative" to
the error message style guide, to help writing the new error messages.
When postgres_fdw option fetch_size was set to zero, previously
the error message "fetch_size requires a non-negative integer value"
was reported. This error message was outright buggy. Therefore
back-patch to all supported versions where such buggy error message
could be thrown.
Reported-by: Hou Zhijie
Author: Bharath Rupireddy
Reviewed-by: Kyotaro Horiguchi, Fujii Masao
Discussion: https://postgr.es/m/OS0PR01MB5716415335A06B489F1B3A8194569@OS0PR01MB5716.jpnprd01.prod.outlook.com
Commit 0563a3a8b changed how partition constraints were generated such
that this function no longer computes the mapping of parent attnos to
child attnos.
This is an external function that extensions could use, so this is
potentially a breaking change. No external callers are known, however,
and this will make it simpler to write such callers in the future.
Author: Hou Zhijie
Reviewed-by: David Rowley, Michael Paquier, Soumyadeep Chakraborty
Discussion: https://www.postgresql.org/message-id/flat/OS0PR01MB5716A75A45BE46101A1B489894379@OS0PR01MB5716.jpnprd01.prod.outlook.com
Similar to 50e17ad28, which allowed hash tables to be used for IN clauses
with a set of constants, here we add the same feature for NOT IN clauses.
NOT IN evaluates the same as: WHERE a <> v1 AND a <> v2 AND a <> v3.
Obviously, if we're using a hash table we must be exactly equivalent to
that and return the same result taking into account that either side of
the condition could contain a NULL. This requires a little bit of
special handling to make work with the hash table version.
When processing NOT IN, the ScalarArrayOpExpr's operator will be the <>
operator. To be able to build and lookup a hash table we must use the
<>'s negator operator. The planner checks if that exists and is hashable
and sets the relevant fields in ScalarArrayOpExpr to instruct the executor
to use hashing.
Author: David Rowley, James Coleman
Reviewed-by: James Coleman, Zhihong Yu
Discussion: https://postgr.es/m/CAApHDvoF1mum_FRk6D621edcB6KSHBi2+GAgWmioj5AhOu2vwQ@mail.gmail.com
In each of the create_*_bound() functions for LIST, RANGE and HASH
partitioning, there were a large number of palloc calls which could be
reduced down to a much smaller number.
In each of these functions, an array was built so that we could qsort it
before making the PartitionBoundInfo. For LIST and HASH partitioning, an
array of pointers was allocated then each element was allocated within
that array. Since the number of items of each dimension is known
beforehand, we can just allocate a single chunk of memory for this.
Similarly, with all partition strategies, we're able to reduce the number
of allocations to build the ->datums field. This is an array of Datum
pointers, but there's no need for the Datums that each element points to
to be singly allocated. One big chunk will do. For RANGE partitioning,
the PartitionBoundInfo->kind field can get the same treatment.
We can apply the same optimizations to partition_bounds_copy(). Doing
this might have a small effect on cache performance when searching for the
correct partition during partition pruning or DML on a partitioned table.
However, that's likely to be small and this is mostly about reducing
palloc overhead.
Author: Nitin Jadhav, Justin Pryzby, David Rowley
Reviewed-by: Justin Pryzby, Zhihong Yu
Discussion: https://postgr.es/m/flat/CAMm1aWYFTqEio3bURzZh47jveiHRwgQTiSDvBORczNEz2duZ1Q@mail.gmail.com
Commit efbfb6424 added logic for reporting exactly which existing
partition conflicts when complaining that a new hash partition's
modulus isn't compatible with the existing ones. However, it
misunderstood the partitioning data structure, and would select
the wrong partition in some cases, or crash outright due to fetching
a bogus table OID in other cases.
Per bug #17076 from Alexander Lakhin. Fix by Amit Langote;
some further work on the code comments by me.
Discussion: https://postgr.es/m/17076-89a16ae835d329b9@postgresql.org
Also "make reformat-dat-files".
The only change worthy of note is that pgindent messed up the formatting
of launcher.c's struct LogicalRepWorkerId, which led me to notice that
that struct wasn't used at all anymore, so I just took it out.
During queries coming from ri_triggers.c, we need to omit partitions
that are marked pending detach -- otherwise, the RI query is tricked
into allowing a row into the referencing table whose corresponding row
is in the detached partition. Which is bogus: once the detach operation
completes, the row becomes an orphan.
However, the code was not doing that in repeatable-read transactions,
because relcache kept a copy of the partition descriptor that included
the partition, and used it in the RI query. This commit changes the
partdesc cache code to only keep descriptors that aren't dependent on
a snapshot (namely: those where no detached partition exist, and those
where detached partitions are included). When a partdesc-without-
detached-partitions is requested, we create one afresh each time; also,
those partdescs are stored in PortalContext instead of
CacheMemoryContext.
find_inheritance_children gets a new output *detached_exist boolean,
which indicates whether any partition marked pending-detach is found.
Its "include_detached" input flag is changed to "omit_detached", because
that name captures desired the semantics more naturally.
CreatePartitionDirectory() and RelationGetPartitionDesc() arguments are
identically renamed.
This was noticed because a buildfarm member that runs with relcache
clobbering, which would not keep the improperly cached partdesc, broke
one test, which led us to realize that the expected output of that test
was bogus. This commit also corrects that expected output.
Author: Amit Langote <amitlangote09@gmail.com>
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m/3269784.1617215412@sss.pgh.pa.us
ScalarArrayOpExprs with "useOr=true" and a set of Consts on the righthand
side have traditionally been evaluated by using a linear search over the
array. When these arrays contain large numbers of elements then this
linear search could become a significant part of execution time.
Here we add a new method of evaluating ScalarArrayOpExpr expressions to
allow them to be evaluated by first building a hash table containing each
element, then on subsequent evaluations, we just probe that hash table to
determine if there is a match.
The planner is in charge of determining when this optimization is possible
and it enables it by setting hashfuncid in the ScalarArrayOpExpr. The
executor will only perform the hash table evaluation when the hashfuncid
is set.
This means that not all cases are optimized. For example CHECK constraints
containing an IN clause won't go through the planner, so won't get the
hashfuncid set. We could maybe do something about that at some later
date. The reason we're not doing it now is from fear that we may slow
down cases where the expression is evaluated only once. Those cases can
be common, for example, a single row INSERT to a table with a CHECK
constraint containing an IN clause.
In the planner, we enable this when there are suitable hash functions for
the ScalarArrayOpExpr's operator and only when there is at least
MIN_ARRAY_SIZE_FOR_HASHED_SAOP elements in the array. The threshold is
currently set to 9.
Author: James Coleman, David Rowley
Reviewed-by: David Rowley, Tomas Vondra, Heikki Linnakangas
Discussion: https://postgr.es/m/CAAaqYe8x62+=wn0zvNKCj55tPpg-JBHzhZFFc6ANovdqFw7-dA@mail.gmail.com
There was some code in gen_prune_steps_from_opexps that needlessly
checked a list was not empty when it clearly had to contain at least one
item. This prompted a further cleanup operation in partprune.c.
Additionally, the previous code could end up adding additional needless
INTERSECT steps. However, those do not appear to be able to cause any
misbehavior.
gen_prune_steps_from_opexps is now no longer in charge of generating
combine pruning steps. Instead, gen_partprune_steps_internal, which
already does some combine step creation has been given the sole
responsibility of generating all combine steps. This means that when
we recursively call gen_partprune_steps_internal, since it always now adds
a combine step when it produces multiple steps, we can just pay attention
to the final step returned.
In passing, do quite a bit of work on the comments to try to more clearly
explain the role of both gen_partprune_steps_internal and
gen_prune_steps_from_opexps. This is fairly complex code so some extra
effort to give any new readers an overview of how things work seems like
a good idea.
Author: Amit Langote
Reported-by: Andy Fan
Reviewed-by: Kyotaro Horiguchi, Andy Fan, Ryan Lambert, David Rowley
Discussion: https://postgr.es/m/CAKU4AWqWoVii+bRTeBQmeVW+PznkdO8DfbwqNsu9Gj4ubt9A6w@mail.gmail.com
Allow a partition be detached from its partitioned table without
blocking concurrent queries, by running in two transactions and only
requiring ShareUpdateExclusive in the partitioned table.
Because it runs in two transactions, it cannot be used in a transaction
block. This is the main reason to use dedicated syntax: so that users
can choose to use the original mode if they need it. But also, it
doesn't work when a default partition exists (because an exclusive lock
would still need to be obtained on it, in order to change its partition
constraint.)
In case the second transaction is cancelled or a crash occurs, there's
ALTER TABLE .. DETACH PARTITION .. FINALIZE, which executes the final
steps.
The main trick to make this work is the addition of column
pg_inherits.inhdetachpending, initially false; can only be set true in
the first part of this command. Once that is committed, concurrent
transactions that use a PartitionDirectory will include or ignore
partitions so marked: in optimizer they are ignored if the row is marked
committed for the snapshot; in executor they are always included. As a
result, and because of the way PartitionDirectory caches partition
descriptors, queries that were planned before the detach will see the
rows in the detached partition and queries that are planned after the
detach, won't.
A CHECK constraint is created that duplicates the partition constraint.
This is probably not strictly necessary, and some users will prefer to
remove it afterwards, but if the partition is re-attached to a
partitioned table, the constraint needn't be rechecked.
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Reviewed-by: Amit Langote <amitlangote09@gmail.com>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Discussion: https://postgr.es/m/20200803234854.GA24158@alvherre.pgsql
For the error message "every hash partition modulus must be a factor
of the next larger modulus", add a detail message that shows the
particular numbers and existing partition involved. Also comment the
code more.
Reviewed-by: Amit Langote <amitlangote09@gmail.com>
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Discussion: https://www.postgresql.org/message-id/flat/bb9d60b4-aadb-607a-1a9d-fdc3434dddcd%40enterprisedb.com
perform_pruning_combine_step() was not taught about the number of
partition indexes used in hash partitioning; more embarrassingly,
get_matching_hash_bounds() also had it wrong. These errors are masked
in the common case where all the partitions have the same modulus
and no partition is missing. However, with missing or unequal-size
partitions, we could erroneously prune some partitions that need
to be scanned, leading to silently wrong query answers.
While a minimal-footprint fix for this could be to export
get_partition_bound_num_indexes and make the incorrect functions use it,
I'm of the opinion that that function should never have existed in the
first place. It's not reasonable data structure design that
PartitionBoundInfoData lacks any explicit record of the length of
its indexes[] array. Perhaps that was all right when it could always
be assumed equal to ndatums, but something should have been done about
it as soon as that stopped being true. Putting in an explicit
"nindexes" field makes both partition_bounds_equal() and
partition_bounds_copy() simpler, safer, and faster than before,
and removes explicit knowledge of the number-of-partition-indexes
rules from some other places too.
This change also makes get_hash_partition_greatest_modulus obsolete.
I left that in place in case any external code uses it, but no core
code does anymore.
Per bug #16840 from Michał Albrycht. Back-patch to v11 where the
hash partitioning code came in. (In the back branches, add the new
field at the end of PartitionBoundInfoData to minimize ABI risks.)
Discussion: https://postgr.es/m/16840-571a22976f829ad4@postgresql.org
Since this function is used as a CHECK constraint condition,
returning NULL is tantamount to returning TRUE, which would have the
effect of letting in a row that doesn't satisfy the hash condition.
Admittedly, the cases for which this is done should be unreachable
in practice, but that doesn't make it any less a bad idea. It also
seems like a dartboard was used to decide which error cases should
throw errors as opposed to returning NULL.
For the checks for NULL input values, I just switched it to returning
false. There's some argument that an error would be better; but the
case really should be can't-happen in a generated hash constraint,
so it's likely not worth more code for.
For the parent-relation-open-failure case, it seems like we might
as well let relation_open throw an error, instead of having an
impossible-to-diagnose constraint failure.
Back-patch to v11 where this code came in.
Discussion: https://postgr.es/m/24067.1605134819@sss.pgh.pa.us
We failed to pass down the query string to check_new_partition_bound,
so that its attempts to provide error cursor positions were for naught;
one must have the query string to get parser_errposition to do anything.
Adjust its API to require a ParseState to be passed down.
Also, improve the logic inside check_new_partition_bound so that the
cursor points at the partition bound for the specific column causing
the issue, when one can be identified.
That part is also for naught if we can't determine the query position of
the column with the problem. Improve transformPartitionBoundValue so
that it makes sure that const-simplified partition expressions will be
properly labeled with positions. In passing, skip calling evaluate_expr
if the value is already a Const, which is surely the most common case.
Alexandra Wang, Ashwin Agrawal, Amit Langote; reviewed by Ashutosh Bapat
Discussion: https://postgr.es/m/CACiyaSopZoqssfMzgHk6fAkp01cL6vnqBdmTw2C5_KJaFR_aMg@mail.gmail.com
Discussion: https://postgr.es/m/CAJV4CdrZ5mKuaEsRSbLf2URQ3h6iMtKD=hik8MaF5WwdmC9uZw@mail.gmail.com
Includes some manual cleanup of places that pgindent messed up,
most of which weren't per project style anyway.
Notably, it seems some people didn't absorb the style rules of
commit c9d297751, because there were a bunch of new occurrences
of function calls with a newline just after the left paren, all
with faulty expectations about how the rest of the call would get
indented.
Previously, the partitionwise join technique only allowed partitionwise
join when input partitioned tables had exactly the same partition
bounds. This commit extends the technique to some cases when the tables
have different partition bounds, by using an advanced partition-matching
algorithm introduced by this commit. For both the input partitioned
tables, the algorithm checks whether every partition of one input
partitioned table only matches one partition of the other input
partitioned table at most, and vice versa. In such a case the join
between the tables can be broken down into joins between the matching
partitions, so the algorithm produces the pairs of the matching
partitions, plus the partition bounds for the join relation, to allow
partitionwise join for computing the join. Currently, the algorithm
works for list-partitioned and range-partitioned tables, but not
hash-partitioned tables. See comments in partition_bounds_merge().
Ashutosh Bapat and Etsuro Fujita, most of regression tests by Rajkumar
Raghuwanshi, some of the tests by Mark Dilger and Amul Sul, reviewed by
Dmitry Dolgov and Amul Sul, with additional review at various points by
Ashutosh Bapat, Mark Dilger, Robert Haas, Antonin Houska, Amit Langote,
Justin Pryzby, and Tomas Vondra
Discussion: https://postgr.es/m/CAFjFpRdjQvaUEV5DJX3TW6pU5eq54NCkadtxHX2JiJG_GvbrCA@mail.gmail.com