Commit 558a9165 taught _bt_delitems_delete() to produce its own XID
horizon on the primary.  Standbys no longer needed to generate their own
latestRemovedXid, since they could just use the explicitly logged value
from the primary instead.  The deleted offset numbers array from the
xl_btree_delete WAL record was no longer used by the REDO routine for
anything other than deleting the items.

This enables a minor optimization:  We now treat the array as buffer
state, not generic WAL data, following _bt_delitems_vacuum()'s example.
This should be a minor win, since it allows us to avoid including the
deleted items array in cases where XLogInsert() stores the whole buffer
anyway.  The primary goal here is to make the code more maintainable,
though.  Removing inessential differences between the two functions
highlights the fundamental differences that remain.

Also change xl_btree_delete to use uint32 for the size of the array of
item offsets being deleted.  This brings xl_btree_delete closer to
xl_btree_vacuum.  Furthermore, it seems like a good idea to use an
explicit-width integer type (the field was previously an "int").

Bump XLOG_PAGE_MAGIC because xl_btree_delete changed.

Discussion: https://postgr.es/m/CAH2-Wzkz4TjmezzfAbaV1zYrh=fr0bCpzuJTvBe5iUQ3aUPsCQ@mail.gmail.com

Escape non-printable characters in failure reports, by using Data::Dumper
in Useqq mode.  Also, bump $Test::Builder::Level so the diagnostic
references the calling line, and use diag() instad of note(),
so it shows even in non-verbose mode (per request from Christoph Berg).

Also, give up on trying to test for the specific way that readline
chooses to overwrite existing text in the \DRD -> \drds test.
There are too many variants, it seems, at least on the libedit
side of things.

Dagfinn Ilmari Mannsåker and Tom Lane

Discussion: https://postgr.es/m/20200103110128.GA28967@msg.df7cb.de

Debian unstable is shipping a broken version of libedit: it de-escapes
words before passing them to the application's tab completion function,
preventing us from recognizing backslash commands.  Fortunately,
we have enough information available to dig the original text out of
rl_line_buffer, so ignore the string argument and do that.

I view this as a temporary workaround to get the affected buildfarm
members back to green in the wake of 7c015045.  I hope we can get
rid of it once somebody fixes Debian's libedit; hence, no back-patch,
at least for now.

Discussion: https://postgr.es/m/20200103110128.GA28967@msg.df7cb.de

Author: Fabien COELHO <coelho@cri.ensmp.fr>

Reported-by: Jon Jensen
Author: Jon Jensen
Reviewed-by: Amit Kapila and Robert Haas
Backpatch-through: 10
Discussion: https://postgr.es/m/nycvar.YSQ.7.76.1912301807510.9899@ybpnyubfg

Adjust a comment that describes how alignment of the new left page high
key works in btree_xlog_split(), the nbtree page split REDO routine.
The wording used before commit 2c03216d is much clearer, so go back
to that.

Satisfy perlcritic, mostly.  Per buildfarm.

The expectation within _bt_delitems_vacuum() is that caller has a
super-exclusive/cleanup buffer lock (not just a pin and a write lock).

When row triggers exist in partitioned partitions that are not either
part of FKs or deferred unique constraints, they are not correctly
cloned to their partitions.  That's because they are marked "internal",
and those are purposefully skipped when doing the clone triggers dance.
Fix by relaxing the condition on which internal triggers are skipped.

Amit Langote initially diagnosed the problem and proposed a fix, but I
used a different approach.

Reported-by: Petr Fedorov
Discussion: https://postgr.es/m/6b3f0646-ba8c-b3a9-c62d-1c6651a1920f@phystech.edu

Up to now, psql's tab-complete.c has had exactly no regression test
coverage.  This patch is an experimental attempt to add some.

This needs Perl's IO::Pty module, which isn't installed everywhere,
so the test script just skips all tests if that's not present.
There may be other portability gotchas too, so I await buildfarm
results with interest.

So far this just covers a few very basic keyword-completion and
query-driven-completion scenarios, which should be enough to let us
get a feel for whether this is practical at all from a portability
standpoint.  If it is, there's lots more that can be done.

Discussion: https://postgr.es/m/10967.1577562752@sss.pgh.pa.us

On further reflection about commit 4d02eb01, it occurs to me that
expandRTE() had better agree with what addRangeTableEntryForFunction()
is doing.  So teach that about functions possibly having typmods, too.

Make the function prototype order consistent with the definition order
in nbtpage.c.

One code path in addRangeTableEntryForFunction() neglected to assign
a collation to the tupdesc entry it constructs (which is a bit odd
considering the other path did do so).  This didn't matter before commit
5815696b, because nothing would look at the type data in this tupdesc;
but now it does.

While at it, make sure we assign the correct typmod as well.  Most
function expressions don't have a determinate typmod, but some do.

Per buildfarm, which showed failures in non-C collations, a case
I'd not thought to test for this patch :-(

When I added the ParseNamespaceItem data structure (in commit 5ebaaa49),
it wasn't very tightly integrated into the parser's APIs.  In the wake of
adding p_rtindex to that struct (commit b541e9ac), there is a good reason
to make more use of it: by passing around ParseNamespaceItem pointers
instead of bare RTE pointers, we can get rid of various messy methods for
passing back or deducing the rangetable index of an RTE during parsing.
Hence, refactor the addRangeTableEntryXXX functions to build and return
a ParseNamespaceItem struct, not just the RTE proper; and replace
addRTEtoQuery with addNSItemToQuery, which is passed a ParseNamespaceItem
rather than building one internally.

Also, add per-column data (a ParseNamespaceColumn array) to each
ParseNamespaceItem.  These arrays are built during addRangeTableEntryXXX,
where we have column type data at hand so that it's nearly free to fill
the data structure.  Later, when we need to build Vars referencing RTEs,
we can use the ParseNamespaceColumn info to avoid the rather expensive
operations done in get_rte_attribute_type() or expandRTE().
get_rte_attribute_type() is indeed dead code now, so I've removed it.
This makes for a useful improvement in parse analysis speed, around 20%
in one moderately-complex test query.

The ParseNamespaceColumn structs also include Var identity information
(varno/varattno).  That info isn't actually being used in this patch,
except that p_varno == 0 is a handy test for a dropped column.
A follow-on patch will make more use of it.

Discussion: https://postgr.es/m/2461.1577764221@sss.pgh.pa.us

The comment was apparently copy-and-pasted and did not reflect the
actual test outcome.

Currently while decoding changes, if the number of changes exceeds a
certain threshold, we spill those to disk.  And this happens for each
(sub)transaction.  Now, while reading all these files, we don't close them
until we read all the files.  While reading these files, if the number of
such files exceeds the maximum number of file descriptors, the operation
errors out.

Use PathNameOpenFile interface to open these files as that internally has
the mechanism to release kernel FDs as needed to get us under the
max_safe_fds limit.

Reported-by: Amit Khandekar
Author: Amit Khandekar
Reviewed-by: Amit Kapila
Backpatch-through: 9.4
Discussion: https://postgr.es/m/CAJ3gD9c-sECEn79zXw4yBnBdOttacoE-6gAyP0oy60nfs_sabQ@mail.gmail.com

_bt_delitems_vacuum() comments claimed that it isn't worth another scan
of the page to avoid falsely unsetting the BTP_HAS_GARBAGE page flag
hint (this happens to be the same wording that was removed from
_bt_delitems_delete() by my recent commit fe97c61c).  The comments made
little sense, though.  The issue can't have much to do with performing a
second scan of the target leaf page, since an LP_DEAD test could easily
be performed in the first scan of the page anyway (the scan that takes
place in btvacuumpage() caller).

Revise the explanation.  It makes much more sense to frame this as an
issue about recovery conflicts.  _bt_delitems_vacuum() cannot easily
generate an XID cutoff in the same way that _bt_delitems_delete() is
designed to.

Falsely unsetting the page flag is not ideal, and is likely to happen
more often than was supposed by the original comments.  Explain why it
usually isn't a problem in practice.  There may be an argument for
_bt_delitems_vacuum() not clearing the BTP_HAS_GARBAGE bit, removing the
question of it being falsely unset by VACUUM (there may even be an
argument for not using a page level hint at all).  This can be revisited
later.

If we have, say, an int column that is left-joined to a bigint column
with USING, the merged column is the int column promoted to bigint.
GROUP BY's tests for whether grouping on the merged column allows a
reference to the underlying column, or vice versa, should know about
that relationship --- and they do.  But I nearly broke this case with
an ill-advised optimization, so the lack of any test coverage for it
seems like a bad idea.

Commit fe97c61c updated LP_DEAD item deletion comments, but missed a
minor discrepancy on the REDO side.  Fix it now.

In passing, don't talk about the btree_xlog_vacuum() behavior within
btree_xlog_delete().  The reliance on XLOG_HEAP2_CLEANUP_INFO records
for recovery conflicts is already discussed within btvacuumpage() and
mentioned again in passing above btree_xlog_vacuum(), which seems
sufficient.

Backpatch-through: update all files in master, backpatch legal files through 9.4

Change the exception syntax used in the documentation to use the more
current

    except Exception as ex:

rather than the old

    except Exception, ex:

We keep the old syntax in the test code since Python <2.6 is still
supported there, but the documentation might as well use the modern
syntax.

Use __builtin_clz() where available.  Where it isn't, we can still win
a little by using the pg_leftmost_one_pos[] lookup table instead of
having a private table.

Also drop the initial right shift by ALLOC_MINBITS in favor of
subtracting ALLOC_MINBITS from the leftmost-one-pos result.  This
is a win because the compiler can fold that adjustment into other
constants it'd have to add anyway, making the shift-removal free.

Also, we can explain this coding as an unrolled form of
pg_leftmost_one_pos32(), even though that's a bit ahistorical
since it long predates pg_bitutils.h.

John Naylor, with some cosmetic adjustments by me

Discussion: https://postgr.es/m/CACPNZCuNUGMxjK7WTn_=WZnRbfASDdBxmjsVf2+m9MdmeNw_sg@mail.gmail.com

This currently works, but add this test to ensure it continues to work.
Lack of this test became evident after a recent bugfix submission that
would have inadvertently broken it, in
https://postgr.es/m/CA+HiwqFM2=i+uHB9o4OkLbE2S3sjPHoVe2wXuAD1GLJ4+Pk9eg@mail.gmail.com

Unique expression indexes can constrain data in creative ways, so show
two examples.

Reported-by: Tuomas Leikola

Discussion: https://postgr.es/m/156760275564.1127.12321702656456074572@wrigleys.postgresql.org

Backpatch-through: 9.4

The previous docs referenced these distinct ideas confusingly.

Reported-by: Eugen Konkov

Discussion: https://postgr.es/m/376945611.20191026161529@yandex.ru

Backpatch-through: 9.4

This operation was possible for the owner of the schema or a superuser.
Down to 9.4, doing this operation would cause inconsistencies in a
session whose temporary schema was dropped, particularly if trying to
create new temporary objects after the drop.  A more annoying
consequence is a crash of autovacuum on an assertion failure when
logging information about an orphaned temp table dropped.  Note that
because of 246a6c8f (present in v11~), which has made the removal of
orphaned temporary tables more aggressive, the failure could be
triggered more easily, but it is possible to reproduce down to 9.4.

Reported-by: Mahendra Singh, Prabhat Sahu
Author: Michael Paquier
Reviewed-by: Kyotaro Horiguchi, Mahendra Singh
Discussion: https://postgr.es/m/CAKYtNAr9Zq=1-ww4etHo-VCC-k120YxZy5OS01VkaLPaDbv2tg@mail.gmail.com
Backpatch-through: 9.4

This follows multiple complains from Peter Geoghegan, Andres Freund and
Alvaro Herrera that this issue ought to be dug more before actually
happening, if it happens.

Discussion: https://postgr.es/m/20191226144606.GA5659@alvherre.pgsql

When revalidate_rectypeid() acts to update a stale record type OID
in plpgsql's data structures, it fixes the active PLpgSQL_rec struct
as well as the PLpgSQL_type struct it references.  However, the latter
is shared across function executions while the former is not.  In a
later function execution, the PLpgSQL_rec struct would be reinitialized
by copy_plpgsql_datums and would then contain a stale type OID,
typically leading to "could not open relation with OID NNNN" errors.
revalidate_rectypeid() can easily fix this, fortunately, just by
treating typ->typoid as authoritative.

Per report and diagnosis from Ashutosh Sharma, though this is not his
suggested fix.  Back-patch to v11 where this code came in.

Discussion: https://postgr.es/m/CAE9k0Pkd4dZwt9J5pS9xhJFWpUtqs05C9xk_GEwPzYdV=GxwWg@mail.gmail.com

Mark the fields that should be accessed via partitioning-related
functions, as we already did for some other fields.

Amit Langote

Discussion: https://postgr.es/m/CA+HiwqFnK6LbVMACMCaqwWrvoSFTecZzufKRahg2qGvLPYMX=g@mail.gmail.com

Instead of passing around a pointer to the RangeTblEntry that
provides the desired column, pass a pointer to the associated
ParseNamespaceItem.  The RTE is trivially reachable from the nsitem,
and having the ParseNamespaceItem allows access to additional
information.  As proof of concept for that, add the rangetable index
to ParseNamespaceItem, and use that to get rid of RTERangeTablePosn
searches.

(I have in mind to teach the parser to generate some different
representation for Vars that are nullable by outer joins, and
keeping the necessary information in ParseNamespaceItems seems
like a reasonable approach to that.  But whether that ever
happens or not, this seems like good cleanup.)

Also refactor the code around scanRTEForColumn so that the
"fuzzy match" stuff does not leak out of parse_relation.c.

Discussion: https://postgr.es/m/26144.1576858373@sss.pgh.pa.us

confirmed_flush is part of a replication slot's information, but not
confirmed_lsn.

Author: Kyotaro Horiguchi
Discussion: https://postgr.es/m/20191226.175919.17237335658671970.horikyota.ntt@gmail.com
Backpatch-through: 11

The part in charge of doing the vacuum on all the indexes of a relation
was duplicated, with the same handling for progress reporting done.
While on it, update the progress reporting for heap vacuuming in the
subroutine doing the actual work, keeping the status update local.  This
way, any future caller of lazy_vacuum_heap() does not have to worry
about doing any progress reporting update.

Author: Justin Pryzby, Michael Paquier
Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com

Column defaults may be specified separately for each partition.
But INSERT via a partitioned table ignores those partition's default values.
The former is documented, but the latter restriction not.
This commit adds the note about that restriction into the document.

Author: Fujii Masao
Reviewed-by: Amit Langote
Discussion: https://postgr.es/m/CAHGQGwEs-59omrfGF7hOHz9iMME3RbKy5ny+iftDx3LHTEn9sA@mail.gmail.com

In the wake of commit 5b931237, there's no particular reason
for this restriction (previously, it was problematic because of
the implied rowtype reference).  A simple constraint on a whole-row
Var probably isn't that useful, but conceivably somebody would want
to pass one to a function that extracts a partitioning key.  Besides
which, we're expending much more code to enforce the restriction than
we save by having it, since the latter quantity is now zero.
So drop the restriction.

Amit Langote

Discussion: https://postgr.es/m/CA+HiwqFUzjfj9HEsJtYWcr1SgQ_=iCAvQ=O2Sx6aQxoDu4OiHw@mail.gmail.com

This is dead code in the wake of the previous commit.
We can always add it back if we need it again someday.

Discussion: https://postgr.es/m/CA+HiwqFUzjfj9HEsJtYWcr1SgQ_=iCAvQ=O2Sx6aQxoDu4OiHw@mail.gmail.com

Formerly the rd_partkey and rd_partdesc data structures were always
populated immediately when a relcache entry was built or rebuilt.
This patch changes things so that they are populated only when they
are first requested.  (Hence, callers *must* now always use
RelationGetPartitionKey or RelationGetPartitionDesc; just fetching
the pointer directly is no longer acceptable.)

This seems to have some performance benefits, but the main reason to do
it is that it eliminates a recursive-reload failure that occurs if the
partkey or partdesc expressions contain any references to the relation's
rowtype (as discovered by Amit Langote).  In retrospect, since loading
these data structures might result in execution of nearly-arbitrary code
via eval_const_expressions, it was a dumb idea to require that to happen
during relcache entry rebuild.

Also, fix things so that old copies of a relcache partition descriptor
will be dropped when the cache entry's refcount goes to zero.  In the
previous coding it was possible for such copies to survive for the
lifetime of the session, as I'd complained of in a previous discussion.
(This management technique still isn't perfect, but it's better than
before.)  Improve the commentary explaining how that works and why
it's safe to hand out direct pointers to these relcache substructures.

In passing, improve RelationBuildPartitionDesc by using the same
memory-context-parent-swap approach used by RelationBuildPartitionKey,
thereby making it less dependent on strong assumptions about what
partition_bounds_copy does.  Avoid doing get_rel_relkind in the
critical section, too.

Patch by Amit Langote and Tom Lane; Robert Haas deserves some credit
for prior work in the area, too.  Although this is a pre-existing
problem, no back-patch: the patch seems too invasive to be safe to
back-patch, and the bug it fixes is a corner case that seems
relatively unlikely to cause problems in the field.

Discussion: https://postgr.es/m/CA+HiwqFUzjfj9HEsJtYWcr1SgQ_=iCAvQ=O2Sx6aQxoDu4OiHw@mail.gmail.com
Discussion: https://postgr.es/m/CA+TgmoY3bRmGB6-DUnoVy5fJoreiBJ43rwMrQRCdPXuKt4Ykaw@mail.gmail.com

The following renaming is done so as source files related to index
access methods are more consistent with table access methods (the
original names used for index AMs ware too generic, and could be
confused as including features related to table AMs):
- amapi.h -> indexam.h.
- amapi.c -> indexamapi.c.  Here we have an equivalent with
backend/access/table/tableamapi.c.
- amvalidate.c -> indexamvalidate.c.
- amvalidate.h -> indexamvalidate.h.
- genam.c -> indexgenam.c.
- genam.h -> indexgenam.h.

This has been discussed during the development of v12 when table AM was
worked on, but the renaming never happened.

Author: Michael Paquier
Reviewed-by: Fabien Coelho, Julien Rouhaud
Discussion: https://postgr.es/m/20191223053434.GF34339@paquier.xyz

Using \ is unnecessary and ugly, so remove that.  While at it, stitch
the literals back into a single line: we've long discouraged splitting
error message literals even when they go past the 80 chars line limit,
to improve greppability.

Leave contrib/tablefunc alone.

Discussion: https://postgr.es/m/20191223195156.GA12271@alvherre.pgsql

Since d6c55de1, src/port/snprintf.c is able to use %m instead of
strerror().  A couple of utilities in src/bin/ have already done the
switch, and do it now for pg_waldump as this reduces the workload for
translators.

Note that more could be done, particularly with pgbench.  Thanks to
Kyotaro Horiguchi for the discussion.

Discussion: https://postgr.es/m/20191129065115.GM2505@paquier.xyz

Our algorithm for choosing batch numbers turned out not to work
effectively for multi-billion key inner relations.  We would use
more hash bits than we have, and effectively concentrate all tuples
into a smaller number of batches than we intended.  While ideally
we should switch to wider hashes, for now, change the algorithm to
one that effectively gives up bits from the bucket number when we
don't have enough bits.  That means we'll finish up with longer
bucket chains than would be ideal, but that's better than having
batches that don't fit in work_mem and can't be divided.

Batch-patch to all supported releases.

Author: Thomas Munro
Reviewed-by: Tom Lane, thanks also to Tomas Vondra, Alvaro Herrera, Andres Freund for testing and discussion
Reported-by: James Coleman
Discussion: https://postgr.es/m/16104-dc11ed911f1ab9df%40postgresql.org