This is the final batch of commits from PostgreSQL 9.2 development,
up to the point where the REL9_2_STABLE branch was created, and 9.3
development started on the PostgreSQL master branch.

Notable upstream changes:

* Index-only scan was included in the batch of upstream commits. It
  allows queries to retrieve data only from indexes, avoiding heap access.

* Group commit was added to work effectively under heavy load. Previously,
  batching of commits became ineffective as the write workload increased,
  because of internal lock contention.

* A new fast-path lock mechanism was added to reduce the overhead of
  taking and releasing certain types of locks which are taken and released
  very frequently but rarely conflict.

* The new "parameterized path" mechanism was added. It allows inner index
  scans to use values from relations that are more than one join level up
  from the scan. This can greatly improve performance in situations where
  semantic restrictions (such as outer joins) limit the allowed join orderings.

* SP-GiST (Space-Partitioned GiST) index access method was added to support
  unbalanced partitioned search structures. For suitable problems, SP-GiST can
  be faster than GiST in both index build time and search time.

* Checkpoints now are performed by a dedicated background process. Formerly
  the background writer did both dirty-page writing and checkpointing. Separating
  this into two processes allows each goal to be accomplished more predictably.

* Custom plan was supported for specific parameter values even when using
  prepared statements.

* API for FDW was improved to provide multiple access "paths" for their tables,
  allowing more flexibility in join planning.

* Security_barrier option was added for views to prevents optimizations that
  might allow view-protected data to be exposed to users.

* Range data type was added to store a lower and upper bound belonging to its
  base data type.

* CTAS (CREATE TABLE AS/SELECT INTO) is now treated as utility statement. The
  SELECT query is planned during the execution of the utility. To conform to
  this change, GPDB executes the utility statement only on QD and dispatches
  the plan of the SELECT query to QEs.
Co-authored-by: NAdam Lee <ali@pivotal.io>
Co-authored-by: NAlexandra Wang <lewang@pivotal.io>
Co-authored-by: NAshwin Agrawal <aagrawal@pivotal.io>
Co-authored-by: NAsim R P <apraveen@pivotal.io>
Co-authored-by: NDaniel Gustafsson <dgustafsson@pivotal.io>
Co-authored-by: NGang Xiong <gxiong@pivotal.io>
Co-authored-by: NHaozhou Wang <hawang@pivotal.io>
Co-authored-by: NHeikki Linnakangas <hlinnakangas@pivotal.io>
Co-authored-by: NJesse Zhang <sbjesse@gmail.com>
Co-authored-by: NJinbao Chen <jinchen@pivotal.io>
Co-authored-by: NJoao Pereira <jdealmeidapereira@pivotal.io>
Co-authored-by: NMelanie Plageman <mplageman@pivotal.io>
Co-authored-by: NPaul Guo <paulguo@gmail.com>
Co-authored-by: NRichard Guo <guofenglinux@gmail.com>
Co-authored-by: NShujie Zhang <shzhang@pivotal.io>
Co-authored-by: NTaylor Vesely <tvesely@pivotal.io>
Co-authored-by: NZhenghua Lyu <zlv@pivotal.io>

This reverts commit 4750e1b6.

This reverts commit c242c8b5.

This reverts commit 1a324f69.

Query node now doesnot have intoClause, it has been commented during the
92 merge, so doing the same to fix compilation issues with assert /
debug enabled.

The following pipeline:

    $CAT $SED_TMP_FILE | $TRUSTED_SHELL ${SED_HOST} $DD of=$SED_TMP_FILE > /dev/null 2>&1

when run with a TRUSTED_SHELL that executed the command on the local
host, suffered from a race where cat was reading from the file as dd was
attempting to truncate it. In situations where dd won that race, the
resulting sed command file would be empty, and no change would be made.

Simplify the logic by removing all temporary files, and forwarding the
sed command file contents to TRUSTED_SHELL via the stdin pipe.
Co-authored-by: NShoaib Lari <slari@pivotal.io>

This is the final batch of commits from PostgreSQL 9.2 development,
up to the point where the REL9_2_STABLE branch was created, and 9.3
development started on the PostgreSQL master branch.

Notable upstream changes:

* Index-only scan was included in the batch of upstream commits. It
  allows queries to retrieve data only from indexes, avoiding heap access.

* Group commit was added to work effectively under heavy load. Previously,
  batching of commits became ineffective as the write workload increased,
  because of internal lock contention.

* A new fast-path lock mechanism was added to reduce the overhead of
  taking and releasing certain types of locks which are taken and released
  very frequently but rarely conflict.

* The new "parameterized path" mechanism was added. It allows inner index
  scans to use values from relations that are more than one join level up
  from the scan. This can greatly improve performance in situations where
  semantic restrictions (such as outer joins) limit the allowed join orderings.

* SP-GiST (Space-Partitioned GiST) index access method was added to support
  unbalanced partitioned search structures. For suitable problems, SP-GiST can
  be faster than GiST in both index build time and search time.

* Checkpoints now are performed by a dedicated background process. Formerly
  the background writer did both dirty-page writing and checkpointing. Separating
  this into two processes allows each goal to be accomplished more predictably.

* Custom plan was supported for specific parameter values even when using
  prepared statements.

* API for FDW was improved to provide multiple access "paths" for their tables,
  allowing more flexibility in join planning.

* Security_barrier option was added for views to prevents optimizations that
  might allow view-protected data to be exposed to users.

* Range data type was added to store a lower and upper bound belonging to its
  base data type.

* CTAS (CREATE TABLE AS/SELECT INTO) is now treated as utility statement. The
  SELECT query is planned during the execution of the utility. To conform to
  this change, GPDB executes the utility statement only on QD and dispatches
  the plan of the SELECT query to QEs.
Co-authored-by: NAdam Lee <ali@pivotal.io>
Co-authored-by: NAlexandra Wang <lewang@pivotal.io>
Co-authored-by: NAshwin Agrawal <aagrawal@pivotal.io>
Co-authored-by: NAsim R P <apraveen@pivotal.io>
Co-authored-by: NDaniel Gustafsson <dgustafsson@pivotal.io>
Co-authored-by: NGang Xiong <gxiong@pivotal.io>
Co-authored-by: NHaozhou Wang <hawang@pivotal.io>
Co-authored-by: NHeikki Linnakangas <hlinnakangas@pivotal.io>
Co-authored-by: NJesse Zhang <sbjesse@gmail.com>
Co-authored-by: NJinbao Chen <jinchen@pivotal.io>
Co-authored-by: NJoao Pereira <jdealmeidapereira@pivotal.io>
Co-authored-by: NMelanie Plageman <mplageman@pivotal.io>
Co-authored-by: NPaul Guo <paulguo@gmail.com>
Co-authored-by: NRichard Guo <guofenglinux@gmail.com>
Co-authored-by: NShujie Zhang <shzhang@pivotal.io>
Co-authored-by: NTaylor Vesely <tvesely@pivotal.io>
Co-authored-by: NZhenghua Lyu <zlv@pivotal.io>

Insist() is replaced with either Assert() or ereport(ERROR) as
appropriate.

The test for creating an external table with a custom format was
placed in the upstream create_table.sql test suite.  Move to the
external_table suite instead to keep all external table tests in
one place and to align better with upstream.
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

There is no point in checking for the presence of a formatter, as
it's already tested for in the error path in transformFormatOpts()
when the external table is created.  Remove dead code and make it
fail on the error case should we ever manage to reach here.  This
resolves a FIXME added during the 9.1 merge which highlighted the
issue.
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

Align a few error messages to match upstream coding convention for
error messages (which we are slowly adapting to), and fix a typo.
Spotted while fixing other things in the file.
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

Commit 9b2d6688 removed check_pg_get_expr_args() as was done in
upstream, but Greenplum had an altered version with logic broken out
into the helper function check_pg_get_expr_arg().  This function is
now dead code so remove as well to get rid of compiler warning.

The is_split_part member is compared in _equalIndexStmt(), and is
copied in _copyIndexStmt(), so we should include it in the read
and write functions as well.
Reviewed-by: NJacob Champion <pchampion@pivotal.io>
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

The altconname member defines an alternative name for the index if
another name for a constraint was requested. Including it in the
comparison is mostly for correctness, as the index name is compared,
even if few (if any) situations would fail on this being omitted.
Reviewed-by: NJacob Champion <pchampion@pivotal.io>
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

IndexStmt is defined in src/include/nodes/parsenodes.h with indexOid
before the is_part_child member. Reorder the node functions to match
the definition.
Reviewed-by: NJacob Champion <pchampion@pivotal.io>
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

Translate nestparams passed from ORCA to create the nestparams node in Nested Loop joins.

This feature can be enabled by setting the trace flag EopttraceEnableNestLoopParams.
Co-authored-by: NAbhijit Subramanya <asubramanya@pivotal.io>
Co-authored-by: NEkta Khanna <ekhanna@pivotal.io>

Translate nestparams passed from ORCA to create the nestparams node in Nested Loop joins.

This feature can be enabled by setting the trace flag EopttraceEnableNestLoopParams.
Co-authored-by: NAbhijit Subramanya <asubramanya@pivotal.io>
Co-authored-by: NEkta Khanna <ekhanna@pivotal.io>

The segwalrep test uses a mock WAL receiver to validate what the WAL
sender is streaming out. The mock that we created did not handle AO
XLOG records that have been split across two pages with a continuation
record. We now properly construct the split AO record back to get the
AO information that is needed for test validation.

Also, do some trivial spelling fixes.
Co-authored-by: NAsim R Praveen <apraveen@pivotal.io>

* docs - update guc gp_enable_relsize_collection - ORCA support
Will be ported to 5X_STABLE.

Generated dev pipelines should be included in .gitignore. It's too
easy to accidentally add a generated dev pipeline in a commit.
Authored-by: NJason Vigil <jvigil@pivotal.io>

When constructing the connectionstring for connecting to the QEs,
ensure we don't overflow the assigned buffer.
Reviewed-by: NGang Xiong <gxiong@pivotal.io>

We are well past the 8.5devel version, so go back to using the
actual version from PG_VERSION rather than hardcoding. This
removes a FIXME that was added during the 9.0 merge cycle due
to 9.0 being labelled 8.5 in upstream for some time before it
was renamed.
Reviewed-by: NVenkatesh Raghavan <vraghavan@pivotal.io>

When build depends by run "make", meet an error "pkg_resources.DistributionNotFound:
The 'pylint<1.9.0,>=1.8.1' distribution was not found and is required by conan".
Since we don't run pylint anywhere，it can be removed from dependencies file safely.
Reviewed-by: NChris Hajas <chajas@pivotal.io>
Reviewed-by: NDaniel Gustafsson <dgustafsson@pivotal.io>

Signed-off-by: NDhanashree Kashid <dkashid@pivotal.io>

The commit 8ab6a6b4 from upstream removed the function
check_pg_get_expr_args and all the calls to it.
Nevertheless the merge brought it back and with it the
issue corrected on commit f223bb7aSigned-off-by: NTaylor Vesely <tvesely@pivotal.io>
Co-authored-by: NTaylor Vesely <tvesely@pivotal.io>

gpload test case will run gpload with subprocess, read stdout
and stderr from it and wait for exit. sys.exit in gpload does some
cleanup may cause deadlock between test and gpload. os._exit will
exit immediately, but we need to flush stdout and stderr before
that.

* Remove reference to ability to use port/username

Port and username options do not work in current gpssh release.

* Removed example

* Added missing space

This commit moves hyperloglog aggregates and functions to catalog as
well as migrates any hyperloglog related functions into
`src/backend/utils`.

Between GPDB 5.8.0 and 5.9.0, the recommended Linux System Settings
changed for kernel.sem.
https://gpdb.docs.pivotal.io/580/install_guide/prep_os_install_gpdb.html#topic3
https://gpdb.docs.pivotal.io/590/install_guide/prep_os_install_gpdb.html#topic3

However, the gpdb repository had references to the kernel.sem parameter,
as well as a couple other parameters, which seem to be inconsistent.
Co-authored-by: NAmil Khanzada <akhanzada@pivotal.io>
Co-authored-by: NTrevor Yacovone <tyacovone@pivotal.io>

The regression database has shown to be full of questionable
objects which we probably won't support for upgrades without
manual intervention before the upgrade. With all the changes
going into pg_upgrade it's time to start on a known minimal
dataset and expand from there to the full ICW leftover schema.
As a stop-gap for now, drop the regression databases eand only
upgrade a small subset.

Also disallow upgrades of orhpaned toast tables, and clean
up the invocation of the test_gpdb_pre.sql script to be less
hacky.

This is extracted from a larger patch in 5e92c436adcb32d295e77b3f3.
Co-authored-by: NPaul Guo <paulguo@gmail.com>
Co-authored-by: NMax Yang <myang@pivotal.io>

Make sure we use the psql client from the new bindir (the on in PATH
might well be an upstream postgres psql etc), and use a port different
from the standard postgres port since it's likely to be in use.

Also fix integer value comparison which caused a warning in my Bash,
while it might work in other versions of Bash.

binary-upgrade mode;  instead only skip dumping the current user.

This bug was introduced in during the removal of split_old_dump().  Bug
discovered during local testing.

Previously each QE would dump/restore the schema before upgrading
the datafiles, which consumes a lot of time on large databases.
Instead, use the datadir which was created from the dump/restore
on the QD for the QEs as well and bootstrap the segment upgrade
to jump straight to copying files instead.

This is a first stab at implementing this model of distribution,
follow-up commits will be required to finalize the patch.