* Enable to run behave tests with the installed binary instead of source
* code.

* This is an effort to enable us to run behave tests without relying on
* the source code and instead use the GPDB installed in the system as
* the behave tests are integration tests.

* If the version has non-digit characters, then the float casting fails
* for the minor version.

Author: Chumki Roy <croy@pivotal.io>
Date:   Mon Dec 5 11:06:45 2016 -0800

    Remove error line - Aborting PQO plan generation.
    * Fix ORCA error message and make it same as with Planner.

Author: Haisheng Yuan <hyuan@pivotal.io>
Date:   Wed Jul 20 14:30:38 2016 -0700

    Update kerberos tinc test answer files

Cherry-picked into GPDB ahead of merge to fix compiler warning from
backporting of bytea HEX encode/decode.

Coverity complained that the aoseg_namespace variable potentially
was uninitialized at usage, and while technically correct it would
require the pg_aoseg namespace to be missing in the source cluster
which seems far fatched. Fix by forcing initialization to invalid
and ensuring that the Oid has been changed to a valid Oid before
using.

While looking at the code I also realized it had been blindly
cargoculted to the same design pattern as the remaining queries
which here was pointless. Simplify the logic and avoid allocating
an extra PQExpBuffer.

While the type cache used in binary upgrades isn't a dumpable
object, it has a DO_ identifier associated to be able to use
the indexing code. Skip it when looking through the DO objects
to keep compilers and static analyzers happy.

This was already done in pg_dump, do so for cdb_dump_agent too
and expand the documentation on why.

When adding operators to an existing operator family, ensure to
dispatch the Oids of the created objects to the segments. This
also backports the upstream alt_opf17 test case from alter_generic.

Reported by Dhanashree Kashid

- Update answer file for
  tincrepo/mpp/gpdb/tests/catalog/basic/security_definer/pre_script.ans
- Index on pg_aoseg was removed. Update
  tincrepo/mpp/gpdb/tests/catalog/mpp25256/mpp25256.sql and its answer file to
  reflect that.
- UDF exception handling test cases:
  - Some of the answer files had to be updated for the (file: line no) diff
  - The error messages during gang creation has changed. A few answer files had
    to be changed to account for the diff
- AOCO alter table
  - Fix the diff in the \d+ output for aoco table with cidr and inet columns

Update the .ans file to conform to the test output for 5.0.

The patch aims to cleanup and fix tests in the
tincrepo/mpp/gpdb/tests/storage/basic folder.
- Fix the external table tests.
- Fix transaction management test and remove the max_prepared_transactions test
  since it is no longer valid.
- Fix the xidlimits compilation.
- Fix the partition ddl tests.

Update the .ans file to conform to the test output for 5.0.

If a CTAS query uses a query with an InitPlan, and the InitPlan contains
a call to a function that dispatches another query to the segments, the
list of Oids is dispatched to the segments prematurely, in the inner query.
As a result, when the CTAS query is dispatched, the OID of the new table
is not included, and you get a "no pre-assigned OID for relation" error.

To fix, attach the OID of the new table to the query descriptor before
executing the InitPlans.

Fixes github issue #1543, reported with test case by Abhijit Subramanya.

Currently, the walsender process is not properly stopped during
gpstop. The SignalSomeChildren function was changed in the Postgres
8.3 merge in commit 20ac0be1. It was
previously at a Postgres 9.2 state to support sending signals to
walsender as a part of our master and standby master replication.

We don't build this as part of the CI, and we still won't pass the PL/TCL
regression tests (doesn't seem like there's anything serious there at a
quick glance though), but let's at least make it compile.

Per github issue #1524

For simplicity. This is less error-prone, too, in the face of future changes
to ExprContext.

DynamicTableScanInfo is an extension of EState, so always allocate it in the
same memory context. The DynamicTableScanInfo.memoryContext field always
pointed to es_query_cxt, so that is what we in fact always did anyway, this
just removes the unnecessary abstraction, for simplicity.

Pass the EState that contains it to where it's needed, instead.

This doesn't save too much time, it is just the pedant in me cringing
when I looked at why we had a mandatory proprietary dependency.

Alas this probably should have been done when we first introduced the
`--enable-netbackup` configure flag, but better late than never.

The four BSA agents seem to (inherently, as opposed to incidentally)
share the same dependencies. If that's true, we might as well declare
them together in one place.

Those binaries need the gpbsa shared object to build. This wasn't
exposed until EITHER

0. We changed the order we build them in the previous commit; OR
0. We ran a parallel build

This commit simply adds the proper dependency declaration to all the
offending targets.

This should not break anything given the declarative nature of Make. But
it does (especially noticeable if you are not running a parallel
`make`). Because we are not declaring a dependency on gpbsa libraries on
the the dump agents executables.

When building a scan descriptor for appendonly tables, the compression type for
a table is copied from the pg_appendonly row from relcache. When the
pg_appendonly row for the AO table gets updated, as part of relcache
invalidation, the old relation object will be destroyed. However the new
relation object would still point to the original compresstype object. So
appendonly_beginrange scan_internal() should use pstrdup to copy the
compression type instead of copying the pointer directly.

Connect to a segment in utility mode and run the following query

The test case that's included is not for the bug that was fixed, but for an
older bug with default_tablespace, that doesn't seem to be covered by any
existing tests. Seems good to add a test for it since we're touching the
code.

The bug that this actually fixes, is difficult to test, hence no regression
test for that included. But to reproduce manually:

---
PGOPTIONS="-c gp_session_role=utility" psql -p 25432 gptest
psql (8.3.23)
Type "help" for help.
gptest=# create table t1(a int);
NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Greenplum Database data distribution key for this table.
HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
CREATE TABLE
gptest=# insert into t1 values (1), (2), (3), (4);
INSERT 0 4
gptest=# create table t2 as select * from t1;
server closed the connection unexpectedly
	This probably means the server terminated abnormally
	before or while processing the request.
---

Fixes github issue #1511, reported with test case by Abhijit Subramanya.

When a superuser tries to split a partition table which was created by a non
superuser, the split should complete successfully and the resulting partitions
should have the owner set to the owner of the parent table.

To avoid hang issue, rewrite AnalyzerWorker to follow the same logic as Worker class. Besides, haltWork()
need to be called after join() to shutdown the whole WorkerPool, fix some incorrect usage for this.

Instead of logging errno, inlude the %m format specifier which pulls
in the human readable errno translation. Also fix typos and make the
error logging a bit more consistent with regards to style, string
quoting and spelling.

From a report by Github user laixiong in issue #1507

If the format_str variable is left uninitialized we are in trouble
(partly because palloc never returns NULL on failure to allocate).
Remove useless checks and also remove initialization on declaration
since that can hide us breaking the code subtly as it stands today.

* Implement bytea HEX encode/decode
this is from upstream commit a2a8c7a6
The primary reason for back patching is to allow pgadmin4 to work
Unfortunately there are a few hacks to make this work without Enum Config GUCs
mostly around setting the GUC as the target is an integer but the guc value is a string

* fixed regression tests \
removed unused code

The code in rewriteDefine.c didn't know about AO tables, and tried to use
heap_beginscan() on an AO table. It resulted in an ERROR, because even for
an empty table, RelationGetNumberOfBlocks() returned 1, but we tried to
scan it like a heap table, heap_getnext() couldn't find the block. To fix,
add an explicit check that you can only turn heap tables into views this
way. This also forbids turning an external table into a view. (We could
possible relax that limitation, but it's a pointless feature anyway we only
support for backwards-compatibility reasons in PostgreSQL.)

Also add test case for this.

Constraints are not uniquely identified by namespace and name alone.
It's possible to have multiple constraints with the same name, but attached
to different tables or domains. So include parent relation and domain OIDs
in the key.

This was uncovered by a regression test in the TINC test suite. Move the
OID consistency tests from TINC into the normal regression test suite, to
have test coverage for this.

1. Add custom external table dependency with protocol to keep the dumped
   SQL ordered correctly. (create protocol must execute before creating custom
   external table)

2. Remove \" from dumped location URI

3. Support dump multiple locations
Signed-off-by: NPeifeng Qiu <pqiu@pivotal.io>
Signed-off-by: NYuan Zhao <yuzhao@pivotal.io>

Commit 2a3df5d0 introduced broken changes to the ereport() calls in the
external table command. Fix by including the proper format specifier.
Also concatenate the error message to a single row while at it to
improve readability.

Spotted by Heikki Linnakangas

Update the .ans file to conform to the test output for 5.0.

Seen failures multiple times in pipeline at point for checking if database is in
changetracking or not for failover_to_mirror scenarios. For this case primary is
paniced and mirror takes over but due to work-load running, master may see
panics due to 2PC erroring out after retries. The test should be able to
continue working in such a condition, hence adding wait_for_database to be up
after master panic.

* Timeout was removed from the WorkerPool on commit
  c7dfcc45