Authors: Chris Hajas and Todd Sedano

read_bytes will be negative if read() returns a negative value on error.
Therefore, this should be a signed int.

This is a backup-breaking failure and the dump should exit rather than
simply breaking from the loop.

These .sql files assumed that the timezone for machines running the tests would
have the timezone as PST/PDT. This commit explicitly sets the timezone to PST in
the .sql files and updates the .ans files to handle this new behavior.

This resolves #1880. Thanks Lirong Jian for the inital PR #1952.

Details:
Fixed a typo of string comparison in cdbexplain
Fix the typo for UNINITIALIZED_SORT.
Add new method to convert sort space type str to enum.

These were not that exciting tests, because ORCA falls back to the Postgres
planner for these queries. But it's a reasonable test case of partition
pruning in general, after adding a WHERE clauses to the query, so add one
such test case to the main test suite.

These tests are from the legacy cdbfast test suite, from
writable_ext_tbl/test_* files. They're not very exciting, and we already
had tests for some variants of these commands, but looking at the way all
these privileges are handled in the code, perhaps it's indeed best to have
tests for many different combinations. The tests run in under a second, so
it's not too much of a burden.

Compared to the original tests, I removed the SELECTs and INSERTs that
tested that you can also read/write the external tables, after successfully
creating one. Those don't seem very useful, they all basically test that
the owner of an external table can read/write it. By getting rid of those
statements, these tests don't need a live gpfdist server to be running,
which makes them a lot simpler and faster to run.

Also move the existing tests on these privileges from the 'external_table'
test to the new file.

The gpcheckcat 'inconsistent' check should skip the pg_index.indcheckxmin column
for now as due to the HOT feature, the value of this column can be different
between the master and the segments.

A more long term fix will resolve the underlying issue that makes the
indcheckxmin column value to be different between the master and the segments.

rolresgroup was added to pg_authid in previous commits, we also should
add it to pg_authid.
Signed-off-by: NNing Yu <nyu@pivotal.io>

There is the same restrictions with the CREATE ROLE syntax, only
superuser can be assigned to the admin_group.

    -- assign the role r1 to resource group g1
    ALTER ROLE r1 RESOURCE GROUP g1;
Signed-off-by: NNing Yu <nyu@pivotal.io>

There are two default resource groups, default_group and admin_group,
roles will be assigned to one of them depending on they are superuser or
not unless a group is explicitly specified. Only superuser can be
assigned to the admin_group.

    -- create a role r1 and assign it to resgroup g1
    CREATE ROLE r1 RESOURCE GROUP g1;
Signed-off-by: NNing Yu <nyu@pivotal.io>

[ci skip]

dblink is a bulit-in module of postgresql, this patch enable it in GPDB.

1. Enable compiling dblink with GPDB
2. Update and enable dblink test cases

Known issue:

1. The following functions are not supported by GPDB and disabled temporarily
   due to GPDB behavior changed.
   dblink_send_query()
   dblink_is_busy()
   dblink_get_result()

2. Anonymous connection and connection by name are not supported in a nested statement with
   both write and read. E.g.
   INSERT INTO tblname SELECT * FROM dblink(['connName'],'SELECT * FROM otherdb.tblname') as tbl;
   Currently, only connection with connection string is supported in such statements.
Signed-off-by: NPeifeng Qiu <pqiu@pivotal.io>
Signed-off-by: NYuan Zhao <yuzhao@pivotal.io>
Signed-off-by: NJianxia Chen <jchen@pivotal.io>
Signed-off-by: NAdam Lee <ali@pivotal.io>

v22 of the pull_request resource removes the default behavior making
shallow clones. A shallow clone will not fetch tags, since no tagged
commits are walked over during the fetch. We need tags to be present to
build and determine the version string.
Signed-off-by: NJingyi Mei <jmei@pivotal.io>

Signed-off-by: NDavid Sharp <dsharp@pivotal.io>

* Transaction id support

* Removed extra line
[ci skip]

- Also remove unwanted test: the abort test should be skipped,
  based on feedback from Storage team
Signed-off-by: NC.J. Jameson <cjameson@pivotal.io>

These things are covered by the 'transactions' test, and the
'uao_dml/uao_dml_column' test, in the main test suite.

When `explain` traces a VAR to a CONST, it assumes (rightly so) that
it's most likely the only constant being projected at that operator,
e.g.

```
EXPLAIN SELECT a AS c FROM (SELECT 1 AS a) AS bar ORDER BY 1;
                   QUERY PLAN
------------------------------------------------
 Sort  (cost=0.03..0.04 rows=1 width=4)
   Sort Key: (1)
   ->  Result  (cost=0.00..0.01 rows=1 width=0)
 Optimizer status: legacy query optimizer
(4 rows)

```

For a query containing the `VALUES` construct, planner would generate
`ValuesScan`, whereas ORCA generates an `Append` with each constant as a
`Result` node underneath. Notice how the ORCA plan shape breaks the
assumption of the `EXPLAIN` rendering logic, resulting in somehow
confusing output like this:

```
explain select * from foo, (values (1), (2)) f(i) where foo.a=f.i;
                                  QUERY PLAN
------------------------------------------------------------------------------
 Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..431.00 rows=1 width=8)
   ->  Hash Join  (cost=0.00..431.00 rows=1 width=8)
         Hash Cond: (1) = a
         ->  Append  (cost=0.00..0.00 rows=1 width=4)
               ->  Result  (cost=0.00..0.00 rows=1 width=4)
                     ->  Result  (cost=0.00..0.00 rows=1 width=4)
               ->  Result  (cost=0.00..0.00 rows=1 width=4)
                     ->  Result  (cost=0.00..0.00 rows=1 width=4)
         ->  Hash  (cost=431.00..431.00 rows=1 width=4)
               ->  Table Scan on foo  (cost=0.00..431.00 rows=1 width=4)
 Settings:  optimizer=on
 Optimizer status: PQO version 2.7.0
(12 rows)
```
In the above case, `(1)` is less than helpful because that is not the
only constant in the relation.

This commit renders a `VAR` node that trace back to a `CONST` that has a
non empty `resname` with the `resname` instead of the constant value.

An unintended (but good) consequence of this change is that some
constants are now rendered as their projected column instead of the
constant value, e.g.:

```
EXPLAIN SELECT a AS c FROM (SELECT 1 AS a) AS bar ORDER BY 1;
                   QUERY PLAN
------------------------------------------------
 Sort  (cost=0.03..0.04 rows=1 width=4)
   Sort Key: c
   ->  Result  (cost=0.00..0.01 rows=1 width=0)
 Optimizer status: legacy query optimizer
(4 rows)
```
Signed-off-by: NJesse Zhang <sbjesse@gmail.com>
Signed-off-by: NBhunvesh Chaudhary <bchaudhary@pivotal.io>

The TINC plperl91 tests in package/language were mostly duplicates
of what is already tested in src/pl/plperl (especially since sources
were properly synced recently). There were also massive duplication
within the tests of testing every single datatype and various
combinations thereof. This leaves some superusers tests for plperlu
which still holds some value due to them operating with different
users.

The Naive Bayes functionality is deprecated, and removed, in 5.0 in
favour of using MadLib. Remove test.

This only tested that functions in various languages returning
SETOF could be called with SELECT func() as well as SELECT *
FROM func(). We have ample coverage of this in ICW already so
remove.

These tests were killed a few years ago when they started failing
on error message output. The verdict back then was they weren't at
all interesting to keep running anyways so they were skipped but
not removed. Remove from the repo.

These tests were almost redundant with existing tests, but there were
a couple of things they did test:

* gp_create_table_random_default_distribution = on works, and
* you cannot add a PRIMARY KEY to a randomly distributed table.

You can't do it, but it would still be good to have a test case for it.
Spotted while reviewing legacy "cdbfast" tests - there was a test for this
there.

Long time ago, they apparently weren't in the catalogs, but they're not
going to suddenly disappear from there now, any more than any other
function. The functions themselves are called elsewhere in the test file,
so if they are missing or broken, the test will fail that way.

This allows us to have the exact same error message and hint for errors,
as what the traditional planner produces. That makes testing easier, as
you don't need to have a different expected output file for ORCA and
non-ORCA. And allows for more structured errors anyway.

Use the new function for the case of trying to read from a WRITABLE
external table. There was no test for that in the main test suite
previously. There was one in the gpfdist suite, but that's not really the
right place, as that error is caught the same way regardless of the
protocol. While we're at it, re-word the error message and change the error
code to follow the Postgres error message style guide.

Like in commit 08f1d8e8. Float8, time, timestamp, and timestamptz have
the same typlen, typyval, typtype, typalign and typstorage attributes as
int8. It is therefore sufficient to run these tests with int8.

Likewise, date has the same type attributes as int4, so the tests on int4
cover the same ground.

The original TINC tests where these were copied for tested specifically
that ORCA can use Dynamic Index Scans for these queries. To be closer to
that spirit, even when running without ORCA, try to use index scans when
possible. It exercises more codepaths in the planner anyway, and allows
us to notice more easily if the planner loses the ability to use an index
for some reason.

`gpinitsystem`, `gpstart` and `gpstop` had different prompting before:

$ gpinitsystem -c gpinitsystem_config 
...
Continue with Greenplum creation **Yy/Nn>**
n
...

$ gpstart
...

Continue with Greenplum instance startup **Yy|Nn (default=N):**
**>** n
...

$ gpstop
...

Continue with Greenplum instance shutdown **Yy|Nn (default=N):**
**>** n
...

This commit makes them consistent:

$ gpinitsystem -c gpinitsystem_config 
...

Continue with Greenplum creation Yy|Nn (default=N):
> n

Signed-off-by: NTom Meyer <tmeyer@pivotal.io>

* Re-enable UPDATE tests
* After statements that fail on GPDB but not in upstream, or vice versa,
  add fixup-statements, to restore the test table's state to what it is in
  the upstream, so that the expected output for subsequent tests match up
  with upstream.
* Change row order in expected output to match upstream.

Rename checkpoint.c to checkpointer.c. And move the code from bgwriter.c to
checkpointer.c and also renames most of corresponding data structures to refect
the clear ownership and association. This commit brings it as close as possible
to PostgreSQL 9.2.

Reference to PostgreSQL related commits:
commit 806a2aee
    Split work of bgwriter between 2 processes: bgwriter and checkpointer.
commit bf405ba8
    Add new file for checkpointer.c
commit 8f28789b
    Rename BgWriterShmem/Request to CheckpointerShmem/Request
commit d843589e5ab361dd4738dab5c9016e704faf4153
    Fix management of pendingOpsTable in auxiliary processes.