cd vagrant
vagrant up
vagrant ssh
cd /gpdb/vagrant
./vagrant-build.sh

Fix minor errors related to the new location of the CentOS vagrant file.

Fix path names and add credits

Move the Vagrant instructions to /vagrant and add a note to that in the main README.md file.

Rename file.

Move instructions to the vagrant directory

Add instructions to bring up a larger (gpdemo) cluster in Vagrant.

Change to using absolute paths.

Fix formatting for SQL statements.

Make the explanation easier to read.

Correct typo.

Fix section numbering

Fix an error.

Convert the list in the last step to regular paragraphs.

Editing pass

Fix image

This set of commits was mostly already cherry-picked in, the main new patch
brings support for cross-type hashing in hash joins and hash index searches.

 Conflicts:
	doc/FAQ
	doc/TODO
	doc/src/FAQ/FAQ.html
	doc/src/FAQ/TODO.html
	doc/src/sgml/array.sgml
	doc/src/sgml/datatype.sgml
	doc/src/sgml/func.sgml
	doc/src/sgml/installation.sgml
	doc/src/sgml/libpq.sgml
	doc/src/sgml/maintenance.sgml
	doc/src/sgml/plperl.sgml
	doc/src/sgml/plpgsql.sgml
	doc/src/sgml/pltcl.sgml
	doc/src/sgml/ref/vacuum.sgml
	doc/src/sgml/rowtypes.sgml
	doc/src/sgml/xfunc.sgml
	src/backend/executor/nodeHash.c
	src/backend/executor/nodeHashjoin.c
	src/backend/postmaster/postmaster.c
	src/include/executor/nodeHash.h
	src/pl/plpgsql/src/pl_comp.c
	src/port/exec.c

Re-run of tidycat to bring in the changes from the recently merged
support for UDF default parameters.

This patch brings in a small fix for PL regression tests and an optimization
in build_regtype_array(). The remainder of the upstream commits were either
already cherry picked or was rejected.

 Conflicts:
	src/pl/plpgsql/src/pl_exec.c
	src/pl/plpython/plpython.c
	src/tools/msvc/pgbison.bat

The gpcrondump -c and -o options deleted the oldest backup sets based
on the backup timestamp. The options --cleanup-date and --cleanup-total
have been added to allow users more control over maintaining their
backup sets.

Authors: Karen Huddleston and Jimmy Yih

To make merging tiny bit easier.

Makes merging sligthly easier.

It was dead, because printTList field was never set. It was broken, because
when show_plan_tlist() called deparse_context_for_plan(), it passed
Plan nodes as arguments, while deparse_context_for_plan() expects them to
be RangeTblEntrys.

Too bad, the ability to print each node's target list would've been really
helpful. EXPLAIN VERBOSE in later PostgreSQL versions does that, so we'll
get there eventually...

The old behaviour was to store it in the master node.

All python code actually uses the pygresql module for database access.

Remove references to pcre and gperf from Ivy config and elsewhere. They
were not used for anything.

Remove license blurp for clapack from NOTICE ; it's not used or included
anywhere.

Remove references to lapack, f2c, blas from src/test/regress/Makefile.
They were used in a filter row to remove those libraries from LIBS, so this
was harmless, but also unnecessary when they're never added to LIBS in the
first place.

This patch brings UUID as a native datatype according to RFC4122.
Functionality for generating UUIDs and building against popular UUID
libraries will come in subsequent commits in order to keep the merges
somewhat along to upstream. To avoid build issues this commit also
contains 74a1a2b8 which fixes building
on OS X and Windows.

 Conflicts:
	src/backend/utils/adt/Makefile
	src/include/catalog/catversion.h
	src/include/catalog/pg_amop.h
	src/include/catalog/pg_amproc.h
	src/include/catalog/pg_cast.h
	src/include/catalog/pg_type.h
	src/include/utils/builtins.h