# TODO for MonetDB/SQL
# This file is provided AS-IS, no claims can be made based upon this
# file.
# Please don't remove entries from this file.  It functions as base for
# a Changelog.  It's format is like a Changelog, and the newest items
# are on top.  Please keep the status of items up-to-date, e.g. OPEN,
# DONE, PARTIAL.
# $Header$

*sql-2.14 (15 Aug 2006)

  29 Sep 2006; Niels Nes <nielsnes>; OPEN
  Instead of one global tmp schema we need a list of local temporary tables
  per schema (each schema can have temps)

  02 Aug 2006; Fabian Groffen <mr-meltdown>; OPEN
  Add SSL support in JDBC.

  18 Jul 2006; Niels Nes <nielsnes>; OPEN
  Change symbol/list into tree with dynamic number of children.

  16 Jul 2006; Fabian Groffen <mr-meltdown>; OPEN
  Add COPY INTO which requires no "x RECORDS" part.  For this, extend
  the COPY INTO syntax stuff to the PostgreSQL type, such that string
  escaping is reduced to a minimum too.

  20 Jun 2006; Niels Nes <nielsnes>; OPEN
  Keep list of related sequence number with tables.  For two reasons:
  1) ease of cleanup/drop handling, and 2) cheaper implementations of
  identity columns.  Also related (?) detect dependencies of tables upon
  sequences to warn/cascade during DROP.
  
  20 Jun 2006; Fabian Groffen <mr-meltdown>; DONE
  "Generated columns" which are evaluated when they are needed.  The
  idea is to have some piece of MAL code that is just inserted before
  the actual bind.  This would solve problems for the authorisation
  system in M5 and can solve SEQUENCES which now have sneaky inplace
  transactions. (isview -> table_type, table 0, view 1, viewasbats 2, 
	sqlgenerated 3, mgenerated 4)

	table types are now
		0 table
		1 view
		2 generated (this means there is a function which generates
		             a table of this form, ie is not a table but a type)

  01 Jun 2006; Fabian Groffen <mr-meltdown>; OPEN
  Mapi speed-up.  The tuples in Mapi output can (and should) be sent in
  a more effective way.  This requires some big changes, but should
  bring us closer to MySQL and friends by lowering our protocol
  overhead.
