1.22 2013-08-10
    No code changes.

    - Updated test suite to pass with Perl 5.18 hash randomization.  (Mark Stosberg)

1.21 2012-01-20
    No code changes.

    - Sync'ed version of DBIx::Interp and SQL::Interp to avoid test failure (RT#73756)

1.20 2011-11-30

    No changes since 1.11-TRIAL. Releasing as stable.

1.11-TRIAL 2011-05-03

    - This is no longer supported for enabling SQL tracing: 
         SQL::Interp TRACE => 1;
      You can still set $ENV{TRACE_SQL} = 1  to enable SQL tracing.

    - Add sql_interp_strict() as a first stab at an API for adding a strict mode. 
    - Switched to using Sub::Exporter and Exporter. This allows you to do this:

      use SQL::Interp 'sql_interp_strict' => { -as => 'sql_interp' };

      So that strict mode is enabled throughout your code. Before you do that,
      read the docs on sql_interp_strict-- some of your code may need to be
      refactored for it work, so you may want to start with the incremental

      While sql_interp() will work as before for backwards compatibility, you
      can do this to rename them on import now:

      use SQL::Interp
         'sql_interp'        => { -as => 'sql_interp_insecure' },
         'sql_interp_strict' => { -as => 'sql_interp' };

      If you would like a shortcut for this in the form of SQL::Interp::Strict,
      let me know. 

    - The documentation has been updated to clarify that DBIx::Simple's
      SQL::Interp integration is recommended over DBIx::Interp.

    - Noted that SQL::Interp now has the primary repo hosted on Github:

1.10 2009-06-28

    - Add support for MySQL's "REPLACE INTO". (TBONE, Mark Stosberg, RT#43477)

    - When SQL snippets are provided as hashrefs, we now always sort them to
      insure consistent results for the benefit of possible statement caching.
      This shouldn't change your application behavior, as hashs are not to be
      considered an ordered data structure. If you want to force an element to
      be first in the where clause, that's better done with the pre-existing
          sql_interp("... WHERE a = 1 AND ",\%rest_of_where_hash);

      Thanks to Andrew Kirkpatrick for the patch (RT#33310).

    - The documentation said that we supported a reference to an arrayref in 
      the "IN ()" context, but there didn't appear to be code or tests for this.
      Both have been added now. (Mark Stosberg, RT#39336).

    - Update formatting and language about the historical fork with SQL::Interpolate.
    - Fix documentation typos (TBONE, RT#42726, RT#42733, RT#42336)

1.06 2007-12-20

    - Fix warnings introduced in 1.05
    - Fix Build.PL to create Makefile.PL

1.05 2007-11-15

    - Support "NOT IN" with an empty list (Aristotle, test by Mark Stosberg, RT#30752)
    - Support MySQL's "ON DUPLICATE KEY UPDATE", so it works like "SET". (Aristotle, RT#30652)

1.01 2007-07-15
    - No code changes. Just some doc improvements.

    - Add details about the fork to "See Also"  (suggestion from Juerd).
    - mention that SQL::KeywordSearch supports SQL::Interp-style results
    - mention DBIx::Simple support for SQL::Interp

1.00 2007-07-06

    - forked as SQL::Interp / DBIx::Interp from SQL::Interpolate
    - TRACE_SQL can now be set through an environment variable
    - sql_var() renamed to sql_type()
    - removed sql_literal(). It had been previously superceded by sql(). 
    - removed make_sql_interp and make_dbi_interp.
        These are just basic curry patterns that you probably won't need,
        and can be easily created as-needed.

    - significant doc update, mostly to simplify it
    - removed Macro support
    - removed filter support

0.40 2007-05-08
    - No code changes. Declaring stable.
    - List DBI as a dependency to squelch a test failure.

0.40_03 2006-09-27

    - Further documentation refinements. No code changes.

0.40_02 ??

	## SQL::Interp

	- POD improvements.

	- Expanded (table) identifier match:
	  /[a-zA-Z_.]+/ --> /[a-zA-Z_][a-zA-Z0-9_\$\.]*/
	  (reported by mark tiefenbruck)

	## DBIx::Interp

	- WARNING: BREAKS compatibility with 0.33.

	- DBIx::Interp now inherits from DBI.

	- Changed SQL::Interp::Key and SQL::Interp::Attr
	  to DBIx::Interp::Key and DBIx::Interp::Attr
	  respectively to reduce chance of name conflict.

	- Fixed circular references to allow garbage collection.

	- Now requires Scalar::Util.

0.33	2005-11-10

	## SQL::Interp

	- Build.PL - fixed version syntax so that META.yml is valid.

	- Added support for result sets:
	  [[1, 2], [3, 4]], 'UNION', [{x => 10, y > 11}, {x => 12, y => 13}]

	- Added support for temporary table references:
	  "SELECT * FROM", [[1, 2], [3, 4]], 'JOIN',
	                   [{x => 10, y > 11}, {x => 12, y => 13}]

	- Added support for "IS NULL" in
	    IN:  "WHERE", {bla => undef}
	    OUT: "WHERE bla IS NULL"
	  Recommended by (slaven) in

	- POD improvements in all modules.

0.32	2005-11-05

	## SQL::Interp

	This module has been simplified, and it incorporates
	defect corrections and style improvements.

	- Fixed ("x IN", []) to return "1=0" not "1=1".

	- Fixed ("x in", \@v) to not fail when "in" lowercase.

	- Fixed sql_interp.t on some platforms
	  (perl.cpan.testers, 254029)

	- Fixed sub-module version numbers for CPAN.

	- New sql() function (based on SQL::Interp::SQL
	  from Macro.pm).  sql_literal() is now depreciated
	  in favor of sql().

	- sql_interp() now handles recursion correctly.

	- sql_flatten() moved into SQL::Interp::Macro.

	- Added pod.t - Test::Pod checks

	- Added pod-coverage.t - Test::Pod::Coverage checks

	- Added dist.t - distribution checks

	- POD and code style improvements.

	## DBIx::Interp

	- Removed AUTOLOAD usage.

	## SQL::Macro

	The macro framework has been refined.

	- Fixed sql_and(sql()) and sql_or(sql()) to
	  now return '1=1' and '1=0' respectively (rather than "()").

	- sql_flatten() now expands depth-first rather than

	- sql_fragment() is now depreciated.  Use sql() instead.

	- relations() is now depreciated.  Use sql_rel_filter() instead.

	- expand() no longer has $filter param.  Macros now
	  query the $interp object for state.

	- SQLFilter::macro_names() no longer used.  The $interp
	  object no longer knows which macro object use which filters.

	- SQLFilter::filter_text_fragment() no longer has $changing_ref
	  in parameter list.

	## SQL::Interp::Filter

	- sql// objects now transform into calls to make_sql().

0.31	2005-09-27

	## SQL::Interp

	- Improved error reporting by sql_interp
	  (recommended by mark stosberg)

	- Added support for ("WHERE", {x => \@v, ...})
	  ==> "WHERE (x IN (?,...) AND ...)".
	  (recommended by multiple people)

	- Generate 'WHERE id = 5 and 1=1' and 'WHERE id = 5 or 1=0'
	  rather than 'WHERE id = 5 and 1' and 'WHERE id = 5 or 0'
	  for Oracle compatibility.
	  (reported by wojciech pietron)

	- Fixed some improper handling of "use"
	  parameters, e.g. $x in "use SQL::Interp FILTER=>$x"

	- Fixed various Exporter problems from custom import
	  (reported by mark stosberg)

	- sql_interp.t and dbi.t - Fixed test case errors due to
	  different hash order on MacOS.
	  (reported by sean davis)

	- Documentation improvements.

	## SQL::Interp::Macro

	- Fixed: '^' and '$' are no longer
	  automatically placed around regexes in 'relations' parameter.
	  (reported by wojciech pietron)

	## DBIx::Interp

	- Added support for transparent
	  caching of statement handles (no need to prepare()).

	- Fixed error in STX::fetchrow_hashref returning an
	  arrayref rather than a hashref.
	  (reported by mark tiefenbruck)

	- Added Carp

0.30	2005-01-10

	## General

	- Refactored SQL::Interp module into multiple modules.
	  Everything except sql_interp-related functionality
	  moved out of SQL::Interp and into DBIx::Interp,
	  SQL::Interp::Filter, and SQL::Interp::Macro (new module).
	  (recommended by mark s)

	- Reorganized macro processing.  SQL::Interp contains
	  only hooks for macro processing, and SQL::Interp::Macro
	  implements various macros.

	## SQL::Interp

	- added sql_var type

	- added sql_literal type (thanks mark s)

	- ("WHERE x in", []) now generates "WHERE 1" not "WHERE NULL"

	- croak not die

	- macros, sql_var, and sql_literal can now exist in
	  elements of aggregates (hashref and arrayref)

	## DBIx::Interp

	- DBIx::Interp is now derived from SQL::Interp.

	- Support for statement handle sets (stx/prepare).

	- Renamed select_key() to key_field().

	## SQL::Interp::Macro

	- SQL::Interp::SQL is now a macro for uniformity.

	- new macros: relations, sql_and, sql_fragment, sql_if,
	  sql_link, sql_or, sql_paren, sql_rel

	- removed limit() macro since MySQL supports Postgres syntax.
	(thanks mark s)

	## Other

	- reorganized POD (thanks mark s)

	- reorganized test cases

0.29	2004-12-25

	### SQL::Interp features

	- Added support for ("IN", $scalarref) (thanks mark s)

	- Added support for ("INSERT INTO mytable", $scalarref)

	- Added support for ("INSERT INTO mytable", $arrayref)

	- Added support for ("WHERE {x = 3, y = 4}")

	- Added support for ("INSERT HIGH_PRIORITY DELAYED INTO mytable", $ref)
	  for mysql.

	- Added simplified syntax for table joins using
	  R(...) and LINK(...) macros via new method filter_sql().

	- Added OO-interface in addition to the functional one.

	- Added support for \%attr to be returned by dbi_interp()
	  and sent to DBI when new SQL::Intepolate::Attr object is given.
	  This object may be created by new attr() function.

	- Added support for extra $key_field to be returned by
	  dbi_interp() when instance of new SQL::Interp::Key object
	  is given.  This object may be created by new attr() function.
	  This is for using in DBI::selectall_hashref and similar methods.

	- Added exports:
           attr, dbi_interp, select_key, limit, make_sql_interp,
           make_dbi_interp, filter_sql, sql_flatten, limit

	- Added debugging options in use statement: TRACE_SQL, TRACE_FILTER.

	- Changed: less extra whitespace around SQL generated by sql_interp().

	- Changed: In limit(), for mysql, don't use bind values since they
	  don't always work right.

	- Changed: In limit(), for mysql, $start and/or $count may be undef.

	### SQL::Interp internal changes

	- Dereferencing extra reference that source filtering produces from
	  sql[$x] when $x is an arrayref or hashref is now performed more
	  immediately in SQL::Interp::SQL::new() rather than sql_interp().

	- new SQL::Interp::Limit extends SQL::Interp::Macro.

	- Moved SQL::Interp::SQL into SQL::Interp

	- Fixed some documentation (thanks mark s)

	- Added more tests, more complete docs

	### DBIx::Interp features

	- Added new() method in addition to connect() to attach existing
	  database handle (thanks mark s)

	- Added dbh() and interp() methods to expose underlying DBI and
	  SQL::Interp objects.

0.20	2004-01-03
	- Added SQL::Interp::Filter module for source filtering
	  with string-like sql// operator.

	- Added DBIx::Interp module. (thanks jim c)

	- Added SQL::Interp::SQL module.

	- Support macro calls inside SQL (e.g. LIMIT macro).

	- Renamed sql_interpolate and dbi_interpolate to sql_interp
	  and dbi_interp respectively (thanks terrence).

	- Changed SQL::Interp to not export symbols by default.

	- Limited cross-database support for LIMIT clauses.

0.10	2003-12-26
	- Initial version.