[Bug-gnubg] New Mac OS/X Builds [20110808 Snapshot]

[Michael Petch]

Howdy All,

I have produced new Mac OS/X builds from the 20110808 Snapshot. These versions have been updated:

- PPC release for Tiger (10.4)/Leopard(10.5) ,
- 32Bit Intel versions for Leopard (10.5), and Snow Leopard (10.6)
- 64Bit Intel versions for Snow Leopard(10.6), and Lion (10.7).

I have not officially tested the build on Lion but I have user feedback that it works in general.The README1ST file (in the DMG) contains information that may help if you experience OS/X Lion related crashes. If you experience other major problems with Lion please contact me at address@hidden .

The previous Mac release was April 10th, 2011. There are speed improvements, GUI changes, and general bug fixes. The new releases can be found here: http://gnubg.org/index.php?itemid=102
 
The Changelog entries for this release are:

2011-08-07  Michael Petch  <address@hidden>
	* set.c: Some imported XGID cube positions were not properly
	swapped when the board was displayed. Reported by Philippe (plm)

2011-08-07  Michael Petch  <address@hidden>
	* play.c: Fix bug #33143 reported by Kevin McDonough where 
	interrupting a forced move in 2D got into wrong state causing
	bot not to continue automatic play.

2011-08-03  Michael Petch  <address@hidden>
	* gnubgmodule.c: Add PythonParseMove function

2011-08-03  Michael Petch  <address@hidden>
	* set.c: Fix the display of the board when importing XGID's
	so that they appear *similar* to those HTML positions pasted to
	websites. Reported by Leo Bueno (on BGO). Assume that most
	people are using XG2 and exporting with default setting "Player
	on roll always at the bottom".

2011-08-03  Philippe Michel  <address@hidden>

	* gnubg.c, gtkboard.c, play.c, set.c: More fixes to swap players
	feature.

2011-08-01  Philippe Michel  <address@hidden>

	* play.c, gnubg.c, set.c: Fix display of SETBOARD records in the
	game record panel (bug #32778). Fix Game|Swap players for games
	ending with a SETBOARD record (in practice, games created from a
	pasted id).

2011-07-29  Michael Petch  <address@hidden>
	* analysis.c: Previous getLuckRating change had unintended side
	effects when we stopped using hard coded values. Hard coded values
	were off by an approximate factor of 10. Although this has been fixed
	it may wish to be revisited as to why the values were hardcoded in
	this fashion originally. Reported by Chris (pck).

2011-07-17  Michael Petch  <address@hidden>
	* gnubgmodule.c: Fix Python position key functions to internally call 
	the old style position keys functions. Fixes compiler errors due to 
	new position key changes made on 2011/07/15.

2011-07-15  Philippe Michel  <address@hidden>

	* positionid.c, positionid.h, lib/cache.c, lib/gnubg-types.h: Use
	a simpler position key for internal use. PositionKey() and
	PositionFromKey() are much faster. The global speed improvement is
	about 10%.
	A side effect is that the evalution cache becomes 38% larger.

2011-07-09  Philippe Michel  <address@hidden>

	* gnubg.c, gtkgame.c: Set players names from default_names at
	program launch and when pasting a position id, in addition to when
	starting a match.

2011-07-08  Philippe Michel  <address@hidden>

	* play.c: Update players names in game record panel when starting
	a new match.

2011-06-13  Michael Petch  <address@hidden>
	* analysis.c: getLuckRating was returning values based
	on old hard coded luck thresholds. Use the user supplied 
	ones instead 

2011-06-08  Michael Petch  <address@hidden>
	* gnubgmodule.c: Add new python function "DiceRolls" to 
	return the next "n" rolls from the current RNG. 

2011-05-18  Michael Petch  <address@hidden>
	* configure.in:  --enable-threads and --enable-extmatchid are 
	now the default. amd64 cpu types will automatically enable sse2

2011-05-17  Michael Petch  <address@hidden>
	* gtkgame.c: Modify the Menu to support copying of GNUBGID's,
	as well as the MatchId or positionID's individually.

2011-05-17  Michael Petch  <address@hidden>
	* configure.in: Modify the ./configure process to now enable
	Extended match ID's by default. They can be disabled with
	--disable-extmatchid

2011-05-17  Michael Petch  <address@hidden>

	* gnubg.c, backgammon.h, gtkfile.c, gtkrelational.c: The word 
	"Relational" has been removed from the GUI when referring to the 
	database. This makes more sense to the casual users. Fixed the 
	situation where a Batch sometimes had to be preceded by a clear 
	match/session. Fixed bug reported by Leo that prevented matches 
	being analysed with incomplete analysis settings. Fixed bug
	reported by Leo Bueno that prevented matches that may not have been
	played to completion from being analysed. Added Warning if you 
	use incomplete analysis settings and have requested to save the 
	matches to the database

2011-05-13  Philippe Michel  <address@hidden>

	* eval.c, lib/inputs.c (new file), lib/Makefile.am: Computation of
	baseInputs() with SSE. The code was already there in eval.c, but
	commented out.

2011-05-08  Philippe Michel  <address@hidden>

	* lib/gnubg-types.h, lib/cache.h, lib/cache.c, backgammon.h,
	positionid.h, positionid.c, analysis.c, eval.c, export.c, eval.h,
	gnubg.c, gnubgmodule.c, gtkboard.c, gtkpanels.c, import.c, play.c,
	rollout.c, set.c, sgf.c, makebearoff.c, makehyper.c:
	Use a defined type for the position key.
	Small speed up with faster EqualKeys and CopyKey macros.

--

Michael Petch 
CApp::Sysware Consulting Ltd. 
OpenPGP FingerPrint=D81C 6A0D 987E 7DA5 3219 6715 466A 2ACE 5CAE 3304