|
|
| (1) Posted by Thomas Maeder [Tuesday, Jun 10, 2008 00:05] |
Popeye 4.47
[Back from the Stade de Suisse (congratulations to the Dutch team for the great show!), I find that sourceforge.net is finally working again, allowing me to make the following overdue announcement:]
The Popeye team is happy to announce Popeye 4.47, available on http://sourceforge.net/project/showfiles.php?group_id=200122
As always, feedback is very welcome; please use the appropriate "Tracker" on http://sourceforge.net/tracker/?group_id=200122.
Additions
=========
Platform
--------
- pywin64.exe (part popeye-4.47-windows.zip): executable for 64bit Windows with considerable speed improvement compared to pywin32.exe
Condition
---------
- Losing chess (Schlagschach, QuiPerdGagne)
Bug fixes
=========
Solving behaviour
-----------------
- Take&make: pawns can't "make" to their base line any more
- Sentinelles + magic pieces: Sentinelles pawn doesn't change colour in the move that produces it
- Masand proof games are now solved correctly
Input handling
--------------
- reduced cases where option intelligent is accepted when it shouldn't
- condition RoyalDynasty can be specified after pieces even if there is more than 1 king per side
Known limitations
=================
pywin64.exe doesn't support the maxtime option
Thanks to Vaclav Kotesovec for his support in testing pywin64.exe!
|
|
| (2) Posted by Joost de Heer [Thursday, Jun 12, 2008 15:29] |
boolean.h seems to be missing from the source distribution.
Joost
|
|
| (3) Posted by Thomas Maeder [Thursday, Jun 12, 2008 20:20] |
QUOTE
boolean.h seems to be missing from the source distribution.
Thanks, Joost!
I have just updated the source package; this one should work now.
|
|
| (4) Posted by Joost de Heer [Friday, Jun 27, 2008 10:24] |
r# (and perhaps more?) solving seems to be broken. I was playing around a bit with a reflexmate from Schiffmann (see Michael McDowell's post), and that resulted in the following:
BeginProblem
Author Unknown
Origin Unknown
Stipulation r#2
Protocol D:appsschakenpopeyefancypy.txt
Pieces
White Pe7 Pf7 Ka4 Bb4
Black Sf8 Kc7 Pb6 Pd6 Pd5 Rg4 Ph4 Sa3 Pb3 Pc3 Qh3 Ra2 Bd2 Pe2 Bd1
Option Try Variation NoBoard
EndProblem
Popeye Windows-32Bit v4.47 (4 MB)
1.e7-e8=Q ! threat:
2.Qe8-b5
2...b3-b2 #
2...Sa3-b1 #
2.Qe8-e3
2...b3-b2 #
2.Qe8-e5
2...b3-b2 #
2.Qe8-e6
2...b3-b2 #
2.Qe8-a8
2...b3-b2 #
2.Qe8*f8
2...b3-b2 #
[..]
1...Sa3-c4 +
refutes.
A correct key with a refutation?
|
|
| (5) Posted by Juraj Lörinc [Friday, Jun 27, 2008 12:47] |
The same strange behaviour in Popeye Windows-32Bit v4.45. I suppose Joost has called that from Fancy interface, just like me, but input file looks ok in the window anyway.
|
|
| (6) Posted by Thomas Maeder [Friday, Jun 27, 2008 20:18] |
QUOTE
r# (and perhaps more?) solving seems to be broken.
Thanks Joost and Juraj!
The bug you found looks similar to this one:
http://sourceforge.net/tracker/index.php?func=detail&aid=1914585&group_id=200122&atid=972237
[Would you please report this new bug in the bug tracker? Thanks!]
Typically, the reason for these problems is the "hash table" Popeye uses to avoid solving the same "sub-problem" twice (let alone multiply). Popeye solves direct problems in two stages:
1) determine whether a particular white move is a continuation
2) find the entire solution after this move
Note that 1) doesn't require looking for duals; a lot of time is saved by considering every defence dealt with when one continuation has been found. In stage 2), all continuations are searched, and the solution (or try) is printed.
There are certain situations where something remembered in phase 1) is mis-interpreted in phase 2).
|
|
| (7) Posted by Thomas Maeder [Sunday, Jun 29, 2008 13:44] |
QUOTE
The bug you found looks similar to this one:
... but turned out to be something completely different ...
The bug has probably existed since Popeye solves reflex stipulations.
Fixed for releases >4.47.
Cf. http://sourceforge.net/tracker/index.php?func=detail&aid=2004485&group_id=200122&atid=972237
|
|
No more posts |
MatPlus.Net  Forum Internet and Computing Popeye 4.47 |
ISC 2024
