Summary: | Solver gives wrong information | ||
---|---|---|---|
Product: | [Applications] kpat | Reporter: | KOLI <karola.bvb> |
Component: | solver | Assignee: | Stephan Kulow <coolo> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | kde-games-bugs-null, MarathonToo |
Priority: | NOR | ||
Version First Reported In: | 3.6 | ||
Target Milestone: | --- | ||
Platform: | Ubuntu | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
KOLI
2015-06-24 21:59:48 UTC
In a Freecell game, I reached a state in which the solver in the bottom left of the screen said "this game is winnable", but every time I made a non-reversible move, the solver said "this game is no longer winnable." Eventually, I gave up, and clicked the "demo". After a few moves, the demo made an illegal move, placing a red 3 on a black 9. I undid some moves and ran the demo again. This time, after a few moves, the demo entered a cycle in which the sequence of states of the game were duplicated. I did this twice. The first time, the cycle was very short: It moved a red queen back and forth between two black kings. I am not able to give the steps to reproduce this, because, in this version, the game number is not available. In previous versions, the game number appeared on the window title bar. :-( I just tried the demo again. This time, after many moves, it placed a black 8 on a black 8. The solver has always been reliable before. I wish my comment to be considered withdrawn. I recalled that I had taken a break during that game, and saved it. In loading that file, I cannot recreate the software errors I reported. I played some moves and clicked the "demo" button a few times. I never did catch it making an illegal move. It did solve the game, but it took over 500 moves. There were some strange sequences, such as moving a red 2 from a black 3 to a free cell, and then moving the other red 2 from a free cell to the black 3. Inefficient, but not illegal. KPat is now using a completely new solver. Let's assume the new solver fixes the issue. If not, please report to freecell-solver developers via https://github.com/shlomif/fc-solve |