At first I thought I just couldn't do it. But then I looked up a solution and it appears that not every combination is even solvable. It appears that most of the time I end up with an unsolvable pattern (everything correct except for the last row where 14 and 15 are exchanged). I once solved the puzzle, although since I can't reproduce that I'm not so sure any more. There was no notification that I made it, I don't know if there is supposed to be one. Reproducible: Sometimes
Created attachment 104697 [details] pull request for patch of fifteenPuzzle logic I submitted a pull request with a fix for the problem but realized this is probably not the correct way to do it, since my pull request is the only open one. Could someone please advise me how to submit this patch?
I experienced the same problem, I never got a solvable puzzle and am convinced the developer has just switched around the logic, so only unsolvable puzzles are generated. The patch is just to change one 0 to a 1, see my other comment.
@Florian: https://community.kde.org/Infrastructure/Phabricator
Your patch doesn't match your comment. Note that test for inversions being even or not is not the == 1 at the end but rather then (inversionsMod == 0) which is already there in the current code. Note the two sets of == in the same statement I think you've accidentally changed it to: if the size is even and the inversions are even, it's not solvable.
Git commit 3516eda8107bda0c91fa8fa5ef7e6da34d995f6b by Igor Poboiko. Committed on 22/05/2017 at 15:51. Pushed by poboiko into branch 'master'. Fix fifteen puzzle solveability Differential Revision: https://phabricator.kde.org/D5913 M +2 -2 applets/fifteenPuzzle/package/contents/ui/FifteenPuzzle.qml https://commits.kde.org/kdeplasma-addons/3516eda8107bda0c91fa8fa5ef7e6da34d995f6b