https://bugs.kde.org/show_bug.cgi?id=92097
--- Comment #15 from JStrebel <joerg.stre...@gmail.com> --- > I was also thinking of implementing this option. So, to me at least, this is > wanted. However, I was simply going to port over the algorithm given in > https://www.chiark.greenend.org.uk/~sgtatham/puzzles/js/mines.html, which I > see you've already considered (and rejected?). The thing with this algorithm is that it claims to always produce solvable minefields, but I am not so sure. This algorithm is also used in the Android app "Antimine" (https://play.google.com/store/apps/details?id=com.logical.minato&hl=en_US&gl=US). I have been playing ca. 30 games there on medium difficulty and I ran into numerous situation, where I had the impression that only guessing would help me (and I have screenshots for others to check my assumption). So I am not so sure that this algorithm is really effective. I said above that KMines should be easy to extend with the Gecode library, but I am not so sure anymore. KDE games use a pretty intricate build system. I tried to build the current version of KMines on openSUSE Leap 15.3, but I failed because of outdated dependencies. Maybe I need to build a fork first and then investigate how to incorporate my changes into the official KMines repo. -- You are receiving this mail because: You are watching all bug changes.
