- User Since
- Mar 5 2016, 10:19 AM (237 w, 6 d)
May 4 2020
Apr 18 2020
@ngraham : I applied the change now, but was unable to build and test the change due to git master requiring newer KDE deps. I hope it works.
- Unify signals - returnPressed->focusViewRequest.
Apr 14 2020
Apr 12 2020
@aacid : hi! The get_next_move() function already returns non-zero when there are no more moves left, so we can rely on checking its return value.
Closing then. I don't have any immediate plans for reworking this patch.
Apr 10 2020
Apr 9 2020
@sitter : you also need to revert the changes inside the main.cpp's if (ok) conditional. Also note that the GolfSolver autotest fails with your patch if one configures kpat to build using this lib of mine: https://www.shlomifish.org/open-source/projects/black-hole-solitaire-solver/ .
Changing fprintf(stdout to KPAT_LOG in main.cpp is likely wrong because it interferes with the output of the commandline range solver. The other changes seem fine.
Mar 24 2020
@ngraham @meven : here is a new patch: https://www.shlomifish.org/Files/files/code/dolphin--shlomif-patch-D26362--modified-v0.2.0.patch but note that it makes no difference and om fedora 32 x64 it seems that pressing enter twice invokes the top item in the results which was already selected. It also can be done as "down arrow; enter" using either of my patches.
@ngraham : hi! I am interested in finishing it up, but will be fine with someone else doing that. I'll try to get to it soon.
Jan 19 2020
Jan 16 2020
- Readability improvement per coates' advice.
- fix the non-randomised deal problem.
- Use specialized QRandomGenerator methods.
Jan 3 2020
Convert to emit signal per @ngraham 's comment.
Jan 2 2020
Here is a fixed startRandom():
Dec 24 2019
Thanks for the patch! It looks fine at first glance, but I noticed that when you deal the first game after the game variant carousel ([kpat start] → Freecell) the deal index in the window's title is always "1". It doesn't happen with the current git master ( 96de699e8fe1b6c56881b29761384b5a600ad6ab ).
Aug 14 2019
Aug 11 2019
Aug 9 2019
Jul 31 2019
- Initialize to a default value.
Jul 15 2019
Jun 1 2019
Seems like it was already pushed in https://phabricator.kde.org/R410:d8c4403f115a69a6b1365ab63d1c7c6787d3f6c3 .
@aacid : hi, can we land this?
May 14 2019
May 9 2019
May 3 2019
May 1 2019
Apr 30 2019
- Remove unneeded "virtual"s due to override.
Already applied as part of commit 9b86b67bb33cc011e01ff5466ef88fe80f514583 to git master.
Apr 29 2019
Apr 25 2019
Apr 23 2019
@arojas: I pushed it to 19.04 too - thanks.
Apr 22 2019
- Add a comment to explain.
I think we can have an autotest for it, but it will require some work.
Apr 21 2019
Apr 16 2019
Apr 14 2019
Apr 8 2019
Mar 15 2019
Mar 12 2019
- Add an assert and a test for bz405187
Mar 11 2019
Mar 8 2019
Feb 19 2019
Feb 14 2019
- Avoid having a virtual solverFormat().
@aacid : well, there are similar solverFormat() methods in Freecell and in Simple Simon, but they all can be non-virtual and I guess a static_cast will work.
Feb 13 2019
@aacid : re making it pure virtual that would require every subclass to supply it and it isn't needed or used by many of the game variant subclasses.
Feb 9 2019
Convert to qCCritical and fix a typo.
Feb 8 2019
Feb 7 2019
@aacid : OK, thanks for the review. I think I will commit my changes.
Feb 5 2019
- avoid compile warnings.
First 20,000 deals:
Better formatted table:
For the record, the kpat cmd line I used was ./kpat --start 1 --end 1000000 --solve 12 2>&1 | tee -a new-golfs.txt
I found out that kpat has a command line range solving mode. Here are the results of the first 5,000 deals:
Feb 4 2019
@aacid : for deal no. 5 (starting position) the old solver reports that "This game cannot be won", while my solver solves it quickly and to completion. I can try collecting statistics on a range of deals wrt the performance of the two solvers.