The New Sudoku Players' Forum

[yzfwsf]

With default settings ,Hodoku can't get any Hints for :0372:7:2..1.98.7..9.2.1.3.718.5.2..3.29164.412.86.7.96....2.1.269.35...9.658..25...12.96:513 415 422 435 763 464 764 365 568 871 493 497:771::

[SpAce]

With default settings ,Hodoku can't get any Hints for :0372:7:2..1.98.7..9.2.1.3.718.5.2..3.29164.412.86.7.96....2.1.269.35...9.658..25...12.96:513 415 422 435 763 464 764 365 568 871 493 497:771::

Interesting. I haven't seen that problem before. Maybe a bug with the XY-Chain implementation? I got it to work when I switched off XY-Chains in "Solver" so it hinted an AIC instead. Curiously, the "Find all steps" didn't work until I switched off Remote Pairs, X-Chains, and XY-Chains in "All possible steps", so perhaps not just an XY-Chain problem, though.

[yzfwsf]

Bug seems to stay in "Ristrict chain lengths".If uncheck steps>>Chains (X,XY,Remote Pairs)>>Ristrict chain lengths then Hodoku work fine.

[yzfwsf]

Another puzzle:003000060500061000070200008400007000020600010005000003050042000800090070004800320
When paste this puzzle,Hodoku stuck.When uncheck "Resrict chain lengths"

[StrmCkr]

its a size limit for chains that prevents it from searching for the solution path

in the" all possible steps" tab left click on the gear button -> left click on the steps tab
under the

chains (x,xy, remote pair) either change the [0 - 39] chain length limit drop down list to a number higher then 24

or

uncheck the
restrict chain lengths box.

and it then finds a solution easily.

[SpAce]

its a size limit for chains that prevents it from searching for the solution path
...
chains (x,xy, remote pair) either change the [0 - 39] chain length limit drop down list to a number higher then 24

Thanks, StrmCkr! That fixes it and also explains why the problem only affected those three types of chains. So the actual bug is in the UI because it seems to hang instead of stopping the search for those chains. Doesn't seem like a common problem anyway as I've never seen it before.

[SpAce]

What about this:

Code: Select all

.-------------.-------------.------------.
| 18   9   18 | 2    5   4  | 7  6    3  |
| 7    3   2  | 68   18  16 | 4  5    9  |
| 4    5   6  | 9    3   7  | 8  2    1  |
:-------------+-------------+------------:
| 23   17  5  | 347  9   8  | 6  134  27 |
| 9    6   38 | 1    47  2  | 5  34   78 |
| 28   17  4  | 37   6   5  | 9  13   28 |
:-------------+-------------+------------:
| 6    8   9  | 5    2   3  | 1  7    4  |
| 15   2   7  | 46   14  9  | 3  8    56 |
| 135  4   13 | 78   78  16 | 2  9    56 |
'-------------'-------------'------------'


The "Find all steps" hangs with "Searching: ALS". Curiously, if I uncheck ALS-XY-Wings it works (and finds several ALS-XZs, ALS-Chains, and Death Blossoms). I would have expected the problem to be with the longer ALS-Chains instead of ALS-XY-Wings. Is there a fix for this?

[yzfwsf]

At this point, Hodoku cannot find any grouped AICs, but my solver will find them.

Code: Select all

:0000:x:+421+35...6...7..3.+4.+3...4.1..1.89.4.+33..2...9+1+9.6..3..5..3..71..8...+3..4.+19.....+37:523 626 737 248 753 655 674 675 582 586 686 295 596 696::

Code: Select all

:0000:x:..3...2.....8...9..6..+3...4+28+6+79+3+4+155+3+9+614+7+28..72..+63+9..43+7.9..6+9.+42....+37.+9.6.4.:111 411 515 116 516 122 525 126 526 729 131 731 133 533 136 536 187 189::

[jco]

Hello,

This is related to the last post. I am reporting this hoping that the this problem has being fixed (in this case, please let me know how).

In the configuration below

Code: Select all

.--------------.-------------.----------------.
| 9   46  2346 | 37  28  5   | 1    678  378  |
| 8   35  7    | 6   4   1   | 2    35   9    |
| 26  1   235  | 37  28  9   | 678  4    3578 |
:--------------+-------------+----------------:
| 27  48  249  | 5   1   237 | 39   278  6    |
| 3   68  256  | 4   9   27  | 78   1    2578 |
| 1   57  259  | 8   6   237 | 39   257  4    |
:--------------+-------------+----------------:
| 67  2   36   | 1   5   8   | 4    9    37   |
| 4   9   8    | 2   37  6   | 5    37   1    |
| 5   37  1    | 9   37  4   | 68   268  28   |
'--------------'-------------'----------------'


HoDoKu is unable to find the move:

(2=7)r5c6 - r5c79 = r46c8 - r8c8 = r7c9 - (7=6)r7c1 - r3c1 = (6-7)r3c7 = 7r5c7 => -7 r5c6

YZF_Sudoku finds it and finds many others not shown in HoDoKu's set of possible moves.

Regards,
jco

[1to9only]

This is related to the last post. I am reporting this hoping that the this problem has being fixed (in this case, please let me know how).

HoDoKu has not been updated for many years, because hobiwan (Bernhard Hobiger) passed away some time ago.
wookie2u2 (Keith Corlett) has an 'improved' HoDoku, download link: https://sourceforge.net/projects/sudoku-explainer/files/HoDoKu.2.4.2.2020-11-23.7z/download
For the OP grid:

Code: Select all

:0372:7:2..1.98.7..9.2.1.3.718.5.2..3.29164.412.86.7.96....2.1.269.35...9.658..25...12.96:513 415 422 435 763 464 764 365 568 871 493 497:771::


wookie2u2's HoDoKu (at default settings) offered this hint:

Code: Select all

AIC: 8 => r4c13,r6c8<>8


[jco]

Hello 1to9only,

Thank you for the information and link. Unfortunately this version does not work in my machine.
I have the previous versions of HoDoKu working fine but this one, after executing the comand "java -jar HoDoKu.jar" (I type "sudo" before that command), it opens the program but the menu "preferences" does not work. Also, only part of the solution path for the puzzle appears. I also have SE installed and working fine (I discard problems with Java for this reason), so it seems to be something with this particular version.

Best regards,
jco

[1to9only]

With default settings ,Hodoku can't get any Hints for :0372:7:2..1.98.7..9.2.1.3.718.5.2..3.29164.412.86.7.96....2.1.269.35...9.658..25...12.96:513 415 422 435 763 464 764 365 568 871 493 497:771::

This is an old thread!

The fix is to uncomment these lines in SolutionStep.java:

Code: Select all

2010,2014c2010,2014
< //            // change 20110512: short chains first!
< //            int chainDiff = compareChainLengths(o);
< //            if (chainDiff != 0) {
< //                return chainDiff;
< //            }
---
>             // change 20110512: short chains first!
>             int chainDiff = compareChainLengths(o);
>             if (chainDiff != 0) {
>                 return chainDiff;
>             }


However, those lines were commented out by hobiwan a long time ago to fix an ALS-Chains sorting problem, see here.

Code: Select all

Commits on Mar 31, 2012
Fixed: #3513603: Search for all steps ignores ALS types
fixed a problem with sorting: ALS-Chains were sometimes sorted incorrectly


So fixing this issue will likely bring back the old ALS-Chains sorting problem!!

PseudoFish's Hodoku fixed the XY-Chain problem here.

I have the fix (uncommenting the 5 lines) in a private Hodoku build - 'Next hint' gives this XY-Chain:

Code: Select all

XY-Chain: 1 1- r7c1 -7- r7c5 -4- r9c4 -7- r9c7 -3- r8c8 -1 => r7c8,r8c1<>1


I have not checked the other grids posted in this thread.

[ghfick]

I can confirm that Bernhard Hobiger died in 2014. There was an obit posted :

trauer.kleinezeitung.at/todesanzeige/bernhard-hobiger

dated 07.05.2014 [either July 5 or May 7 ?]

I had been in some contact' with him through several of the last revisions to HoDoKu until 2012. I think he had a number of plans for HoDoKu but he had advised me that he needed to concentrate on his 'real' work. In any case, he did not return to HoDoKu.

ZF Yang is carrying the torch with YZF_Sudoku. I think Bernhard would have been pleased with Yang's advances.

Gordon