Alith-Game-Studio / WhatTheWitness

GNU General Public License v3.0
17 stars 2 forks source link

Answer check for deformed panel #2

Open competor opened 1 year ago

competor commented 1 year ago

Hi. I'm making my puzzles using this tool and have one trouble. I tried to make a deformed panel of a 4x4 board with some edge parts missing, but this software gives me NG processing. スクリーンショット 2023-06-07 161014

The screen on the right is from the site. This is the only intended one. How can I fix this problem?

This is my .wit file. https://gist.github.com/competor/32e89cd5b590c83bd2fcfa161609b58e

instr3 commented 1 year ago

Hi competor,

The reason is that there are extra facets (cells) in the puzzle. Facets are associated with vertices instead of edges, so by deleting the edges, the facets will not be automatically deleted.

You can see the facets are still there by dragging the map in the editor. If a place shows a dot then there exists a facet.

image

You can simply select them and delete them, then the puzzle should work as intended.

image

competor commented 1 year ago

Oh, thank you it worked! I overlooked the "Delete elements" function in the editor.

competor commented 1 year ago

There is one more related question. I would like to create a board with directly erased Edges without using BrokenDecorator to make the board view simpler and easier to understand. As in the example below. スクリーンショット 2023-06-19 094453

This is the path that I expected. スクリーンショット 2023-06-19 094032

But when I erase the edges, this one is judged as incorrect. スクリーンショット 2023-06-19 094909

This is my .wit file. https://gist.github.com/competor/a495f18a8655bcdad241467c083a6994

instr3 commented 1 year ago

I think now the problem comes with the game client. The game client assumes the graph is topologically "correct":

The editor was created long before the game client so they have different restrictions.

While the code can be modified to judge Tetris correctly in an "ill-defined" graph, I would not recommend doing so since there are other components that rely on the "correctness" of the graph (e.g., correct region segmentation, triangle judging etc).

Therefore, I recommend avoiding this for custom levels.

competor commented 1 year ago

okay!