Bug with L piece

[FlanGrande]

Hello there,

I was planning to use your project as a starting base for a game I am making. While I decided to make a lot of the code from scratch, seeing your Godot implementation helped me structure my code better.

However, when playing yours, I noticed that when you rotate the orange L piece it goes into the wall for a short moment and then it gets pushed back into the level but in an awkward offset that takes it out of the grid.

Interestingly, when I was developing mine, I wasn't tracing your code line by line or anything, but somehow I have the same bug where the L piece is able to clip into the walls. The weirdest thing is that in both of our projects it doesn't happen with the blocks that are already placed.

In my case I'm using world boundaries, so I thought that might be it, but in your case you are using StaticBody2Ds for those too.

I am so puzzled by this that I needed to share it with you. Maybe we can come with an explanation as to why this happens? I sorted it out on my code with an ugly hack, but I'd like to understand what's going on.

Thanks for putting your code out there btw

[BEsharkounet]

Hi. First of all, thank you, i take your message as a huge compliment. I have no idea about the L bug, i'll get into it tomorrow and tell you the issue for sure (if i can figure it) I did that project few time ago, i may do a big update if people use it as structure helper. Le sam. 1 juin 2024, 17:53, Flan ***@***.***> a écrit :

…

Hello there, I was planning to use your project as a starting base for a game I am making. While I decided to make a lot of the code from scratch, seeing your Godot implementation helped me structure my code better. However, when playing yours, I noticed that when you rotate the orange L piece it goes into the wall for a short moment and then it gets pushed back into the level but in an awkward offset that takes it out of the grid. Interestingly, when I was developing mine, I wasn't tracing your code line by line or anything, but somehow I have the same bug where the L piece is able to clip into the walls. The weirdest thing is that in both of our projects it doesn't happen with the blocks that are already placed. In my case I'm using world boundaries, so I thought that might be it, but in your case you are using StaticBody2Ds for those too. I am so puzzled by this that I needed to share it with you. Maybe we can come with an explanation as to why this happens? I sorted it out on my code with an ugly hack, but I'd like to understand what's going on. Thanks for putting your code out there btw — Reply to this email directly, view it on GitHub <#1>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AO7EUY7UOJW32Q7CNRF2IJLZFHU6TAVCNFSM6AAAAABIUID56KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGMZDSMRRGA3TENI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[FlanGrande]

Thanks fir replying, I really appreciate it.

If you can't recreate the bug, just note that I'm using 4.2 and if you still can't, then I can make a video and we can try to think about what's going on.

[BEsharkounet]

Na the code is to messed UP. I would also suggest to not use collision with physics for that kind of game. Also, it's very wrong because of the precision it require. To explain : to put a block between 2 others, it require to not collide pixel perfectly with the 2 others blocks to go between. I reproduced the bug. Cann't fix it but the whole code is to be fixes. Would you like me to put a new version highly improved ? Le lun. 3 juin 2024, 01:51, Flan ***@***.***> a écrit :

…

Thanks fir replying, I really appreciate it. If you can't recreate the bug, just note that I'm using 4.2 and if you still can't, then I can make a video and we can try to think about what's going on. — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AO7EUY3CYUET3JV3EE6BZCLZFOVWLAVCNFSM6AAAAABIUID56KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBUGA3DIMBYGU> . You are receiving this because you commented.Message ID: ***@***.***>

[FlanGrande]

I am reaching the same conclusion in regards to the physics. Dunno why in most cases the collision checks are reliable, but some weird edge cases just don't detect a collision.

I am working on keeping track of collisions using a grid and then I stick to the Tetris Guideline when it comes to edge cases. I might end up hardcoding some behaviours but that's probably the right way to do it without relying on collision detection.

If you want to improve the code knock yourself out! I'd be interested on seeing it. However, I don't have an use for it anymore. I really appreciate your work already it helped me lots!

I really liked what you did with the 2 "fixed blocks" and the 2 "moving" ones. I took that from your design.

Thanks for replying!

[BEsharkounet]

Hello. I could still clean a bit my code, but i did a complete rework of tetris simple game. Feel free to check and give me your thoughts. Regards.

…

On Mon, Jun 3, 2024 at 10:18 PM Flan ***@***.***> wrote: I am reaching the same conclusion in regards to the physics. Dunno why in most cases the collision checks are reliable, but some weird edge cases just don't detect a collision. I am working on keeping track of collisions using a grid and then I stick to the Tetris Guideline when it comes to edge cases. I might end up hardcoding some behaviours but that's probably the right way to do it without relying on collision detection. If you want to improve the code knock yourself out! I'd be interested on seeing it. However, I don't have an use for it anymore. I really appreciate your work already it helped me lots! I really liked what you did with the 2 "fixed blocks" and the 2 "moving" ones. I took that from your design. Thanks for replying! — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AO7EUYYMFIZXC42BMP5SXILZFTFQPAVCNFSM6AAAAABIUID56KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBWGA2DMNRZGI> . You are receiving this because you commented.Message ID: ***@***.***>