Change the big piece coordinate system.

[millabasset]

Change the big piece coordinate system.

[Oshisaure]

Wouldn't it be easier at this point to take the other approach and actually make the piece bigger?
You can procedurally turn a piece big by taking its blocks and making a new piece with blocks at (2x;2y), (2x+1;2y), 2x;2y+1), (2x+1;2y+1), then maybe have each ruleset define an optional wallkick behaviour for big. It could default to doubling the original kicks, but can be overwritten if a ruleset wants to have half steps for kicks or even handle big pieces completely differently.

[MillaBasset]

The mass cleanup for every ruleset would be what I'm trying to avoid, as instead of returning a wallkick result like Nullpomino does (coordinates) wallkicks operate on a piece directly, therefore I'd have to go in and patch all existing rulesets. But if you have any thoughts on how to avoid this I'm open to listen.

[Oshisaure]

Hmm... Not sure it would help anyway cause what I'm thinking of really only applies if there is a kick table with a common name... Or maybe a getKickTable function, perhaps, returning the list of kicks to attempt from the piece and input states? It would allow us to have a list we can iterate on for the basic rotation behaviour, and also would allow us to essentially use the same kick table with doubled offsets. Heck that might let us make regular rotation an optional behaviour to define too, since the vast majority of rulesets can work as a kick table.
Although stuff that checks on the field (centre column rule in ARS, best fit in kon, the entirety of MTC rotation) will need to have a custom rotation code still. But in the case of the ARS rulesets it's fine, since they all extend from one another, they'll inherit the same code.

[MillaBasset]

I'm curious to see your approach in code on this one.

[MillaBasset]

After some fairly extensive testing with my method, I've determined that @Oshisaure's method would be more robust far into the future.

The bad news: A lot of rulesets will need to be refactored. Or pretty much everything using Grid:isOccupied calls. That will fall under the mode/ruleset sweeping bugfix that is detailed in #18, but expect delays on beta6.