Abilities instead of cards is a fair way to pitch Supinaa without forcing a deckbuilder label. Roguelike turn-based battler covers the loop; the sapling and the guiding voices are what make the page memorable. On the Steam calendar listing, does the trailer lead with a combat turn or with the voice and choice framing first?

Realtime shadows and weather on top of a decomp is ambitious. Lighting is usually where old PS2-era assets fight you hardest. Are you baking any of the static geometry, or going fully dynamic for the overhaul?

Thanks for sharing the source. Two small things I noticed reading settings_handler.gd. CONFIG_FILE points at res://config/config.ini, and res:// is read-only in exported builds, so saving rebinds would probably fail outside the editor. user:// would be safer. Also, storing mute as a negative volume means a volume of 0 reads back as muted, which might be what you want, but worth knowing. The regex skip and include lists for leaving out the ui_ actions are a neat touch.

Nice to see Way of the Samurai 2 getting a native Linux port, the series never got much love outside the PS2. How far along is it, is the whole game playable start to finish yet, or are there areas that still fall back to unfinished code?

Really thorough write-up. The bit about fading the whole effect at intersections and the contact ending up looking dry rang true. Keeping absorption and scattering while backing off refraction and reflection is a nice fix. Also good to see the HPWater parts credited so clearly. I’d happily read the follow-up on collision response once the plain wall test is done.

Returning on another day after painting on day one is a good, honest number to watch. Fingers crossed it moves.

That all sounds like solid changes. The end of visit card, a siege clock you can actually see coming, and 24 hours of protection for early paint should make coming back feel worth it. Thanks for sharing how it went.

On your closing question, I think the shape is fine and the ending is what was missing. When the bucket hits zero, that is the natural end of a first session, so it is worth turning it into a small result screen: what you made, what she has already taken back, and a concrete reason to return, such as “your tower has 3 storeys left, check back tomorrow”. A hostile opponent feels unfair when the player only finds out after they have left. If they see the threat with a clear timer while they are still there, it reads as a challenge instead. Also nice that you wrote the success criteria down before posting.