png2rex VS OneMoreLevel

I've ported my old Png2Rex for C++ over to Rust. It's currently exactly the same, just native Rust - so i don't have to fire up a C++ environment to use it. I expect it'll improve a bit, I have a few ideas - and the Rust image system is really powerful. It was interesting porting such a simple tool, and highlighted the differences in development philosophies. The C++ version baked in its own PNG reader, because dependencies are such a pain. This time, I just imported image with the png feature flag. Likewise, instead of including a Rex Paint reader/writer, I just imported the one from bracket-lib and used it. The result? I made this tool, including unit tests and far better error handling, in about 30 minutes. :-)

It is not a technical issue. I draw four maps with REXPaint, which is quite handy. I use two Godot scripts (FileIOHelper.gd and DungeonPrefab.gd) to read a txt file and output terrain data. I can flip the map horizontally and/or vertically if necessary. Problem is, in my mind, ninjas should enclose on all sides and our ninja slayer keeps dodging attacks until the time is right. He fights back with lightning speed and covers narrow corridors and chock points with soul fragments (yellow question marks). But the game does not play like that. I need more time to find a solution.

This week I have raised the difficulty of a few dungeons. I have also added more colors based on players' feedback which greatly improves game experience (see demo above). Therefore I published a nightly release on GitHub and itch.io. I picked all the colors ([grey](f8f9fa-e9ecef-dee2e6-ced4da-adb5bd-6c757d-495057-343a40-212529), green and orange) from Coolors. Below are their hash values.

Demo | Github | itch.io | Twitter | Dev Blog