As a break in between programming sessions working on my game I thought that writing posts on the features that i’ve been working on and any issues (being Rust related or not) that i’ve encountered whilst implementing them would be a fun idea. As i’ve already started working on this project before starting this blog the first few posts will most likely be on things i’ve implemented a while ago and may happen slightly out of the order they were implemented in. Hopefully as these go on i’ll get better at writing these posts too.
As you may have guessed from the title the game is being implemented in Rust. This is mainly because i’ve enjoyed working in it for the past year so I decided now would a good time to try a larger project in it.
Coming from a background with languages like Java, Go (etc) jumping into Rust was challenging for me. Getting used to not relying on a GC to clean up for me and trying to get the hang of borrowing rules took a bit of time and I still haven’t completely mastered them yet.
One of my first projects in Rust was a game I created for my final year project at university. The game was a platforming game with destructible terrain and multiplayer. With deadlines it didn’t end up being the neatest thing i’ve programmed but results turned out well.
This project was the first time I used Glutin vs something like SDL or glfw. It worked well for the cases I needed for this project but I discovered inconsistancies across platforms when I tried it for Steven (i’ll talk about that project later). I haven’t really looked at glutin lately but admittedly I used the project pretty early on into Rust’s stable lifetime as it had had only just hit 1.0 when I started the project.
Even though the project hasn’t been touched in months it still builds on the latest version of Rust (after a cargo update because some dependencies did break between versions).
The other largish project i’ve done in Rust is a re-implementation of the Minecraft client in Rust called Steven which I started around the same time as the previous project. This was actually a rewrite of my client I wrote in Go a while back (apparently I like learning new languages by programming Minecraft clients in them) as it seemed easier to work off a existing codebase than start from scratch. This project is public unlike the other one which allowed me to get some feedback about the implementation.
Steven did really show how slow compiling can be in rust as your project gets larger. A large part of the compile times was down to the block macro define_blocks!. Even with that the CI used to take over an 1.5 hours to build a debug (cargo build) and a release cargo build --release build. Looking at the CI now however (it always uses the nightly build of rustc) its down to ~30 minutes which is a huge improvement (more noticable when not doing a clean build) but still can be unbearable when trying to test changes quickly. Incremental builds are still being worked on for rust but hopefully should be arriving soon.
Planning not being my strong point i’m just making up the game as I go with a rough idea in my head of whats going to happen. Its going to be a management game set in a school/university type enviroment. I’m basing the art style of the game of the game Theme Hospital which is a game that loved to play when I was younger, it had a great sense of humour and a lot of character.
The game is rendered in 3d at an isometric perspective instead of 2d isometric sprites which greatly simplifies the sorting issues isometric games generally have. I also get the benefit of using techniques like shadow mapping to get real-time dynamic shadows as well.