Environment Setup

Setting up a programming environment

You will need an x86-64 Unix environment to run the class tools and ensure that we can build your compiler on the grading server.

Our officially supported programming languages are Java, Scala, Rust, and Typescript. If you are unsure which language to choose, you should default to Java.

If you’re not already on a Unix environment (e.g., Linux, MacOS, WSL, MinGW), you can set up a virtual machine or use WSL. Alternatively, you can use the Athena environment.1

Since you’ll be doing a lot of coding in your chosen environment, it’s worth investing effort into customizing it to be comfortable and productive for you. This might include installing your favorite text editor, tweaking your VM settings, or figuring out how to make your favorite IDE work with the class tools. A common solution is to develop on your local machine and run the compiler on the Unix environment (synchronyzing code with a shared folder, scp, rsync, sshfs, or git).

A note: when developing across multiple environments, it is very easy to lose your code (including overwriting local git history). Push your code early and often. Take advantage of git branches to develop code without breaking your partners’ builds.

You may wish to install the same version of your programming language and build system as the grading server. Details about the grading server are available on the Autograder page.

ARM Environments

If you have an ARM environment, you can install an x86-64 Linux VM using OrbStack. If you are using a recent (M1/M2) Mac, Rosetta 2 should allow you to run x86-64 programs. You can manually run a command in emulation mode using arch -x86_64 <program> (this is helpful for calling an x86-64 gcc/clang on your assembly output). If you are using another ARM environment without the ability to emulate x86-64 programs, you can develop on Athena. In either case, you may be able to use QEMU or equivalent to run an x86-64 Linux.

It is your responsibility to ensure that your compiler is entirely compatible with the x86 grading environment – you can test this on Athena. If you are having issues, please contact the TAs.

Ubuntu

Subsequent instructions assume that you’re on a default Ubuntu installation. If you’re completely new to Linux, take some time familarize yourself with it. You will need to use the terminal at least to run the test scripts.

We will be using Git to distribute test scripts and handle code submission. You should also use Git to collaborate with your teammates. If you’re not familiar with basic Git usage like resolving merge conflicts, read Using Git. That also includes instructions for setting up Git.

Finally, set up the environment for your chosen language. If you’re using Java or Scala, install the JDK:

sudo apt install -y openjdk-17-jdk

If you’re using Scala, also install it:

sudo apt install -y scala

Make sure the environment variable SCALA_HOME is set to “/usr/share/scala”, and that scala -version reports that you have Scala 2.13.

Further, follow these instructions to install Scala Build Tool

macOS

If you are on macOS, the equivalent commands would look like this:

brew tap homebrew/cask-versions
brew install --cask temurin17

If you run into trouble because you have multiple OpenJDK versions installed, see this article about switching between the different versions of Java on macOS.

If you are using, Scala, make sure you install it and Scala Build Tool as follows:

brew install scala@2.13

Make sure the environment variable SCALA_HOME is set to “/usr/local/opt/scala@2.13”, and that scala -version reports that you have Scala 2.13.

Further, follow these instructions to install Scala Build Tool

Athena

To install Java 17:

mkdir -p ~/6035-deps
pushd ~/6035-deps
curl -L https://download.oracle.com/java/17/archive/jdk-17.0.6_linux-x64_bin.tar.gz | tar xz
echo "export PATH=\"$(pwd)/jdk-17.0.6/bin:${PATH}\"" >> ~/.bash_environment
export "PATH=$(pwd)/jdk-17.0.6/bin:${PATH}"
popd

  1. While the class was traditionally run on Athena, we recommend against using Athena this semester. The dialup can be slow and isn’t suited for computationally intensive work like optimizing programs. ↩