These projects are meant to get you warmed up with programming in the C++/UNIX environment. None are meant to be particularly hard, but should be enough so that you can get more comfortable programming.
Realize the best thing you can do to learn to program in any environment is to program a lot. These small projects are only the beginning of that journey; you'll have to do more on your own to truly become proficient.
Tip of the hat to Remzi and Andrea Arpaci-Dusseau for the amazing textbook and projects, which we borrow from heavily.
In this class, the supported configuration is to install Docker Desktop and Visual Studio Code on your local computer, or to use the Visual Studio Code instances on the CSIF lab machines.
To install these packages, follow the installation instructions here. We already have your devcontainer and other settings configured, and Visual Studio Code will manage Docker for you, but the instructions have links to instructions for installing Docker on Windows, Mac, and Linux machines.
Conceptually, Visual Studio Code is an application running on your computer. It creates a virtual machine, using Docker, that has the right version of Linux for this class. Visual Studio Code handles synchronization between files on your local computer and the virtual machine and exposes a terminal where you can use the command line to interact with your code running within the virtual machine, which is also called a dev container.
In a typical configuration you will edit your code using Visual Studio Code, and compile and run your code using the terminal.
To debug your code using the Visual Studio Code debugger, you need to run your program using the terminal and then use one of our pre-set configurations to attach to your running program. Conceptually, there is a server running in your container that manages the Linux side of debugging, and Visual Studio Code connects to this server to issue debugging commands and display the state of your program within the editor.
To debug a program, you run it as follows:
First compile it without address sanitizer
make clean; DEBUGGER=true make
In the Visual Studio Code terminal
- [Linux, Windows, Mac (x86) host]
gdbserver localhost:1234 ./gunrock_web - [Mac (Apple Silicon) host]
ROSETTA_DEBUGSERVER_PORT=1234 ./gunrock_web
Then in Visual Studio Code editor
- Select "Run and Debug" on the left hand control pane
- Select the "GDB Gunrock Web" configuration and hit the play button
This will connect your Visual Studio Code debugger to your program running in your container.
We added a debug target for Mac hosts that will use the Mac native debugger without Docker. Note, this configuration is not supported, but if you're stuck and just need to get your project done many have had success with it. To use it you:
- Open your project and do not open in a dev container.
- On the terminal, compile with
make clean; DEBUGGER=true make - Make sure that the utility you want to debug is visible in your editor. For example, if you're debugging
ds3catthends3cat.cppshould be open in your editor. - Go to the VSCode debugger panel and select the "Mac Debug Current File" option.
Just to reiterate, we can't support this configuration yet, but if you're stuck you can try it as an option.
Here are a few hints to help, but Visual Studio Code is widely used software so checking online for help when you get stuck will be useful:
- Select "Auto Save" in Visual Studio Code to ensure that it saves your files as you edit.
- In our examples we use port
1234for the debugger, but you can change this port if needed, just make sure that you update yourlaunch.jsonfile to match. - To get started with a project, you will typically clone our ECS 150 repo on GitHub using Visual Studio Code "Clone git repository" option and Visual Studio Code will set up a Dev Container for you based on the confiugrations we included in that repo.
- If we push an update in the middle of a project, you can use
git pullin your host terminal to fetch the latest updates. There is probably a way to do this in Visual Studio Code as well.