F# is a cross-platform, functional programming language. It features a strong type system, type inference, type providers and a lot of the niceties of OCaml. These Docker images let you get started with F# quickly and build applications based on F#.
The mono image contains the latest mono version with F#, whereas the core
image contains the latest .NET Core SDK and runtime.
You can choose between using a SDK image or a Runtime image. The Runtime image is a minimal image with only the dependencies needed to run your already-compiled F# code, while the SDK image also has capabilities for compiling F#.
You can create runtime images by using a continuous integration services to
build the software; reference FSharp.Core from your app/service and copy the
final output into the Runtime container to create a container that contains
just your compiled app.
The up-side of doing it this way is a drastically reduced footprint of your image which translates to lower storage-costs and quicker deploys. When that bug hits the fan, you want to be able to quickly roll back your version or push a fix – which means you benefit from lower image sizes.
Once Docker is installed on your machine, type: docker pull fsharp as root.
When Cannot connect to the Docker daemon at {..} Is the docker daemon running? appears:
Run systemctl start docker in order to start it. In order to autostart the docker service, use systemctl enable docker.
- Mono 5.20.1.19
- .NET Core 2.2.203 SDK
The starting command for the mono image is fsharpi so running the image
will result in an F# interactive REPL.
You can get an FSharp interactive session by simply running the fsharp image:
# docker run -it fsharp
Microsoft (R) F# Interactive version 10.2.3 for F# 4.5
Copyright (c) Microsoft Corporation. All Rights Reserved.
For help type #help;;
> ["FSharp"; "in"; "Docker"] |> String.concat " " |> printfn "Hello from %s";;
Hello from FSharp in Docker
val it : unit = ()
You can also build and run some F# code in a container. This command will write a short snippet to a Test.fs file, compile it, and then run it.
# docker run -it fsharp bash -c "echo 'let [<EntryPoint>] main argv = printfn \"Hello from FSharp in Docker\"; 0' > Test.fs && fsharpc Test.fs && mono Test.exe"
Microsoft (R) F# Interactive version 10.2.3 for F# 4.5
Copyright (c) Microsoft Corporation. All Rights Reserved.
Hello from FSharp in Docker
With host volume mounts, you can attach a local source directory from
your host machine, then do the build in a Docker image. Below is an
FSharp source file in the current directory. Map that directory on
the host to the /src/ directory on the container with
--volume `pwd`:/src
Then run the compiler and finally the application in the container.
After the applications exits, the resulting executable is left.
# ls
Program.fs
# docker run -it --rm --volume `pwd`:/src fsharp bash -c "cd /src; fsharpc Program.fs && mono Program.exe"
Microsoft (R) F# Interactive version 10.2.3 for F# 4.5
Copyright (c) Microsoft Corporation. All Rights Reserved.
Perimeter of Circle 2.2 = 27.646015
Perimeter of Triangle (3.0,4.6,2.8) = 10.400000
Perimeter of Square 9.1 = 36.400000
# ls
Program.exe Program.fs
To build an image of your own based on the fsharp image, create a Dockerfile
such as the following:
FROM fsharp
WORKDIR /app
COPY Program.fs /app
RUN fsharpc Program.fs
ENTRYPOINT ["mono", "/app/Program.exe"]
This Dockerfile will copy the Program.fs file from the current directory into the image. Then it will change to that directory and compile it. The final step sets the entrypoint, so the resulting Docker image always runs the application.
Build this with docker build -t myapp ..
Then run with docker run --rm myapp and see that the application is run
directly. The compilation step was performed when building the image, so this
image is ready to execute immediately:
# docker run --rm myapp
Perimeter of Circle 2.2 = 27.646015
Perimeter of Triangle (3.0,4.6,2.8) = 10.400000
Perimeter of Square 9.1 = 36.400000
The entry point for the core image is also fsharpi as there is no interactive
F# REPL available for .NET Core at this time.
Without a REPL, the primary use cases for the core image are building .NET
Core applications or using it as a base for another image.