The Dlang Online Tour

This is the D language online tour which provides an online introduction to this great system programming language with examples that can be edited, compiled and run online.

The most recent version of this tour can be seen here: http://tour.dlang.org.

Add Content

Please refer to the document CONTRIBUTING.md for further information on how to add or change existing content.

Compile & Run

Make sure dub is installed and simply run this in the dlang-tour folder:

dub

You might want to change the settings in config.yml to change the ports the tour is listening to.

Compiling & running of user source code

The source code on the tour can be compiled and then executed online and its output returned to the user. Therefore three different types of execution drivers have been implemented within the dlang-tour:

off: no online compilation allowed. If the user hits Runs an error message will be returned.
stupidlocal: an unsafe method that just runs rdmd on the local host system and returns its output. Very unsafe and musn't ever be used in production!
docker: runs rdmd within a special Docker container (dlang-tour-rdmd) and returns its output. The driver imposes the Docker container memory and execution time limits. Additional configurations options available in config.yml.

The execution driver is set in config.yml.

Caching of the program output can also be enabled in config.yml (exec.cache = true). Caching will only be enabled for the built-in default examples however.

Docker image

A Docker image is automatically built by Travis CI and pushed to the repository at https://hub.docker.com/r/stonemaster/dlang-tour/. The Dlang-Tour Docker container is configured to use Docker itself to compile and run D source code in the online editor. For that to work the host system's Docker /var/run/docker.sock Unix domain socket has to be mounted to work inside the container. Any sandbox for compiling will then be started on the host system actually, and not within the Dlang Tour Docker container itself.

The code for running RDMD in a Docker sandbox can be found here: dlang-tour-rdmd.

So to run the latest dlang-tour version in a Docker container, run the following command:

docker run -d --rm -p 80:8080 -e GOOGLE_ANALYTICS_ID=123 -v /var/run/docker.sock:/var/run/docker.sock \
	--name dlang-tour stonemaster/dlang-tour

The tour will be available at your host system on port 80. To stop the container again just run:

docker stop dlang-tour

For further Docker foo please refer to the Docker documentation.

Note: The docker version inside the container must be compatible to that on the host system! Generally a newer version might run on the host system than within the container.

The Docker container is also the suggested way of dlang-tour's deployment.

Environment variables

`GOOGLE_ANALYTICS_ID`

default: (empty)

To enable Google Analytics please specify the environment variable GOOGLE_ANALYTICS_ID and set it to the tracking ID provided for your Analytics account.

`EXEC_DOCKER_MEMORY_LIMIT`

default: 256

Memory limit per started rdmd Docker container which compiles and runs the user code.

`EXEC_DOCKER_MAX_QUEUE_SIZE`

default: 10

Maximum number of parallel executions of the rdmd user code Docker container.

`EXEC_DOCKER_TIME_LIMIT`

default: 20

Time limit in seconds the rdmd Docker container with the user code is allowed to take until it is killed.

`EXEC_DOCKER_MAX_OUTPUT_SIZE`

default: 4096

Maximum allowed size in bytes of the generated user program's output.

Deployment

A deployment guide can be found in the deploy folder.

License

Boost License. See LICENSE.txt for more details.

sergiors/dlang-tour