Skip to main content

dog is a simple wrapper for docker run to make it simple to call tools residing inside docker containers

Project description

dog

test Status centos-for-dog package Status

dog is a simple wrapper for docker run to make it simple to call tools residing inside docker containers.

The basic idea is that you just put "dog" in front of the command you normally call.

So if your normally use "make" to compile your code, then if your compiler tools are stored inside a docker container you just do "dog make" instead - and dog will mount you local workspace and run the correct docker container without the developer noticing at all.

The docker container to use is defined in a file called dog.config which is normally positioned in the root of your workspace and contains the name of the docker image.

Key selling points

Also known as the three reasons for using dog.

1. Versioning your tools

By putting the docker your project uses to compile in a dog.config file inside the projects repo, the version of the tools to use is now versioned along with the project itself.

This means that you are able to reproduce (and bug-fix) on a 10 year old project, because the dog.config on your release tag/branch will point to the build tools as they looked 10 years ago, and by the "magic" of docker they will still work the same way they did 10 years ago.

It also makes it easier when the tools needs to be upgraded, you can work on a feature branch to get the tools and project build scripts working and then merge both tools and build scripts to master when they are ready.

Experimenting with your tools without breaking anything also becomes much easier when you do not have to modify a machine, but can just modify a Dockerfile.

2. Making developers happy

Installing tools is never fun, or at least, as a developer you would like your build to "just work" on the developers machine without a lot of work.

By putting the build tools inside a docker image you can suddenly compile your project anywhere, so the developer can decide to use the Linux distro (or even Windows or Mac) that they like, and use the editor of their choice - compilation will be the same on all systems since the build tools are contained and versioned inside the docker image.

3. Avoiding the permission issues that arises from using docker for containing tools

Dog is simply a wrapper around "docker run" but with some sane default like mounting the current project inside the docker since that is where you would normally want the build tools to look.

But docker run will run as "root" by default inside the docker, or any other "default user" that has been set up for that particular docker image.

This is problematic since the output of your build will then suddenly be owned by whichever user is running inside the docker, making life difficult for the developer.

To solve this, dog will set a few environment variables that the docker image entrypoint can then use to "sudo" to the correct user and execute the build as the same user that just called dog.

In essence it just makes it simpler to use docker image containing your build tools.

D9.12 kbnuxaes01 [/scratch]
$ id
uid=4548(rtol) gid=1000(oticon) groups=1000(oticon),44(video),1007(docker)
D9.12 kbnuxaes01 [/scratch]
$ dog touch demo.txt
D9.12 kbnuxaes01 [/scratch]
$ ls -l demo.txt
-rw-r--r-- 1 rtol oticon 0 Oct  2 21:55 demo.txt

Example

By default the docker image is fetched from the default docker registry (hub.docker.com), for instance:

[dog]
full-image=gcc:7.5

Means that:

dog gcc main.c -o main

Will compile main.c using gcc 7.5.

crossbuild-for-dog

You can see a more real-world example by looking at the crossbuild-for-dog test image that extends the standard crossbuild docker image with a dog enabled entrypoint - which can then be used to compile ARM linux executables anywhere.

The source for crossbuild-for-dog is available in tests/dockers/crossbuild-for-dog and there is a pytest verifying it in tests/test_crossbuild_for_dog.py.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

dog-8.tar.gz (6.6 kB view hashes)

Uploaded Source

Built Distribution

dog-8-py3-none-any.whl (10.7 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page