Hey there, I just made a PR for mongodb v2. I was wondering how I could build the image without waiting for the next release?

I noticed that each integration has a dockerfile. However, the content of the dockerfile copies the build output to the image .. 😅 I was wondering what is the motivation about this? Why not build the integration within the dockerfile? Personally I only see the disadvantages: for instance now I first have to setup java sdk before I can build this integration (I’m not a java developer), builds are not deterministic on all machines (admittedly publishing from the CI should be fine), etc.. But am interested to hear the other side of this 🙂