I understand the value of free software, but just to put the repo on self hosted servers, I really don’t think configuring travis (not sure if it supports) or write a entirely different config for running on your self hosted server is worth anyone’s time.
When you talk about a self hosted git server, you have to see the following things really carefully.
First, we’ll have to configure somewhere, and it’s not like it’s not costing anything at all, it will. At least a few memory. Now that should be done by someone at FSMK, it’s not like you do a sudo apt-get install gitlab and it’s done.
Second, if we’re doing some kind of project which someone else should be able to use, we need to have some sort of testing integration (travis, circle ci, jenkins for example). Since we’re hosting code on our server, we’ll not want to use the their server, of course you’ll want to use it on our server.
Now to configure that is also not a child’s play, we can do it, but someone in FSMK would have to do it.
Every PR created should be tested on a container, which also consumes some memory, spinning up a VM and just using them is also not a child’s play.
Now let’s say we configure a container to do a continuous integration for golang, what if someone wants to do a project on RoR? What about Android? What about some completely different language which we don’t have images for yet?
If situations like that arise, someone would have to configure that again. Which I again say, is not worth it.
When let’s say we’re using 10 different languages, we’ll need 10 different images which will spin up every time we do a push or create a pull request, believe me, that’s a lot of memory.
It’s not easy to host our own solution for every single thing.