I have two bare repos super and a, one clone of super – clone_super and one clone of a – clone_a. I’ve added repo a as a submodule in clone_super.
If I now make some changes in clone_a, push those and do a git pull origin master in ~/clone_super/a, it will render in a modify status for ~/clone_super/a, displayed after git status in ~/clone_super.
If I do git submodule update, ~/clone_super/a will revert to previous commit. So, is it intended that I need to commit and push submodules for repos that they exist in each time I update them?
The short answer is yes, you must commit and push a submodule after you update it.
Think of how the directory would work if it wasn’t a submodule, every time you changed a file you would have to commit and push it. A submodule is exactly the same except super does not track the file changes directly. Instead the repository inside the submodule tracks the file changes and the super repository tracks those changes by tracking commits in the submodule.
The nice thing about this is it allows tracking at different granularity. You can make dozens of tracked changes in the submodule without changing the super repository and then make a single commit to super reflecting the changes in the submodule. The cost is the requirement of committing twice.