I would say yes and no, but yes the clone command can do it. But branching and CI get a bit more complicated. Pushing and reviewing changes gets more complicated to get the overview. If the functionality and especially the release cycle is different the submodules still have great values. As always your product and repo structure is a mix of different considerations and always a compromise. I think the additions in git the last years have made the previous really bad pain points with bigger repos less annoying. So that I now see more situations it works well.

I always recommend keeping all testing in the same repo as the code that affects the tests. It keeps tracking changes in functionality easier, needing to coordinate commits, merging, and branches in more than one repo is a bigger cognitive load.

Yet, that is the real Linux phone killer move by the former, Microsoft CEO for Nokia. Also the move that killed Nokia phones.

It’s also easier to work if one simple git command can get everything you need. There is a good case for a bigger nono-repo. It should be easy to debug tests on all levels else it’s hard to fix issues that the bigger tests find. Many new changes in git make the downsides of a bigger repo less hurtful and the gains now start to outweigh the losses of a bigger repo.

I agree that in most cases it’s more of an E2E or integratiuon test, not sure of the need to split into different repo, and well in the end I’m not sure that would have made any big protection anyhow.

The Windows Lumina phone was a battery-draining buggy version with a smaller screen than the Linux version N9, When they turned that into Windows (Lumina 800), they had to use a smaller screen and less memory as Windows couldn’t handle the hardware as MeeGo could.

I’m not sure if Yggdrasil or Slackware, which we tried out at the old university computers. But quickly Debian became so much more flexible.