0

I have a bit of an unusual situation to handle with git repositories.

There is a TI repository with some sensor drivers I'd like to use, it is structured in the following way

  • root
    • other_stuff
      • docs
      • src
        • random_stuff (10x)
    • path_to_driver
      • docs
      • src
        • dependency
        • driver

This repository is maintained on their own git server (git.ti...) and I will be maintaining my repositories in a github group (group/).

Ideally, I would maintain an exact copy of TI's repo on github (group/ti_copy) as well as a repo containing only the sensor driver. The 10x random_stuff folders and the dependency folder can interfere when building, and I think it's generally bad practice to keep them as it's quite a lot of bloat,, especially considering this will need to be deployed on a more compute-restricted platform (not commercial) for testing.

How can I best structure this to avoid too much headache, I will both be making changes to my fork, as well as pulling updates from the TI version. I'd like to try and do this the "right" way, whatever that means.

Morten Nissov
  • 392
  • 3
  • 13
  • The (no well-supported and reportedly flaky) git-subtree command, and other things like git-subrepo, purport to do this. See [Alternatives to Git Submodules?](https://stackoverflow.com/q/6500524/1256452). – torek Oct 18 '22 at 21:46

0 Answers0