Discussion: Branching Strategy for Function Refactoring

[jibarozzo]

Posting as an issue, but it's more of a discussion of strategy.

I thought this a little more about how to best incorporate improvements on functions and make sure all is working before merging into main. I suggest the following branching strategy for refactoring functions (thanks to AI for organizing my early morning thoughts):

1. Main Refactor Branch (function_refactor):

   • All refactoring work will be done in the function_refactor branch.
   • This branch will act as the main branch for function refactoring tasks.
   • When creating a new branch for refactoring a specific function, always branch from function_refactor.

2. Specific Function Refactor Branches:

   • Each function refactor should have its own branch, which will stem from function_refactor.
   • The naming convention for these branches will be functionname_refactor. For example, if refactoring a function called sncm.LL, the branch should be called sncm.LL_refactor.

   Example:

   • function_refactor → base branch for refactoring.
   • sncm.LL_refactor → branch specifically for refactoring the sncm.LL function.
   • bino.LL_refactor → branch for refactoring the bino.LL function.

3. Workflow:

   • Step 1: Create the base branch function_refactor (already created).
   • Step 2: For each function that needs refactoring, create a new branch named functionname_refactor.
   • Step 3: After completing the refactor for a specific function, open a pull request from the specific *_refactor branch to function_refactor.
   • Step 4: Once all refactoring work for individual functions is merged into function_refactor, it can be merged into the main branch by "Squash" method.

See A successful Git branching model for more details on this model.

This method helps keep each function's refactoring isolated, making it easier to review, test, and merge individual changes.

Related to #7 and #4

[jibarozzo]

@B-Kristy

[B-Kristy]

Ok thanks for this -- it's an intuitive workflow and it makes sense to me. I probably should have been more specific when naming my function refactor branch as 'myfeature_function_refactor'.

[jibarozzo]

No problem. I'm also learning a lot about branch management. Trying to minimize convolution for me and all.

I just pushed new changes to function_refactor trying out git merge --no-ff . No PR for those changes, so seeing what works best. I like PRs better to make everyone aware of what happened.