Best Practices for Git Tracking R Packages with .jar dependencies

[raumer]

I am currently developing and maintaining an R package that contains several .jar files as dependencies. I am struggling to find an efficient git workflow for the project.

We have 3 environments in our development cycle: dev, stage, and prod. I'd like to organize these environments as git branches, wherein new features are first developed on the dev branch and promoted "upstream" via pull requests to stage and then ultimately prod . Changes to the prod branch will be continuously integrated with RStudio Package Manager, and distributed from there.

The problem that I am facing is in dealing with the .jar file dependencies. The .jar files are also in development alongside of the R code. My understanding is that git doesn't handle diffs or merges of binary files very well. Yet, I need a good way to keep the dev, stage, and prod versions of the .jar files in sync with the corresponding branches of the R package. I am wondering, had anyone developed or know of a best practice for a situation like this?

[maelle]

A rather general answer in case you don't get a more specific one.

Via the search https://github.com/search?l=&o=desc&q=user%3Acran+extension%3Ajar&s=indexed&type=Code you can find CRAN packages that contain jar files, and some of them have a GitHub repository linked from their mirror repository e.g. https://github.com/OHDSI/SqlRender from https://github.com/cran/SqlRender (where the jar file is in the same repo).