Why git? With objects (IBM i)

[worksofliam]

When using git with ILE source, is it important that each developer still gets their own development library which their own objects (not source code) can go into.

A problem that may be an issue in IBM i shops is that they have hundreds of programs, service programs, tables, indexes, etc, that it may become a serious issue if each developer is re-building them in their own developer library.

If you have a small set of programs (maybe less than 20, depends on your system specs) then it would be okay for each developer to re-build all the objects into their own library. If your shop had more than that, it could be an issue for system performance or system storage.

Luckily the IBM i has a one-up here: the library list. You may have a main branch (typically the master branch) which could be built into a ‘master’ library, which contains all the latest objects compiled from the repo. This means, when a user clones the repo: instead of building each object, they only build the objects they need to change or work on and then depend on the programs out of the ‘master’ library.

In fact, you would be used to this concept if you were already developing in source-members out of your own library and then merging up to the master development library.