I’m starting a new thread about this since the old ones on the topic aren’t really relevant. This is in related to this issue: Documentation for cookiecutter · Issue #196 · bemanproject/exemplar · GitHub
For a variety of reasons I’ve been playing with cookie cutter without the usage instructions – and I think some others have as well. One use of cookie cutter I’ve experimented with is to basically be able to generate the latest from an exemplar and then diff against an earlier repo to discover the CI and other infra type changes to pick up. It’s aleady quite useful for that.
Of course the primary use case is for new repos – here we might still have some rough edges, not sure. As it stands today the (unwritten) guidance to authors on new repos goes like:
- make a repo with exemplar as the ‘template’ via github
- exhaustively run through various files updating for library name…
It’s the second part that is the core that cookiecutter should replace. So the user process looks like: - install cookiecutter from instructions here: https://www.cookiecutter.io/
- using your cloned from exemplar template repo
- run cookiecutter and answer questions
- make the output directory the repo tree?
- make the output directory a different tree and merge/copy over?
I’m a bit worried on the last steps and what cookiecutter is assuming about CM management.
w.r.t. the state of cookecutter here’s enhancements I see:
- all references to ‘identity’ text should be replaced appropriately
- includes test / examples directory and readme.md
- for header only it would be good to remove src directories/builds (can CC do that?)
- the cmake.in file should be renamed for repo
I can write an issue/issues for the last part if other agree.