Smarter handling of fetching subprojects’ data

[strogonoff]

Site build may fail in strange ways if project data fetched by Jekyll from other repositories is found in an unexpected state.

It’s a bit of an edge case, as it’s never an issue in clean CI environment or during routine development. However, it adds unwelcome confusion during initial bootstrapping and troubleshooting of other errors (such as #5).

The code in question is around https://github.com/riboseinc/jekyll-theme-open-project-helpers/blob/master/lib/jekyll-theme-open-project-helpers/project_data_reader.rb#L136, where it expects either “a clean slate” or everything already set up properly, and may get confused by a third possibility.

Goal

The idea is a trivial addition to/around git_sparse_checkout which would check for existing data consistency & perform a cleanup and re-initialization if needed.