My project has many derived files, files generated from others, thus really duplicates. Examples:
Transpiling: I use either coffeescript or es6/babel to convert source to .js files that are redundant, in that they are easily recreated from the CS/ES6 files.
Bundling: The derived .js files are preprocessed (uglify etc) then put into myproject.js & myproject.min.js files for easy download from github.
Documentation: I use Docco to create HTML documentation of each source file, thus redundant .. the source comments create the documentation.
I'd prefer the repo be tiny and not have the derived files.
One thought is using gh-pages to have the complete set of files, using a .gitignore to skip derived files for master, but having a separate gh-pages .gitignore allowing all derived files. Seems a bit extreme though.
Is there good (standard, best practices) way to manage the derived files? They can confuse git -- making pull/merge/fetch report conflicts in derived files which we don't care about, a clean full build always is correct.