I am implementing databases in hardware, means - pure silicon - and i know, how hard that is!
I am following Dgraph now for a while and here are my 2¢:
Test driven development must vastly improve in Dgraph, Badger. I think, Dgraph could massively improve by introducing kind of collective test set. The art of TDD always is in generating test data, of course. Perhaps we all could contribute more, much more here.
Have fun!
Hi @Guido_Stepken. Are there specific improvements you’d like to suggest? We’re always open to PRs including those that improve our tests.
Given a set of functions a am always testing side effects, pairing each single function with each other function: f x g and g x f. For each pair i have a test set.
That, of course has impact on the way of programming. What is not tested, is not allowed to run. E.g. f x g x f or f x g x h. If it’s not testable, it will throw an error.
An Excel sheet would be a good start for us to know, what was tested and what we as community can safely use. That for the beginning.