Jens Ljungblad

Anyone who knows they'll have to require dependencies will open that file in order to require the dependency. They will then see that piece of code and a comment explaining...

Yeah I think `Bundler.require` would explode with that example as well. Well, `Bundler.require` is something that Rails adds, it's not something that's included for every Ruby application. So I guess...

Okay, here's another idea. Let's just not add dependencies to the gemspecs at all? We could stop promoting the practice of putting godmin in `admin/admin.gemspec` in favor or placing it...

Well, I had some memory that one reason we did this was so that engine binstubs, `admin/bin/rails`, would find the godmin generators. I tested this briefly yesterday and there was...

I think this should be solved as part of https://github.com/varvet/godmin/issues/228. If we can decouple sandbox deployment from the template, that would solve this issue.

Good point, forgot that it calls the method that way. Perhaps we should put that in the readme? Is that good enough then or do we have other reasons for...

I think we removed support for associations in columns, and linking to the associated record, but I could be wrong. Would be good to have though, as well as support...

I guess one reason for manually specifying types would be performance? Our previous solution for checking what type a certain field was involved reflection, I think. It's what we do...

If we want to get away from partial overriding because of #221, this could be one way. **Case 1. You only want to tweak the attribute value slightly** We could...

If there would always be a `ResourceController`, generated by the installer, would that solve basic overriding? We tend to always create such a class anyway because we want shared behaviour...