David Heinemeier Hansson

With a default vscode setup for this, I think we're better off leaving integration with other editors as a documentation issue at this point.

Is this an extraction or a speculation? Is anyone here using something else than bcrypt? If so, can we perhaps just way the pros and cons of that before adding...

In that case, I think the bigger value would be if we added support for argon2id as an alternative, and then had a top-level switch to set the default algorithm,...

@rafaelfranca Ive actually come to regret backing the enums by an integer rather than string. It makes the db no longer be self explaining. For most contexts, I don’t think...

Actually, what I'd really like to happen here instead is that we detect the underlying type of the db column. So if the enum column is a string, we use...

We use the dynamic methods all the time. Same with the dynamic scopes. That's the key feature for me with enums, more so than integer mapping. It also provides validation...

Bonus points if we can even map this correct to native DB enum types: https://dev.mysql.com/doc/refman/8.0/en/enum.html But I would take just integer and string mapping too.

They should indeed work the same. The short-form is the correct behavior. Given that it's throwing an error when you do long-form with a block, I would suspect that there's...

Did you find a way to address the Dockerfile as well?

Also, do we need the version split if Ruby 3.1+ is now required?