authentication-zero VS Sorcery

Subsequently, we need a way to authenticate our users to associate prompts with them. Rather than using an incumbent like Devise, I chose to use a different approach. The authentication-zero gem can flexibly generate an authentication system, as opposed to including it as an engine. Conveniently, it comes with options such as:

If you don’t need a good amount of features that Devise brings to the table, I‘d skip it entirely. Look up has_secure_password, that will be enough for a vast amount of applications with authentication. Maybe combine with cancancan for authorization. Once you feel you have a grip on those, re-evaluate devise or take a look at https://github.com/lazaronixon/authentication-zero which will transparently integrate into your app instead of providing a Rube Goldberg machine (that’s what devise will feel like for beginners for a long time).

I keep going back and forth between Devise and something a little more friendly like authentication-zero gem for authentication.

For those that have used the authentication-zero gem or are familiar with its functionality, what is the best way to upgrade it in a project when new functionality is released?

True. I tend to stay away from gems that try to integrate into multiple parts of your app to provide some sort of comprehensive solution. The kinds of gems I recommend are: 1) libraries (you call into them when you need them) 2) mounted apps on a url, isolated from the rest of your app 3) generators (this one seems nice, the author mentioned it in another HN thread: https://github.com/lazaronixon/authentication-zero).

Check out https://github.com/lazaronixon/authentication-zero

I prefer to use authentication-zero, which generates code for me in the same application using has_secure_password, has good security practices, uses the same functions as Rails, and allows me to modify the flow to my liking.

I have been using devise for a while and it has consistently given me issues. I have wistfully been staring at sorcery for a while now but cant justify the switch since devise is already in the project.

I made a 8 minutes video tutorial (following the wiki: https://github.com/Sorcery/sorcery/wiki/Simple-Password-Authentication) to introduce how to build a simple password authentication feature with Sorcery. With some minor modification to please Turbo.

We'll be installing Sorcery based off this tutorial in their wiki. I'm modifying a little bit since we are creating something different, but also because their tutorial is a bit outdated since it is based off an older version of Rails.

But does it have to be so soon? There are other areas where I'm just as ignorant as I was about the inner workings of authentication (see "the database" below), and in these areas there's not a gem that can automatically solve the problem for me—which is what I've ended up doing for authentication in my own project: even though I could build authentication from scratch, instead I'm using an authentication gem because the effect is exactly the same, but with less code in my app for me to maintain. (Rather than Devise, I've chosen the more lightweight alternative Sorcery. It's simple enough that I can still understand and control the authentication flow, while also providing enough conveniences that I don't have to write out implementation details from scratch.)

Also Sorcery is, despite its name, a little less magic than Devise.

sorcery