2023 Rebuild?

[nelsonic]

As noted a couple of years back: dwyl/contact#8 swoosh already had all the features we needed for basic email sending. What it lacked was deliverability tracking. Which was why I didn't use it at the time.
But now that swoosh is included by default in Phoenix 1.6.x it's the "defacto".

That is simultaneously an annoying thing - because there's lots of @dependabot noise https://github.com/search?q=org%3Adwyl+swoosh&type=issues ... - and a really good thing because the decision is made. ✅

So the question is: should email be separate from auth?
Or are we fragmenting our backend too much by making them separate apps? 🤷‍♂️

My original reasoning for making the auth and email apps separate
was so that anyone working on auth didn't have to think about email.
But ... given that swoosh is included in Phoenix by default
(i.e. you have to manually exclude it with --no-email flag)
what is the advantage to keeping them separate? 💭

[iteles]

The only advantage I can see is to use email for other purposes other than auth.
But for us, I don't think it makes sense - we're adding unnecessary complexity (from a layman's perspective) at this stage.

[nelsonic]

@iteles thanks for your feedback. 🙏
My only reluctance with this is: more setup time for auth. ⏳
But I think I've got a solution. 💡 dwyl/auth#271