Lance Ivy

> Perhaps this error relates to the version of Rails? Sure, what versions of Rails and Mail do you have? Would you also try: > mail.method(:smtpapi).source_location Which should look something...

this works for me: ``` require 'bundler/inline' gemfile true do source 'https://rubygems.org' gem 'rails', '4.2.6' gem 'mail', '2.6.4' gem 'mail-x_smtpapi', path: './' end require 'action_mailer' ActionMailer::Base.delivery_method = :test class TestMailer...

@cornerman That does sound like the right solution for two different AuthN servers. 👍 Meanwhile, I'm curious if you're willing to describe when having two different AuthN servers is useful?...

> I prefer having a fully initialized authn instance instead of a singleton where I need to rely on someone having called `setHost` That makes sense. The reason why I...

@fleyai I'm unsure what AuthN would be able to accomplish during SSR, since it would not have access to the session refresh token or localstorage. Any ideas?

I think it's a legacy issue in PhantomJS. The console logger isn't as capable. Speaking of PhantomJS issues -- the expected error is different, too: > HTTP response 'Origin is...

Maybe with a `subject_from!` variant? That would remain backwards compatible and match Ruby conventions nicely. I'm curious what your use case is. What decisions would you make with error data?

Sounds like it was useful for forensics, but not necessarily for application logic. Is that right? Can you say more about why this was easier to diagnose from the frontend...

let's standardize on `int` to match the API. much as it pains me, i think the golang way is to create new functions that return the correct type until cutting...

A new client with support for `context.Context` and improved types does seem like the best upgrade path. I'm not afraid of releasing a v2.0 but it'd be nice to have...