Meteor-feature-requests: Update Nodemailer

[cunneen]

The core email package has a dependency on node4mailer, a fork of an old version of nodemailer.

Meteor no longer runs on node 4, and the new nodemailer has some nifty features, particularly https://ethereal.email/ .

This is great for development and testing, where the email can be viewed and tested with a variety of clients (unlike the status quo with meteor, where a message is logged to the console if the MAIL_URL environment variable isn't set).

So I propose:

1. Updating the core email package to use nodemailer instead of node4mailer, and
2. update the core email package's devModeSend method to use ethereal.email .

[StorytellerCZ]

Agree with point 1. Though for point 2 that is something that I think should be left behind a flag or some other selectable option that by default is false as turning it on by default would break backward compatibility and there are other tools for that functionality as well. So maybe an additional package that if you install it will support this functionality on dev, which would allow for other similar tools to easily plugin.
Still we can do point 1 without needing to address point 2, so I think we should proceed on that.

[cunneen]

Yeah @StorytellerCZ you raise a good point about backwards-compatibility, but I'm a bit skeptical about whether any tools exist to parse the console logs for logged email messages when a MAIL_URL isn't set. Do you know of any?

I agree they are two points and can be considered separately. Thanks for your vote of support on point 1.

For point 2, I favour having ethereal.email as the default with a flag that can be explicitly set to revert to legacy behaviour. It's supported by nodemailer and provides a much better experience out-of-the-box for a new developer.

[StorytellerCZ]

@cunneen There is @msavin Meteor Toys that can catch e-mails.

[StorytellerCZ]

Point 1. of OP has been done. Looking into point 2...

[StorytellerCZ]

Since in dev mode the email is shown in console I think that is fine and satisfies the idea behind point 2. Closing.