Hacker News Reader

Comment by caseysoftware | Hacker News Reader

In theory, providers/servers should ignore everything from the + to the @ so that should work just about everywhere else. And since most forms don't filter that out, you can make numerous "unique" addresses connected to a single one.

Now odds are, someone is going to respond with the one service where the + becomes significant..

View on HN

That isn't the theory. There are almost no rules for the local part of an email address and you must not attempt to "validate" them. Whether a '+' terminates the user name or means nothing is entirely up to the domain operator.

The exception being when you know that the domain operator supports RFC5233 then you're allowed to be aware of subaddresses.

That does this RFC even say? I read it twice and it seems to say "Some mail servers may use the part of an email after any character as a subaddress which doesn't affect which inbox a message is delivered to."

It seems to only talk about the receiving server and gives no way for the sending server to do anything useful as it doesn't specify the character to use if the receiver implements it at all.

I think this is good, I would like to not have senders try to guess anything about the structure of my email, but it makes me wonder what the point of having the RFC at all is.

> providers/servers should ignore everything from the + to the @

Why? I'm not sure it has ever been a standard and it's certainly not a default implementation for most providers?

A quick google search returns https://www.cs.rutgers.edu/~watrous/plus-signs-in-email-addr... . Although not standard, its use is "legal"

There is a standard, RFC 5233 on subaddressing: https://tools.ietf.org/html/rfc5233

Adoption is a mixed bag.

Been using this for a while -

https://github.com/johno/normalize-email

That is a shamefully bad library. There's a million ways this is broken but the worst by far is that it lowercases the local part. Case MUST be preserved in the local part.

It is fine if you want to try to block duplicate emails, but you should always continue to use the original. But yes, it has false positives and false negatives. I don't think I would call it "broken".

But at the end of the day doing this kind of comparison probably raises more questions than it solves.

It's not even fine for that. If you assume that JayNesmith@ and JayneSmith@ are duplicates, you're eventually going to annoy someone.

Does MS Exchange or any other relatively big email server allow multiple emails with mixed case? As far as i'm aware all of the big consumer providers (Google, Yahoo, Outlook) don't allow it.

Reply on news.ycombinator.com