898003 - Puny-encoding username part of email addresses in HTMLInputElement::IsValidEmailAddress?

Status: RESOLVED INACTIVE

(Core :: DOM: Core & HTML, defect, P5)

Description

[Biru [:poiru]]

While working on bug 884332, I noticed that we puny-encode not only the domain part, but also the local part of e-mail addresses in HTMLInputElement::IsValidEmailAddress. I discussed this with Mounir and we decided that a bug should be opened to verify that this is indeed the correct behaviour. Relevant part of the discussion below.

18:43    poiru: I still have a question: is puny-encoding the username valid? I took a
                look at Chrome[0] and they don't seem to puny-encode the username.
18:43    poiru: [0]: http://src.chromium.org/viewvc/blink/trunk/Source/core/html/EmailInputType.cpp?revision=154184#l50
18:51   mounir: it is possible that we do it wrong indeed :-/
18:51   mounir: could you open a bug for that?

Comment 1

[Matthew Wronka]

bug 842523 and bug 618876 discussed intentionally punycoding the userpart, but there's no explanation of why; the eMail RFCs indicate the local-part should only be interpreted by the final host, and I didn't see anything by WhatWG regarding anything to the contrary.

Comment 2

[Matthew Wronka]

The removed code in PunyEncodeEmailAddress from the diff attached to bug 791069 would fix this (lines 6456-6473  at the moment, anything that touches the local "username" variable).

Comment 3

[Matthew Wronka]

This appears to have been fixed recently.

Comment 4

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven’t been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.

Comment 5

[Anne (:annevk)]

Closing this as inactive as we'll follow-up here in a new bug once a related specification issue is sorted.