On Tue, 1 Sep, 2020 at 09:59, Timo Sirainen timo@sirainen.com wrote:
On 1. Sep 2020, at 6.24, TACHIBANA Masashi tachibana@qualitia.co.jp wrote:
Hi,
Is this expected or not?
From: user1@fuga.example.com user1@example.com To: user2@hoge.example.com user2@example.com ↓ a uid fetch 43055 (envelope)
- 1860 FETCH (UID 43055 ENVELOPE ("Thu, 30 Jul 2020 13:52:59 +0900" "test1" ((NIL NIL "user1" "fuga.example.com")) ((NIL NIL "user1" "fuga.example.com")) ((NIL NIL "user1" "fuga.example.com")) ((NIL NIL "user2" "hoge.example.com")) NIL NIL NIL "WLXvFmpAZFNeQbPPITjRwimDyamnBm1@test"))
This is an invalid email address, so it's neither correct nor incorrect to have this output. But this reminded me that I was going to discuss about this with other IETF people. Lets see what others think: https://mailarchive.ietf.org/arch/msg/extra/sqRTdsV_DGBhHu2ghdCDFo_pM8Q/
While it is an invalid email address, in the exact same vein as https://dovecot.org/pipermail/dovecot/2020-August/119658.html Dovecot's approach is unhelpful here, and means MUAs must download complete headers rather than reply on envelope address structures. In fact, unlike in the linked case this example this is actually a security vulnerability: http://mailspolit.com/
As a MUA maintainer, I'd really like to see Dovecot take a more proactive approach to sending useful values in envelope address structure, so we don't have to download headers all the time.
From: "user1@fuga.example.com" user1@example.com To: "user2@hoge.example.com" user2@example.com ↓ a uid fetch 43056 (envelope)
- 1861 FETCH (UID 43056 ENVELOPE ("Thu, 30 Jul 2020 13:53:59 +0900" "test1" (("user1@fuga.example.com" NIL "user1" "example.com")) (("user1@fuga.example.com" NIL "user1" "example.com")) (("user1@fuga.example.com" NIL "user1" "example.com")) (("user2@hoge.example.com" NIL "user2" "example.com")) NIL NIL NIL "WLXvFmpAZFNeQbPPITjRwimDyamnBm2@test"))
This is a valid email address, and ENVELOPE reply is correct.
Agreed.
//Mike
-- Michael Gratton. https://mjog.vee.net