launchpad-dev team mailing list archive

[Barry Warsaw <barry.warsaw@xxxxxxxxxxxxx>]

I didn't quite have the context, so I went back and skimmed a few messages
from the archives.

On Jun 10, 2010, at 09:35 AM, Stuart Bishop wrote:

>On Wed, Jun 9, 2010 at 10:25 PM, Graham Binns <graham@xxxxxxxxxxxxx> wrote:
>
>> I'm currently getting stuck into the better bugnotifications story[1]
>> and I think we need to settle how notifications from Launchpad should
>> be addressed. I'm working on three bugs, specifically:
>
>So I think I recall why it was done the way it is done. We pretty much
>copied Roundup, which at the time had the best reputation for email
>interaction.
>
>Roundup wanted to intercept the conversation between people as
>transparently as possible. So replies go through roundup and are
>relayed to all subscribers. Emails that came from Bugtracker
><xxx@xxxxxxxxxxxxx> obfuscate the sender so you don't know who it came
>from. Emails that came from Real Name <xxx@xxxxxxxxxxxxx> screw up
>address books - you would end up with loads of bogus entries for
>people and end up sending private email to the bug tracker. So Richard
>settled on using the genuine name and email address, but subverting
>the replies to go via the tracker.
>
>(adding Barry to the CC: list since he probably knows the best answer
>from 84 years of empirical studies with Mailman)

"Email sucks" is my snarky answer :).

My other snarkier answer is that your email *will* get harvested and you can't
avoid it.  Resistance is futile.

Still, we can probably do better...

Looking at bug 111147, I recall our thinking for the "Contact this user"
feature.  In that case, the user is directly initiating the email almost as if
they were doing so with a webmail application.  In that case, it made sense to
use the author's email address in the From header, give them an opportunity to
select from among their validated addresses, and warn them that this will show
up in the message that the recipient receives.  I still think this is the
right answer for this use case.

For messages generated as a byproduct of some action taken on Launchpad,
that's a different matter.  In those cases, the user is *not* taking a direct
action to send the email and has no control over the From header.  I would
argue that in those cases Launchpad is the author of the message and the From
field should be from launchpad.net.  It makes the most sense to use the bug
number in the localpart, e.g. 111147@xxxxxxxxxxxxxxxxxx.  But in that case,
I'd argue against putting the commenter's name in the From field exactly
because it will mess up address books.  I think having the commenter's name in
the body of the message is enough to know what's going on.

I'm not sure that contacting the commenter off-line is an important enough
use-case to support via the email headers.  You can always click through to
the actual bug comment, and then to the user, and then to the Contact this
User link.  Yes it's inconvenient, but how often do you need to contact the
user of a bug comment off line?

In bug 31586, I agree with the OP and Bjorn (and ScottK's later opinion), and
disagree with Dafydd that Malone is acting like a mailing list manager.  In
the latter case, if you squint, the MLM is just a repeater of the original
message, but the original author has a lot more control over the format of the
headers and body of the message.  Yes, the MLM does some munging (which is
what makes DKIM so problematic), but to me, it's a different precipitating
event for a mailing list message versions a bug comment.

In summary, I would favor bug comments coming From: 123456@xxxxxxxxxxxxxxxxxx
without the commenter's name.

-Barry

Attachment: signature.asc
Description: PGP signature