Topics

moderated Change DMARC behavior for p=none #misc

 

Hi All,

Right now, we munge the From line of messages for people who have DMARC records of p=reject or p=quarantine. We do not do that for people who have DMARC records of p=none. I propose changing that so that we also munge From lines for people with DMARC records of p=none. Recently, several people have contacted support about this, and I think this change makes things more straightforward.

Please let me know if you have any objections or questions.

Thanks,
Mark

Dave Sergeant
 

My objection would be that it makes it more difficult for people to
reply directly to sender from email. OK, I know there is a 'reply to
sender' at the bottom of the mail but many don't see this. I constantly
get queries as to why a person's email bounces but it is clearly there
in the 'from' line, although munged and they have to edit it to the not
quite obvious correct one. I know how to do this, but many can't work
it out.

At the moment it is good that groups.io doesn't do it for everybody
like Yahoo did (with its slightly easier to d-munge version). I can see
why Mark has to do it but it is unfortunate it has to be extended.

Dave

On 19 Feb 2020 at 17:19, Mark Fletcher wrote:

Hi All,

Right now, we munge the From line of messages for people who have DMARC
records of p=reject or p=quarantine. We do not do that for people who
have DMARC records of p=none. I propose changing that so that we also
munge From lines for people with DMARC records of p=none. Recently,
several people have contacted support about this, and I think this
change makes things more straightforward.

Please let me know if you have any objections or questions.

Thanks,
Mark

http://davesergeant.com

Dave Sergeant
 

Another thought. Currently plain text versions of group mails do not
contain the 'reply to sender' option in the footer, only in the html
version if it exists. If this is to be implemented can we add 'reply to
sender' to the footers in these messages as well please.

Dave

On 20 Feb 2020 at 6:49, Dave Sergeant wrote:

My objection would be that it makes it more difficult for people to
reply directly to sender from email. OK, I know there is a 'reply to
sender' at the bottom of the mail but many don't see this. I constantly
get queries as to why a person's email bounces but it is clearly there
in the 'from' line, although munged and they have to edit it to the not
quite obvious correct one. I know how to do this, but many can't work it
out.

At the moment it is good that groups.io doesn't do it for everybody like
Yahoo did (with its slightly easier to d-munge version). I can see why
Mark has to do it but it is unfortunate it has to be extended.

http://davesergeant.com

 

Dave,

Currently plain text versions of group mails do not contain the 'reply
to sender' option in the footer, only in the html version if it
exists.
The reason has to do with the fact that those are mailto links, and they necessarily include the text of the subject line. They'd be sizeable and quite ugly as URLs in plain text.

If this is to be implemented can we add 'reply to sender' to the
footers in these messages as well please.
Maybe call it "New message to sender" instead. That would be a mailto: link without the Subject parameter. Which would be more honest in a way, as mailto links can't create actual replies (can't invoke the mail interface's Reply function) in any case.

One limitation is that email interfaces vary in whether they will linkify (make clickable) a mailto: URL found in plain text; some do, some don't (whereas most will linkify an http(s): URL detected in plain text). So some of your members may still have to resort to copy/paste; but at least they would have a non-munged address to copy.

It might be better to just include the plain text of the sender's address in the footer, for easy copy/paste.

And then perhaps decouple this suggestion from the DMARC issue. That is, make it a stand-alone suggestion in its own topic.

Shal

 

On Thu, Feb 20, 2020 at 03:22 AM, Mark Fletcher wrote:

Right now, we munge the From line of messages for people who have DMARC
records of p=reject or p=quarantine. We do not do that for people who have
DMARC records of p=none. I propose changing that so that we also munge From
lines for people with DMARC records of p=none. Recently, several people have
contacted support about this, and I think this change makes things more
straightforward.

Please let me know if you have any objections or questions.
I'm objecting to this. This would munge From for everybody. It'd be disruptive, and for what? What's the sense? p=none means no rejection. People (including me) make DMARC records with p=none in the belief that some large receivers consider messages passing DMARC as less likely spam, but we don't want the DMARC disruption, so p=none is deliberate and permanent.

Please give examples of problems because of p=none.

--
Lena

ro-esp
 

On Thu, Feb 20, 2020 at 02:22 AM, Mark Fletcher wrote:

Right now, we munge the From line of messages for people who have DMARC
records of p=reject or p=quarantine. We do not do that for people who have
DMARC records of p=none. I propose changing that so that we also munge From
lines for people with DMARC records of p=none.
Please let me know if you have any objections or questions.
It's hard to have objections.... as I don't know what is talked about.
Does anyone care to explain/translate? "munge", "p=" ?

groetjes, Ronaldo

Michael Halstead
 

This would frustrate our users who receive software patches via the project's mailing list.

Git takes patch author information from e-mail headers. Right now we have git hooks in place on the server to prevent munged author information from reaching the repositories. Our maintainers have to fix author information before pushing code. Right now our maintainers only have to do that for a few senders. This change would mean fixing that information for every patch. 

It is not easy to automate fixing author information given the many unique workflows used by our maintainers. Having correct header information, when possible, is very desirable.
-- 
Michael Halstead

Glenn Glazer
 

On 2/20/2020 13:47, ro-esp wrote:
On Thu, Feb 20, 2020 at 02:22 AM, Mark Fletcher wrote:

Right now, we munge the From line of messages for people who have DMARC
records of p=reject or p=quarantine. We do not do that for people who have
DMARC records of p=none. I propose changing that so that we also munge From
lines for people with DMARC records of p=none. 
Please let me know if you have any objections or questions.
It's hard to have objections.... as I don't know what is talked about.
Does anyone care to explain/translate? 

"munge", 

https://www.dictionary.com/browse/munge

"p=" ?

https://en.wikipedia.org/wiki/DMARC#DNS_record

Best,

Glenn


--
PG&E Delenda Est

 

On Thu, Feb 20, 2020 at 12:31 PM Lena via Groups.Io <Lena=lena.kiev.ua@groups.io> wrote:

I'm objecting to this. This would munge From for everybody. It'd be disruptive, and for what? What's the sense? p=none means no rejection. People (including me) make DMARC records with p=none in the belief that some large receivers consider messages passing DMARC as less likely spam, but we don't want the DMARC disruption, so p=none is deliberate and permanent.


In my experience, people who are interested in adding a DMARC record for their domain start with setting one up with p=none. Then they go through and audit the reports on their domain. They see that we are sending email with unmunged From lines, and contact us asking us to fix our problem. This happened twice this past week and has happened on a somewhat regular basis over the past year or two.

That said, I've rolled back the DMARC change that I had pushed earlier today; we are once again not re-writing From lines for p=none.

Thanks,
Mark