On 5/30/24 09:56, Alessandro Vesely wrote:

Hi all,

I've seen these Subject: lines being changed using utf-8 quoted-printable encoding for every non-alnum character.  The message has: X-Mailman-Version: 3.3.9rc4.

And what Mailman package is this? The GNU Mailman project has never had a version 3.3.9rc4. The only versions we have had in our GitLab repo in this vicinity are 3.3.8, 3.3.9b1 and 3.3.9.

Of course, doing so breaks DKIM signatures, and I'm trying to understand in what cases it is advisable to revert the encoding in order to verify the signature.  Does Mailman 3 always and unconditionally behave like that or what does it depend on?

Normally Mailman only alters and/or changes the encoding of Subject: headers when adding a subject prefix which will break DKIM in any case.

I don't know what's doing it in your case. Maybe something in whatever is calling itself 3.3.9rc4 or the Python email package associated with that or maybe some MTA in the path to you.

P.S. I'm curious to see how this Subject: is going to be encoded by Thunderbird. Original-Subject: Mailman3 replaces [ with =?utf-8?q?=5B, why?

A few headers in the message I received from the list are

> X-Mailman-Version: 3.3.10b1
> Precedence: list
> Subject: [Mailman-Developers] Mailman3 replaces [ with =?utf-8?q?=5B, why?
> List-Id: GNU Mailman developers <mailman-developers.python.org>

-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan