Re: Oversigning Resent-xxx headers
Andrew C Aitchison via Exim-users <[email protected]>
| Newsgroups | gmane.mail.exim.user |
|---|---|
| Message-ID | <[email protected]> |
On Thu, 31 Jul 2025, Mark Hills via Exim-users wrote: > On Wed, 23 Jul 2025, Slavko via Exim-users wrote: > >> Ahoj, >> >> Dňa 23 Jul 2025 09:57:00 +0200 John Levine via Exim-users >> <[email protected]> napísal: >> >>> I was tracing down a strange bug in which mail sent to a role account >>> in an IETF working group was forwarded to the recipient's Gmail >>> account and appeared with a big ugly security warning saying invalid >>> DKIM signature. I found that the sender's mail system adds a DKIM >>> signature that oversigns the Resent-xxx headers (i.e., it asserts >>> that they don't exist.) When the IETF forwards the mail, it >>> correctly adds Resent-xxx headers, which breaks the signature and >>> causes the warning. >> >>> ... >> >>> Does Exim do that by default? If so, please don't. >> >> AFAIK yes ;-) While i do not meet problems with Resent-* headers, the >> same situation is with List-* headers, which i meet already (not caused >> by my server, but noticed in my ML experiments). > [...] > > I agree that Exim's default is actively harmful; That does need fixing. > users should not be expected to change this. RFC 6376 section 5.4 suggests that there are several strategies for choosing which headers to sign. I think it should not be for the exim developers to do more than pick a default which is safe and useful to a reasonable class of users. Users with different needs, and distributions that write their own configs that better suit their users, *should* think about which headers to sign and over-sign. > There is a previous thread, and it shows the headers I have been using > with much greater success: > > https://lists.exim.org/lurker/message/20231103.101601.7232f2f9.en.html Jeremy appears to have little enthusiasm for chosing DKIM headers; if this list could agree a set of defaults which is at least safe, I imagine that would be very helpful. > Also, on a related note, RFC 8058 makes it mandatory to sign the > List-Unsubscribe-Post header, which is in none of these. We had to change > that on a mailing list host. I have created https://bugs.exim.org/show_bug.cgi?id=3153 -- Andrew C. Aitchison Kendal, UK [email protected] -- ## subscription configuration (requires account): ## https://lists.exim.org/mailman3/postorius/lists/exim-users.lists.exim.org/ ## unsubscribe (doesn't require an account): ## [email protected] ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://wiki.exim.org/