On Mon, Sep 07, 2026 at 12:09:36AM -0400, Kurt Hackenberg wrote:
On Sun, Sep 06, 2026 at 19:15 +0800, Kevin J. McCarthy wrote:

Mutt actually generates the header with each reference on a single line. So each reference included will take up a single line. Mutt caps the actual sent header to 10...

Why? To detect certain errors, like an alleged loop in the tree?

I think 10 is much too low. Right now a thread on a mailing list I'm on has gone 30 levels deep. I think a loop-detecting cap should probably be 100.

I couldn't say why, since it's been that way since the first "initial revision" commit by Thomas 28 years ago. We'd have to ask ME or Thomas.

However, I would guess that they thought 10 references were enough to reestablish the message place in the graph, even if a parent of two were missing.

Given the length of time it's been like this, without any reports of actual harm caused by this trimming, I lean against increasing the number.

--
Kevin J. McCarthy
GPG Fingerprint: 8975 A9B3 3AA3 7910 385C  5308 ADEF 7684 8031 6BDA

Attachment: signature.asc
Description: PGP signature

Reply via email to