Breaking a really long thread

[Eric]

On Mon, 04 Apr 2016 14:00:26 +0100, Mark Walters <markwalters1009 at gmail.com> wrote:
> On Mon, 04 Apr 2016, Eric <eric at deptj.eu> wrote:
>> On Sat, 02 Apr 2016 06:56:12 -0700, David Mazieres <dm-list-email-notmuch at scs.stanford.edu> wrote:
>>> David Bremner <david at tethera.net> writes:
>>> 
>>>> David Mazieres <dm-list-email-notmuch at scs.stanford.edu> writes:
>>>>
>>>>> Is there any way to break an existing thread (so as to start over with a
>>>>> smaller thread), or otherwise to tweak the threading rules so that a
>>>>> particular References header gets ignored.
>>>>
>>>> Currently there is no way to do this, as threads are "stateless"
>>>> i.e. created on the fly by _notmuch_create_thread based only on
>>>> immutable mail data.
>>> 
>>> Thanks.
>>> 
>>>>> It's annoyingly slow to open
>>>>> a thread with 10,000 messages just to read one SMS.  I'm almost tempted
>>>>> to mangle the messages on delivery and remove the References header
>>>>> before notmuch sees them, but it would be nice to have a cleaner
>>>>> solution, as there are other situations in which one might want to
>>>>> "reset" a really long thread.
>>>>
>>>> Like this thread ;).
>>> 
>>> Oops, sorry for the irrelevant thread inclusion.  I guess emacs adds the
>>> References header after a message is sent is sent?  In my setup, the
>>> easiest way to post to a mailing list is to reply to an existing message
>>> (since I subscribe to each list under a different email address).  I
>>> tried to start a new thread by deleting the In-Reply-To and header which
>>> was all I saw, but I guess the References header got inserted later...
>>
>> Neither notmuch emacs nor any other email client has any business
>> inserting a References header after the user "presses Send". On a new
>> message it shouldn't exist unless inserted manually, and on a reply it
>> should come from the replied-to message (and be changed) before you start
>> "replying". More likely that (if you didn't miss it) it was not shown
>> to you although it existed - that would count as a bug in my opinion
>> (I don't use emacs for anything, not even notmuch).
> 
> By default the reference header is hidden. It is controlled by
> message-hidden-headers which you can customize. (Note notmuch adds
> user-agent to this list via notmuch-mua-hidden-header.)

Ah. Well of course I didn't know that since I don't use emacs. I guess
that if the OP is going to use reply to start new threads he should
unhide it.

>> Actually the message you replied to has no References header, but notmuch
>> reply (command line) to it generates both References and In-Reply-To
>> (same content). I assume notmuch emacs does the same. I don't believe
>> that References should be generated in this situation, its only use
>> would be by a threading algorithm that doesn't use In-Reply-To, and I
>> would consider that a bug in said algorithm.
> 
> That isn't my reading of RFC2822 (section 3.6.4):
> 
>    The "References:" field will contain the contents of the parent's
>    "References:" field (if any) followed by the contents of the parent's
>    "Message-ID:" field (if any).

OK, I guess it should be there. In which case it shouldn't default to
hidden in any MUA, far too many people use the reply-for-new approach
without understanding it.

>> Actually I think there should be a "reply as new" option which uses
>> the other message but does not add either In-Reply-To or References
>> (and does not carry the latter forward if it exists).
> 
> That would be possible. If you don't actually want to include the
> message itself then "c f" to stash the from, and then pasting that as
> the "to" address gets pretty close.

Eric
-- 
ms fnd in a lbry