Re: Display duplicate messages in emacs front-end

Subject: Re: Display duplicate messages in emacs front-end

Date: Wed, 09 Nov 2022 12:58:18 -0800

To: Daniel Kahn Gillmor,


From: Jameson Graef Rollins

On Wed, Nov 09 2022, Daniel Kahn Gillmor <> wrote:
> On Wed 2022-11-09 09:53:26 -0500, Daniel Kahn Gillmor wrote:
>> The functionality here, switching between duplicates in the emacs
>> frontend, is great.  I needed to use it today, because i had two copies
>> of an e-mail message, one with a valid signature and one with a damaged
>> signature.  It Just Worked™ to let me switch between the different
>> variants.
> Now that i'm playing with this feature specifically, it occurs to me
> that the primary use case is likely to be where there are two or maybe 3
> duplicates of a given message.
> In this use case, the emacs bindings could be a bit more ergonomic.
> The current binding is '%', which then prompts the user for "duplicate:"
> and the user is obliged to type in the number of the duplicate that they
> want.
> When i'm looking at duplicate 1/2, the obvious answer is i want 2/2.
> when i'm looking at 2/2, i want 1/2.  The current interface, while nice
> and fully flexible makes me do extra work :)
> Here's a proposed update:
>  - '%' on its own moves to duplicate n+1 (wrapping around from the max
>    back to 1)
>  - prefixed '%' (that is, 'C-u %') yields the current behavior where the
>    user can enter which specific duplicate they want.
> My elisp remains rudimentary at best, so i don't know that i know how to
> make this fix (in particular, i'm not sure how i retrieve the current
> values N/M, in order to calculate mod_M(N)+1), but i'd be happy to test
> any patch that does this.
> I don't know how people feel about changing the default keybindings on a
> relatively recent feature -- would it be upsetting to users if the
> behavior of % changed in this subtle way?

Personally I think it would make sense to have % cycle through
duplicates, with some indicator of which duplicate we're currently on.
I wouldn't even bother with the suggested prefixed behavior if we just
had the cycle behavior.

Also thanks to folks for working on this.

notmuch mailing list --
To unsubscribe send an email to