oldest-first

oldest-first

[Tom Hirschowitz]

Hi all,

The order returned by notmuch with the oldest-first option looks wrong
to me: as far as I can see, threads are sorted according to their oldest
unread message, but in any case, it is not the converse of the
newest-first ordering.

Is this a bug? And if not, how hard would it be to add an option for
getting the converse of newest-first?

Thanks in advance for any help!
Tom

Re: oldest-first

[David Bremner]

Tom Hirschowitz <tom.hirschowitz@univ-smb.fr> writes:

> Hi all,
>
> The order returned by notmuch with the oldest-first option looks wrong
> to me: as far as I can see, threads are sorted according to their oldest
> unread message, but in any case, it is not the converse of the
> newest-first ordering.
>
> Is this a bug? And if not, how hard would it be to add an option for
> getting the converse of newest-first?
>

There is the following documentation in notmuch-search(1).

     Note: The thread order will be distinct between these two options (beyond being sim‐
     ply reversed). When sorting by oldest-first the threads will be sorted by the oldest
     message  in each thread, but when sorting by newest-first the threads will be sorted
     by the newest message in each thread.

If what you are seeing is consistent with that, then I guess it's
officially not a bug.

I haven't looked at the code in question recently enough to estimate the
difficulty of adding another sort order. Can you explain why it's an
important feature for you?

d

Re: oldest-first

[Tom Hirschowitz]

Thanks for your answer.

This is indeed not a bug then. It's not important, but my preferred sort order would be the opposite of newest-first.
I prefer seeing older threads first, but often find myself missing recent messages in threads because they have older unread messages hence are considered old. 

Tom

Re: oldest-first

[Ryan Tate]

[-- Attachment #1: Type: text/plain, Size: 1181 bytes --]


> On Mar 6, 2020, at 10:47 AM, David Bremner <david@tethera.net> wrote:
> 
> There is the following documentation in notmuch-search(1).
> 
>     Note: The thread order will be distinct between these two options (beyond being sim‐
>     ply reversed). When sorting by oldest-first the threads will be sorted by the oldest
>     message  in each thread, but when sorting by newest-first the threads will be sorted
>     by the newest message in each thread.
> 
> If what you are seeing is consistent with that, then I guess it's
> officially not a bug.

The documentation seems to be in error, assuming you have copied it correctly. It says the thread orders are not strictly inverse between the two options, but then describes them precisely inverse. 

Perhaps the word “unread” was unintentionally elided by the doc author, such that you could correct with the capitalized addition:

> When sorting by oldest-first the threads will be sorted by the oldest UNREAD
>     message  in each thread, but when sorting by newest-first the threads will be sorted
>     by the newest message in each thread.


This would match the behavior described by Tom. 


[-- Attachment #2: Type: text/html, Size: 2005 bytes --]

Re: oldest-first

[Ryan Tate]

> On Mar 6, 2020, at 11:50 AM, Ryan Tate <ryantate@ryantate.com> wrote:

> The documentation seems to be in error,


No, I didn’t think about it hard enough, sorry. (Reversing threads sorted by newest message vs sorting from scratch by oldest message.)

>