nnir - articles order

nnir - articles order

[Bruno Tavernier]

Hello,

I started to used nnir with imap (latest git version) and it works quite
well.
However, I noticed that the order in which the articles appear is
opposite to the way gnus handle them usually.

For instance gnus displays the oldest first and the newest at the bottom
and nnir does the opposite.

gnus
,----
| ⚑     45k 13/06/2004 12:05     Joe Corneli  musical-letters.el (revision) 
| ⚑     14k 22/07/2004 20:29  Marcelo Toledo  chronometer.el --- a chronograph for emacs 
| ⚑     19k 10/09/2005 19:31 Daniel Brockman  Re: volume.el --- tweak your sound card volume from Emacs 
`----

nnir
,----
| ✉.   0.0k 10/09/2005 19:31 Daniel Brockman  Re: volume.el --- tweak your sound card volume from Emacs
| ✉.   0.0k 22/07/2004 20:29  Marcelo Toledo  chronometer.el --- a chronograph for emacs
| ✉.   0.0k 13/06/2004 12:05     Joe Corneli  musical-letters.el (revision)
`----

Is that a bug or a feature? :-)

-- 
Bruno


PS: I search for a variable to tune this behavior but couldn't find one.

Re: nnir - articles order

[Steinar Bang]

>>>>> Bruno Tavernier <tavernier.bruno@gmail.com>:

> For instance gnus displays the oldest first and the newest at the
> bottom and nnir does the opposite.

I'm guessing (I haven't checked the code or earlier postings by Andrew
Cohen on the subject) nnir aims to list by the way the search engine
returns the results (which depends on the search engine, but usually
relevance calculation has something to do with the returned order).

Re: nnir - articles order

[Andrew Cohen]

>>>>> "Steinar" == Steinar Bang <sb@dod.no> writes:

>>>>> Bruno Tavernier <tavernier.bruno@gmail.com>:
    >> For instance gnus displays the oldest first and the newest at the
    >> bottom and nnir does the opposite.

    Steinar> I'm guessing (I haven't checked the code or earlier
    Steinar> postings by Andrew Cohen on the subject) nnir aims to list
    Steinar> by the way the search engine returns the results (which
    Steinar> depends on the search engine, but usually relevance
    Steinar> calculation has something to do with the returned order).

That's right (more or less---there are complications when searching more
than one group).

Although I haven't played with it, you might be able to sort the summary
buffer with the usual sorting functions. I'd be curious to know how well
this works.

Re: nnir - articles order

[Andrew Cohen]

>>>>> "Andy" == Andrew Cohen <cohen@andy.bu.edu> writes:

>>>>> "Steinar" == Steinar Bang <sb@dod.no> writes:
>>>>> Bruno Tavernier <tavernier.bruno@gmail.com>:
    >>> For instance gnus displays the oldest first and the newest at
    >>> the bottom and nnir does the opposite.

    Steinar> I'm guessing (I haven't checked the code or earlier
    Steinar> postings by Andrew Cohen on the subject) nnir aims to list
    Steinar> by the way the search engine returns the results (which
    Steinar> depends on the search engine, but usually relevance
    Steinar> calculation has something to do with the returned order).

    Andy> That's right (more or less---there are complications when
    Andy> searching more than one group).

    Andy> Although I haven't played with it, you might be able to sort
    Andy> the summary buffer with the usual sorting functions. I'd be
    Andy> curious to know how well this works.


And it seems to work just fine. IMAP searching doesn't in fact assign a
retrieval score, so the resulting article list can be in any
order. Since the gnus `default' is increasing article number, and this
can be munged by the user into any order, I'll change it to conform to
this in my next push.

So it was both a feature and a bug:)

Re: nnir - articles order

[Bruno Tavernier]

Andrew Cohen <cohen@andy.bu.edu> writes:

>     Andy> Although I haven't played with it, you might be able to sort
>     Andy> the summary buffer with the usual sorting functions. I'd be
>     Andy> curious to know how well this works.
>
>
> And it seems to work just fine. IMAP searching doesn't in fact assign a
> retrieval score, so the resulting article list can be in any
> order. Since the gnus `default' is increasing article number, and this
> can be munged by the user into any order, I'll change it to conform to
> this in my next push.

Thanks for the explanation. And glad to see that it can be modified.

Gnus is evolving really fast these days, gnus devs you're doing a great
job! Much appreciated. :-)

-- 
Bruno