Wrong number of articles from completely read group

Wrong number of articles from completely read group

[Tommy Kelly]

I'm using nnimap to Gmail.
If I have a group where all the articles are read and I open it, I get
asked how many articles I want to show. The number I'm actually shown is
always less than the number I ask for (unless I ask for 1). For example,
asking for 2 gives me 1. Asking for 4 gives me three. But it's not
always n-1. For example, asking for 20 gives me 16.

Sam thing happens if I supply the number as a prefix arg.

Tommy

Re: Wrong number of articles from completely read group

[Lars Magne Ingebrigtsen]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> If I have a group where all the articles are read and I open it, I get
> asked how many articles I want to show. The number I'm actually shown is
> always less than the number I ask for (unless I ask for 1). For example,
> asking for 2 gives me 1. Asking for 4 gives me three. But it's not
> always n-1. For example, asking for 20 gives me 16.

Gnus doesn't know what articles exist, so it asks for articles in the
range of n-x to n.  For IMAP, there's often fewer than x+1 articles in
that range.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> I'm using nnimap to Gmail.
> If I have a group where all the articles are read and I open it, I get
> asked how many articles I want to show. The number I'm actually shown is
> always less than the number I ask for (unless I ask for 1).

Actually it's worse than that. Sometimes I just get told there's no
unread news (which I already know) and I'm kept in the Groups
buffer. But only sometimes. 

Not sure where to look to start even guessing on a cause.

Tommy

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Tommy Kelly <tommy.kelly@verilab.com> writes:
>
>> If I have a group where all the articles are read and I open it, I get
>> asked how many articles I want to show. The number I'm actually shown is
>> always less than the number I ask for (unless I ask for 1). For example,
>> asking for 2 gives me 1. Asking for 4 gives me three. But it's not
>> always n-1. For example, asking for 20 gives me 16.
>
> Gnus doesn't know what articles exist, so it asks for articles in the
> range of n-x to n.  For IMAP, there's often fewer than x+1 articles in
> that range.

Ah. Where n is the number of the most recent article? And in which case
if any articles have been moved or deleted from the group then there
will be fewer articles than expected. I see.

Still doesn't explain, though, why sometimes I'm just told "No unread news"
no matter how many articles I request.

Tommy

Re: Wrong number of articles from completely read group

[Tassilo Horn]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

Hi Lars,

>> If I have a group where all the articles are read and I open it, I get
>> asked how many articles I want to show. The number I'm actually shown is
>> always less than the number I ask for (unless I ask for 1). For example,
>> asking for 2 gives me 1. Asking for 4 gives me three. But it's not
>> always n-1. For example, asking for 20 gives me 16.
>
> Gnus doesn't know what articles exist, so it asks for articles in the
> range of n-x to n.  For IMAP, there's often fewer than x+1 articles in
> that range.

I think that applies to any backend that allows for modifications,
right?  In org-gnus.el, I have this really beatiful snippet to reliably
open a summary buffer at least showing ARTICLE (article number or
message id).

                   ;; articles is 1 and group-opened nil initially
                   (while (and (not group-opened)
                               ;; stop on integer overflows
                               (> articles 0))
                     (setq group-opened (gnus-group-read-group
                                         articles nil group)
                           articles (if (< articles 16)
                                        (1+ articles)
                                      (* articles 2))))
                   (if group-opened
                       (gnus-summary-goto-article article nil t)
                     (message "Couldn't follow gnus link.  %s"
                              "The summary couldn't be opened."))

It works, but is there something better that doesn't rely on some
optional feature (like the registry) and works independently from the
used backend?

Bye,
Tassilo

Re: Wrong number of articles from completely read group

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Still doesn't explain, though, why sometimes I'm just told "No unread
> news" no matter how many articles I request.

Did you say what kind of IMAP server you were using?  (I'm sure it's in
this thread somewhere, but...).

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:
> Did you say what kind of IMAP server you were using?  (I'm sure it's in
> this thread somewhere, but...).

Gmail (Google apps to be precise).

Tommy

Re: Wrong number of articles from completely read group

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Steinar Bang <sb@dod.no> writes:
>> Did you say what kind of IMAP server you were using?  (I'm sure it's in
>> this thread somewhere, but...).

> Gmail (Google apps to be precise).

Um... does "google apps" means that this is something different than
IMAP access to a regular gmail account?

Because I thought gmail IMAP was on the list of IMAP clients that saw
testing with the new nnimap...?

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:
> Um... does "google apps" means that this is something different than
> IMAP access to a regular gmail account?
>
> Because I thought gmail IMAP was on the list of IMAP clients that saw
> testing with the new nnimap...?

Google Apps is just the "business" version of gmail (plus a bunch of
other stuff). As far as I'm aware it's identical in IMAP-ishness to
regular gmail. 

Tommy

Re: Wrong number of articles from completely read group

[Tibor Simko]

On Mon, 20 Dec 2010, Tommy Kelly wrote:
> Still doesn't explain, though, why sometimes I'm just told "No unread
> news" no matter how many articles I request.

Could it be that there are none because the group underwent many
moves/deletes so the asked interval still does not cover any message
when you `C-u RET' some number?

Is everything OK when you `M-g' the given group and then `C-u RET'?

Best regards
-- 
Tibor Simko

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Tibor Simko <tibor.simko@cern.ch> writes:
> Could it be that there are none because the group underwent many
> moves/deletes so the asked interval still does not cover any message
> when you `C-u RET' some number?

Doubt it. The group has undergone *no* moves/deletes. And it's
intermittent. For example, it could be happening (i.e. now messages get
shown), then I send a message which gets Gcc'ed to the group. From there
on it's fine (until it breaks again). Also, I'm pretty sure if I close
Gnus (and emacs maybe) and then restart, it's fine.

> Is everything OK when you `M-g' the given group and then `C-u RET'?

I'll try next time it happens.

thanks for the suggestions,
Tommy

Re: Wrong number of articles from completely read group

[Tommy Kelly]

Tibor Simko <tibor.simko@cern.ch> writes:
> Is everything OK when you `M-g' the given group and then `C-u RET'?

Tommy Kelly <tommy.kelly@verilab.com> writes:
>> I'll try next time it happens.

OK, I tried it and yes, M-g does "fix" it. That said, it also appeared
to treat as unread some articles I had just read.

In general, the marking and accepting-that-they've-been-marked as read
of articles seems a bit fragile at the moment.

Tommy

Re: Wrong number of articles from completely read group

[Lars Magne Ingebrigtsen]

Tassilo Horn <tassilo@member.fsf.org> writes:

> I think that applies to any backend that allows for modifications,
> right?

Yes.

> In org-gnus.el, I have this really beatiful snippet to reliably
> open a summary buffer at least showing ARTICLE (article number or
> message id).

[...]

> It works, but is there something better that doesn't rely on some
> optional feature (like the registry) and works independently from the
> used backend?

Not at the moment, but there should be.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen