Re: strange behaviour in one group

Re: strange behaviour in one group

[Kai Großjohann]

Kester Clegg <kester@RemoveThisBit.cs.york.ac.uk> writes:

> This is continuing an earlier posting to this group, but I still haven't
> solved the problem.  In one group, I have deleted / expired all
> messages.  The .overview file is empty.

What about the .marks file?

> However, when I hit return or space, I always get asked for how many
> messages I want to download, 395.  Hitting return again gnus scans
> through headers but says 'no mail'.  The group parameters (using G
> E) are:
>
> ("nnml+private:private" 3
>  ((1 . 714))

This means that you have read articles 1 through 714, I think.

>  nil "nnml:private"
>  ((auto-expire . t)
>   (expiry-wait . 31)
>   (display . all)
>   (visible . t)
>   (gnus-summary-line-format "%U%R%I %d %~(cut-left 9)~(max-right 2)o:%~(cut-left 11)~(max-right 2)o %>%(%[%n]%) %s\n")
>   (posting-style
>    (signature-file "~/.signature"))))
>
> Does anyone know what the ((1 . 714)) part refers to?  If I change 714
> to another number, I get a false number of unread messages in that
> group.  If I try to enter the group, the number always gets reset to
> 714.  What is the 714 referring to, where does it get it from and how
> can I get the group to behave normally, like all the others?  Will
> editing the group parameters help me?

Maybe the 714 comes from the .marks file, if there is one.
-- 
A preposition is not a good thing to end a sentence with.

Re: strange behaviour in one group

[Kai Großjohann]

Kester Clegg <kester@RemoveThisBit.cs.york.ac.uk> writes:

> kai.grossjohann@gmx.net (Kai Großjohann) writes:
>
> [...]
>
>> >
>> > ("nnml+private:private" 3
>> >  ((1 . 714))
>> 
>> This means that you have read articles 1 through 714, I think.
>
> OK, so as there are no articles remaining in that group, this is wrong?

This is unclear to me.  Hm.  It could be right.  If you copy a new
message to that group, with `B c', say, does it then get numbered
715?  If this is the case, then all's well, I think.

> [...]
>
>> Maybe the 714 comes from the .marks file, if there is one.
>
> No, I don't have one.  But this means it comes from somewhere else, I
> guess.  As I said, it keeps resetting that figure, even if you change
> it, so somewhere there must a record of something that has got
> corrupted.  The .newsrc.eld I guess, but I'm a bit nervous of editing
> that. 

Yeah.  But you could try M-x gnus-group-clear-data RET.  That might
do something useful without requiring to edit the file.
-- 
A preposition is not a good thing to end a sentence with.

Re: strange behaviour in one group

[Kai Großjohann]

Kester Clegg <kester@RemoveThisBit.cs.york.ac.uk> writes:

> kai.grossjohann@gmx.net (Kai Großjohann) writes:
>
>> Kester Clegg <kester@RemoveThisBit.cs.york.ac.uk> writes:
>> 
>> > Where does it get the magic figure of 397 from?  This figure isn't in
>> > the .newsrc.eld file, but I'm convinced it lies behind gnus getting
>> > muddled up for this group. 
>> 
>> Is there a cached article with number 715 - 397 +/- 2?
>
> No, no cached articles for this group.

Hm.  Really strange.  Then the only thing that comes to my mind is to
read the source to figure out what happens when you hit RET.  I'd
search for the spot where it checks gnus-large-newsgroup.  Then I'd
try to figure out what was compared against gnus-large-newsgroup.
Then I'd try to figure out why was the value so large.

(You could set gnus-large-newsgroup to a large value, of course, but
that'd be cheating, right?)
-- 
A preposition is not a good thing to end a sentence with.