N versus . for unseen messages

N versus . for unseen messages

[Kai Großjohann]

Some of my unseen messages are marked `N', some are marked `.'.  It's
clear to me that `N' is used only in nnimap groups, but then why do I
sometimes see `.' in nnimap groups?

May I suggest to unify the two marks?

Also, I have the strange feeling (nothing substantial, though) that
sometimes not all articles are marked `N' (or `.') that should be
marked like that.  The vague feeling is that the first new message in
a group does not get the N mark, whereas the second and subsequent
messages do.

Oh, boy, what a wonderful report.  I should really know better.  But
then, maybe some of you have already seen this?  If not, I'll try to
figure out what happens for real.

kai
-- 
Symbol's function definition is void: signature

Re: N versus . for unseen messages

[Nuutti Kotivuori]

Kai Großjohann wrote:
> Some of my unseen messages are marked `N', some are marked `.'.
> It's clear to me that `N' is used only in nnimap groups, but then
> why do I sometimes see `.' in nnimap groups?

'N' comes from the IMAP server's recent status, '.' comes from Gnus's
idea of what's new and what's not. Apparently these two are running at
the same time - and accesses outside Gnus atleast do change the state
of both.

The user should be able to select which one (or maybe both?) is to be
used - since for others the IMAP server status might never be recent
and for others it might be the most accurate recent indicator if they
use other mailreaders as well.

-- Naked

Re: N versus . for unseen messages

[Kai Großjohann]

Nuutti Kotivuori <nuutti.kotivuori@smarttrust.com> writes:

> Kai Großjohann wrote:
> > Some of my unseen messages are marked `N', some are marked `.'.
>> It's clear to me that `N' is used only in nnimap groups, but then
>> why do I sometimes see `.' in nnimap groups?
>
> 'N' comes from the IMAP server's recent status, '.' comes from Gnus's
> idea of what's new and what's not. Apparently these two are running at
> the same time - and accesses outside Gnus atleast do change the state
> of both.

I never access my mail from outside Gnus.  At least, not knowingly.

kai
-- 
Symbol's function definition is void: signature

Re: N versus . for unseen messages

[Nuutti Kotivuori]

Kai Großjohann wrote:
> Nuutti Kotivuori <nuutti.kotivuori@smarttrust.com> writes:
>> Kai Großjohann wrote:
>>> Some of my unseen messages are marked `N', some are marked `.'.
>>> It's clear to me that `N' is used only in nnimap groups, but then
>>> why do I sometimes see `.' in nnimap groups?
>>
>> 'N' comes from the IMAP server's recent status, '.' comes from
>> Gnus's idea of what's new and what's not. Apparently these two are
>> running at the same time - and accesses outside Gnus atleast do
>> change the state of both.
> 
> I never access my mail from outside Gnus.  At least, not knowingly.

Fetchmail? Biff? Maybe even some 'g', 'M-g' or other invocation could
unwillingly change the recent flag on server. The RFC says that it is
set for exactly the first seesion to _have been notified_ about this
message - only STATUS command does not alter recent. So unless care is
taken that a mailbox is never selected or examined before actually
entered, it's impossible to properly get the recent flags from the
server. (And I believe getting the smallest/biggest uid for messages
in a group is impossible without selecting the group. ((The status
command can return NEXTUID, which is the next uid to be reserved -
though that is ofcourse a different matter)))

And all this is ofcourse expecting a fully standards compliant server
- I won't start imagining what bogosities some servers do with recent.

But I don't know anything about nnimap really, so perhaps Simon could
answer this one better.

-- Naked

Re: N versus . for unseen messages

[Kai Großjohann]

Nuutti Kotivuori <nuutti.kotivuori@smarttrust.com> writes:

> Fetchmail? Biff? Maybe even some 'g', 'M-g' or other invocation could
> unwillingly change the recent flag on server. The RFC says that it is
> set for exactly the first seesion to _have been notified_ about this
> message - only STATUS command does not alter recent.

Ah, maybe it is like this:

* a message arrives.  I hit g, but then don't look at that part of the
  group buffer, or the summary buffer the message is in.

* another message arrives.  I hit g again.

Maybe the first message isn't marked N anymore because I've hit g
again.  Maybe the first message is still marked . because I haven't
seen it.

kai
-- 
Symbol's function definition is void: signature