Can't mark negative article numbers

Can't mark negative article numbers

[Robin S. Socha]

Okay, so I say ^ in the summary buffer and then !. No workie. Why?

Re: Can't mark negative article numbers

[ShengHuo ZHU]

"Robin S. Socha" <robin-dated-1005003013.83ef58@socha.net> writes:

> Okay, so I say ^ in the summary buffer and then !. No workie. Why?

Because Gnus dunno the exact article number or it is an article in a
foreign group.

ShengHuo

Re: Can't mark negative article numbers

[Robin S. Socha]

* ShengHuo ZHU <zsh@cs.rochester.edu> writes:
> "Robin S. Socha" <robin-dated-1005003013.83ef58@socha.net> writes:

>> Okay, so I say ^ in the summary buffer and then !. No workie. Why?

> Because Gnus dunno the exact article number or it is an article in a
> foreign group.

No. It's local news. Or olds. Whatever. WIBNI it worked?

Re: Can't mark negative article numbers

[ShengHuo ZHU]

"Robin S. Socha" <robin-dated-1005004530.e2f224@socha.net> writes:

> * ShengHuo ZHU <zsh@cs.rochester.edu> writes:
>> "Robin S. Socha" <robin-dated-1005003013.83ef58@socha.net> writes:
>
>>> Okay, so I say ^ in the summary buffer and then !. No workie. Why?
>
>> Because Gnus dunno the exact article number or it is an article in a
>> foreign group.
>
> No. It's local news. Or olds. Whatever. WIBNI it worked?

Is there an Xref header in the article?

ShengHuo

Re: Can't mark negative article numbers

[Matthias Wiehl]

"Robin S. Socha" <robin-dated-1005003013.83ef58@socha.net> writes:
> Okay, so I say ^ in the summary buffer and then !. No workie. Why?

,----[ <vxkae8uu35z.fsf@cinnamon.vanillaknot.com> ]
| Karl Kleinpaste <karl@charcoal.com> writes:
| 
| Yes.  This is a fairly fundamental issue.  You see, when you chase to
| a parent article, it may or may not have come from within the same
| group as the group you're currently reading.  If you're reading
| gnu.emacs.gnus and chase to an article's parent, you may be reading an
| article posted (only) to news.software.readers.  What if you don't
| read n.s.r?  What does it mean to mark an article to which you don't
| even have access in the normal case?
| 
| Generally, if Gnus can figure out where in the current group a
| chased-to article belongs, it will not number it negatively, but will
| show it with its normal within-the-group numbering.  It is only
| articles from outside the current group that get such treatment.
`----

,----[ <vafk7xizui7.fsf@lucy.cs.uni-dortmund.de> ]
| Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
| 
| When you hit `^', Gnus looks to see if it has downloaded the header
| for this message already.  If so, it uses the article number found
| there.  But if not, then Gnus fetches the right article by message-id,
| but it does not look to find the right article number for this
| message-id.  Note that the article might come from a different group,
| so it's not possible to always assign a good number.
`----

Re: Can't mark negative article numbers

[Kai Großjohann]

"Robin S. Socha" <robin-dated-1005004530.e2f224@socha.net> writes:

> No. It's local news. Or olds. Whatever. WIBNI it worked?

Are you volunteering?

kai
-- 
Lisp is kinda like tpircstsoP

Re: Can't mark negative article numbers

[Florian Weimer]

"Yair Friedman (Jerusalem)" <YAIRFR@amdocs.com> writes:

> > > > That does not address the point of Gnus marking messages in groups
> > > > to which you are not subscribed.  Given that it is not obvious what
> > > > group a referred parent is from,
> > > 
> > > Doesn't the Newsgroups header say?
> > 
> > According to some people on the USEFOR list, no.  User agents should
> > not interpret the Newsgroups: header.
> 
> What about Xref then?

The Xref header is optional.

Re: Can't mark negative article numbers

[Russ Allbery]

Yair Friedman (Jerusalem) <YAIRFR@amdocs.com> writes:
> Florian Weimer <fw@deneb.enyo.de> writes:

>> According to some people on the USEFOR list, no.  User agents should
>> not interpret the Newsgroups: header.

> What about Xref then?

That should be accurate if present.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>

Re: Can't mark negative article numbers

[Yair Friedman (Jerusalem)]

Florian Weimer <fw@deneb.enyo.de> writes:

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
> 
> > > That does not address the point of Gnus marking messages in groups
> > > to which you are not subscribed.  Given that it is not obvious what
> > > group a referred parent is from,
> > 
> > Doesn't the Newsgroups header say?
> 
> According to some people on the USEFOR list, no.  User agents should
> not interpret the Newsgroups: header.

What about Xref then?

Re: Can't mark negative article numbers

[Florian Weimer]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> > That does not address the point of Gnus marking messages in groups
> > to which you are not subscribed.  Given that it is not obvious what
> > group a referred parent is from,
> 
> Doesn't the Newsgroups header say?

According to some people on the USEFOR list, no.  User agents should
not interpret the Newsgroups: header.

(Don't flame me, that's not my opinion.)

Re: Can't mark negative article numbers

[Paul Jarc]

Stainless Steel Rat <ratinox@peorth.gweep.net> writes:
> * prj@po.cwru.edu (Paul Jarc)  on Sat, 21 Apr 2001
> | Simple: allow marks to be set/unset on the article in the group that
> | it came from, if you're subscribed to that group.
> 
> That does not address the point of Gnus marking messages in groups to which
> you are not subscribed.

I wasn't attempting to address that point; I'm not interested in doing
so.

> Given that it is not obvious what group a referred parent is from,

It's obvious to Gnus, because nnfoo-retrieve-article reports the group
that the article came from.  So then we could also make it obvious to
the user by tweaking the summary line - different colorization, or a
pseudo-mark, or something.

> that will lead to much confusion on the parts of the users
> because Gnus is silently doing two very different things.

As I see it, Gnus would always be doing the same thing: changing the
marks on the article.  It's just that marking an article with 'tick,
say, won't necesarily make that article show up the next time you
enter this summary buffer, because the article might not be in this
group.  So we could change the summary line to make it clear when this
is the case, and we could also give the user a way to say that Gnus
should pretend that this article is in this group.  Gnus would
maintain a list of "foreign articles" for each group, and fetch their
headers along with the group's articles' headers.  Assuming this was
desired; maybe people would be satisfied with the different summary
line to indicate a foreign article.  I would.


paul

Re: Can't mark negative article numbers

[Stainless Steel Rat]

* Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)  on Sun, 22 Apr 2001
| Doesn't the Newsgroups header say?

Yes.  But I suspect a lot of people trim down gnus-visible-headers because
by default it displays a lot of generally less than useful information,
especially when the current group is already displayed on the summary
buffer mode line.

Even so, it would be not so obvious that there has been a change in groups
unless one is paying specific attention to it.  It would be very easy for
something to slip, leaving Gnus with the blame for a problem instead of
being the fix for one.
-- 
Rat <ratinox@peorth.gweep.net>    \ Happy Fun Ball contains a liquid core,
Minion of Nathan - Nathan says Hi! \ which, if exposed due to rupture, should
PGP Key: at a key server near you!  \ not be touched, inhaled, or looked at.

Re: Can't mark negative article numbers

[Karl Eichwalder]

Early this year I reported something similar; here's my last message:

Karl Eichwalder <keichwa@gmx.net> writes:

> Karl Kleinpaste <karl@charcoal.com> writes:
> > It is only articles from outside the current group that get such
> > treatment.
> 
> Unfortunately, not ;)
> 
> In addition, I like to propose to change the message to make it sound
> more userfriendly.  At the moment it is nothing more than a debugging
> message.  Please, add something like this:
> 
>     Can't mark negative article numbers; article doesn't belong to
>     current group.
> 
> Since forthcoming Emacs will be able to deal nicely with multiline
> minibuffer messages it shuold be allowed to output more verbose error or
> warning messages.

Karl Kleinpaste <karl@charcoal.com> writes:

> Thomas Skogestad <tskogest@jusstud.uio.no> writes:
> > Gotta unmark it as it is now marked as read.
> > 6. No can do. "Can't mark negative article numbers"

> If the parent article came from the same group, it should be given its
> normal article number, and so should be un-markable.

It should but it still doesn't work.

-- 
work : ke@suse.de                          |                   ,__o
     : http://www.suse.de/~ke/             |                 _-\_<,
home : keichwa@gmx.net                     |                (*)/'(*)

Re: Can't mark negative article numbers

[Kai Großjohann]

On 22 Apr 2001, Stainless Steel Rat wrote:

> That does not address the point of Gnus marking messages in groups
> to which you are not subscribed.  Given that it is not obvious what
> group a referred parent is from,

Doesn't the Newsgroups header say?

kai
-- 
The passive voice should never be used.

Re: Can't mark negative article numbers

[Stainless Steel Rat]

* prj@po.cwru.edu (Paul Jarc)  on Sat, 21 Apr 2001
| Simple: allow marks to be set/unset on the article in the group that
| it came from, if you're subscribed to that group.

That does not address the point of Gnus marking messages in groups to which
you are not subscribed.  Given that it is not obvious what group a referred
parent is from, that will lead to much confusion on the parts of the users
because Gnus is silently doing two very different things.

I suppose warnings can be printed, but that is ugly.
-- 
Rat <ratinox@peorth.gweep.net>    \ Happy Fun Ball contains a liquid core,
Minion of Nathan - Nathan says Hi! \ which, if exposed due to rupture, should
PGP Key: at a key server near you!  \ not be touched, inhaled, or looked at.

Re: Can't mark negative article numbers

[Paul Jarc]

Anssi Saari <as@sci.fi> writes:
> On Thu, Apr 19, 2001 at 03:23:35PM -0400, Paul Jarc wrote:
> > > The best you can do is mark the child which led to the interesting
> > > parent, and chase to the parent again when you come back.
>  
> > Yes, but that's not the best that Gnus can do.
> 
> Well, what should Gnus do? Allow marking parents from different groups
> and getting them from the server?

Simple: allow marks to be set/unset on the article in the group that
it came from, if you're subscribed to that group.

> In effect, entering one group might require Gnus to get headers for
> multiple groups because of such marked articles.

That's not necessary.  Using ^ to pull in an article from another
group, and then setting marks on that article, doesn't affect the
group(s) the article belongs to.  It just affects the marks it has, in
whichever groups it *is* in.


paul

Re: Can't mark negative article numbers

[Anssi Saari]

On Thu, Apr 19, 2001 at 03:23:35PM -0400, Paul Jarc wrote:
> > The best you can do is mark the child which led to the interesting
> > parent, and chase to the parent again when you come back.
 
> Yes, but that's not the best that Gnus can do.

Well, what should Gnus do? Allow marking parents from different groups
and getting them from the server? In effect, entering one group might
require Gnus to get headers for multiple groups because of such marked
articles. I think nnkiboze already does this sort of thing? Is there a
problem with this?

Re: Can't mark negative article numbers

[Paul Jarc]

Karl Kleinpaste <karl@charcoal.com> writes:
> E.g. somebody writes to news.software.readers with followups set to
> gnu.emacs.gnus.  You see a followup in g.e.g, but you're not
> subscribed to n.s.r.

Consider the case where you *are* subscribed.  That's the case in
question here.

> The best you can do is mark the child which led to the interesting
> parent, and chase to the parent again when you come back.

Yes, but that's not the best that Gnus can do.


paul

Re: Can't mark negative article numbers

[Karl Kleinpaste]

Thomas Skogestad <tskogest@jusstud.uio.no> writes:
> Gotta unmark it as it is now marked as read.
> 6. No can do. "Can't mark negative article numbers"

If the parent article came from the same group, it should be given its
normal article number, and so should be un-markable.

However, if the parent came from some other group, which might not
even be a group to which you're subscribed, there's no means by which
to assign it a state of "un-readness" because it's not visible in the
right group in the first place.

E.g. somebody writes to news.software.readers with followups set to
gnu.emacs.gnus.  You see a followup in g.e.g, but you're not
subscribed to n.s.r.  You `^' to the n.s.r parent, and you want to
mark it.  What should Gnus do?  There's little it *can* do, short of
subscribing you to n.s.r.

The best you can do is mark the child which led to the interesting
parent, and chase to the parent again when you come back.

Can't mark negative article numbers

[Thomas Skogestad]

1. Person A posts a followup to an article posted to group foo and cross
posts it to group bar.

2. I read the follow-up in group bar, but I also read group foo.

3. I wish to read the original message and press "^".

4. Interesting! And long. But I've got a bus to catch!

5. So I'll read it later. Ooops! Gotta unmark it as it is now marked as
read.

6. No can do. "Can't mark negative article numbers"

--

For nordmenn med universitets-ISP:

Praktisk eksempel: ntb.innenriks.nytt og ntb.diskusjon.

-- 
thomas.skogestad@jusstud.uio.no
http://quimby.gnus.org/circus/problem.gif
http://www.messengers-of-messiah.org/~csebold/emacs/why.phtml

Re: Can't mark negative article numbers

[Karl Eichwalder]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> You could use `A T' to fetch the whole thread.  That fetches all old
> headers, not just a single one.

Why do I always forget about the `A <x>' keys combinations?  Thanks for
the reminder!

Karl Kleinpaste <karl@charcoal.com> writes:

> Karl Eichwalder <keichwa@gmx.net> writes:
> > It isn't possible to mark/tick an article fetched with
> > `gnus-summary-refer-parent-article'.

> Yes.  This is a fairly fundamental issue.  You see, when you chase to
> a parent article, it may or may not have come from within the same
> group as the group you're currently reading.  If you're reading
> gnu.emacs.gnus and chase to an article's parent, you may be reading an
> article posted (only) to news.software.readers.  What if you don't
> read n.s.r?  What does it mean to mark an article to which you don't
> even have access in the normal case?

Okay, I see the point.  To answer your question: If the article isn't
part of the group gnus might try to copy the article (mail groups)
and/or put it into the cache.

> Generally, if Gnus can figure out where in the current group a
> chased-to article belongs, it will not number it negatively, but will
> show it with its normal within-the-group numbering.

No, I doesn't work for me at all.  To put the question different: Why
does my Gnus always number id-fetched articles negatively?  Obviously,
it can't figure out that the fetched article (`^') belongs to the
current group.

> It is only articles from outside the current group that get such
> treatment.

Unfortunately, not ;)

In addition, I like to propose to change the message to make it sound
more userfriendly.  At the moment it is nothing more than a debugging
message.  Please, add something like this:

    Can't mark negative article numbers; article doesn't belong to
    current group.

Since forthcoming Emacs will be able to deal nicely with multiline
minibuffer messages it shuold be allowed to output more verbose error or
warning messages.

Thanks for all your help!

-- 
work : ke@suse.de                          |                   ,__o
     : http://www.suse.de/~ke/             |                 _-\_<,
home : keichwa@gmx.net                     |                (*)/'(*)

Re: Can't mark negative article numbers

[Karl Kleinpaste]

Karl Eichwalder <keichwa@gmx.net> writes:
> It isn't possible to mark/tick an article fetched with
> `gnus-summary-refer-parent-article'.

Yes.  This is a fairly fundamental issue.  You see, when you chase to
a parent article, it may or may not have come from within the same
group as the group you're currently reading.  If you're reading
gnu.emacs.gnus and chase to an article's parent, you may be reading an
article posted (only) to news.software.readers.  What if you don't
read n.s.r?  What does it mean to mark an article to which you don't
even have access in the normal case?

Generally, if Gnus can figure out where in the current group a
chased-to article belongs, it will not number it negatively, but will
show it with its normal within-the-group numbering.  It is only
articles from outside the current group that get such treatment.

Re: Can't mark negative article numbers

[Kai Großjohann]

On 13 Jan 2001, Karl Eichwalder wrote:

> It isn't possible to mark/tick an article fetched with
> `gnus-summary-refer-parent-article'.  [...]  Is it possible to work
> around this problem?

You could use `A T' to fetch the whole thread.  That fetches all old
headers, not just a single one.  But `^' can fetch messages from a
different group, too.

kai
-- 
Be indiscrete.  Do it continuously.

Can't mark negative article numbers

[Karl Eichwalder]

It isn't possible to mark/tick an article fetched with
`gnus-summary-refer-parent-article'.  Trying to mark such an article
results in the error message:

    Can't mark negative article numbers

Is it possible to work around this problem?  If you'll need more info,
please ask.

-- 
work : ke@suse.de                          |                   ,__o
     : http://www.suse.de/~ke/             |                 _-\_<,
home : keichwa@gmx.net                     |                (*)/'(*)