Agent downloading even when there isn't any new articles

Agent downloading even when there isn't any new articles

[Henrik Enberg]

I just upgraded my Gnus to the latest version that contains Kevin's
2002-12-12 changes.  Now the agent enters every groups and does a (a
very small) bit of downloading everytime I do `J s'.  Previously it
only did that when I had `gnus-agent-consider-all-articles' set to t,
but now it does that regardless of that setting.

It's not really an important thing, but it can probably be annoying if
you have a slow and/or expensive connection.

Re: Agent downloading even when there isn't any new articles

[kgreiner]

Henrik Enberg <henrik@enberg.org> writes:

> I just upgraded my Gnus to the latest version that contains Kevin's
> 2002-12-12 changes.  Now the agent enters every groups and does a (a
> very small) bit of downloading everytime I do `J s'.  Previously it
> only did that when I had `gnus-agent-consider-all-articles' set to t,
> but now it does that regardless of that setting.
>
> It's not really an important thing, but it can probably be annoying if
> you have a slow and/or expensive connection.

I was able to reproduce this problem on my computer.  What I found was
that fetch session would, using the agent predicate, select articles
for automatic fetching (So far, so good).  Unfortunately, some of
these articles could not be fetched as they had expired on my server.
The problem is that these articles still satisfy the agent predicate
so, every time I invoke fetch session, gnus-agent tries to fetch them.

So, here's a question for the group, should gnus-agent-fetch-session
mark articles that it fails to fetch as read?  If so, what mark should
I use? ancient,killed,del,expirable

Kevin

Re: Agent downloading even when there isn't any new articles

[Henrik Enberg]

kgreiner@xpediantsolutions.com writes:

> Henrik Enberg <henrik@enberg.org> writes:
>
>> I just upgraded my Gnus to the latest version that contains Kevin's
>> 2002-12-12 changes.  Now the agent enters every groups and does a (a
>> very small) bit of downloading everytime I do `J s'.

[...]

> So, here's a question for the group, should gnus-agent-fetch-session
> mark articles that it fails to fetch as read?  If so, what mark should
> I use? ancient,killed,del,expirable

I'd say killed, since they don't exist.

Re: Agent downloading even when there isn't any new articles

[Kai Großjohann]

kgreiner@xpediantsolutions.com writes:

> So, here's a question for the group, should gnus-agent-fetch-session
> mark articles that it fails to fetch as read?  If so, what mark should
> I use? ancient,killed,del,expirable

Sometimes, Gnus shows yellow fg on black bg and, I think, a G (instead
of R/Y/K/...) for canceled articles.  That appears to be the right
mark.  But I don't know which mark it is.

-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: Agent downloading even when there isn't any new articles

[Lars Magne Ingebrigtsen]

kai.grossjohann@uni-duisburg.de (Kai Großjohann) writes:

> Sometimes, Gnus shows yellow fg on black bg and, I think, a G (instead
> of R/Y/K/...) for canceled articles.  That appears to be the right
> mark.  But I don't know which mark it is.

`gnus-canceled-mark'.

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

Re: Agent downloading even when there isn't any new articles

[Lars Magne Ingebrigtsen]

kai.grossjohann@uni-duisburg.de (Kai Großjohann) writes:

> Sometimes, Gnus shows yellow fg on black bg and, I think, a G (instead
> of R/Y/K/...) for canceled articles.  That appears to be the right
> mark.  But I don't know which mark it is.

`gnus-canceled-mark'.

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

Re: Agent downloading even when there isn't any new articles

[Lars Magne Ingebrigtsen]

kgreiner@xpediantsolutions.com writes:

> The current status of the patch, which has NOT been committed, is that
> I've fixed the fetch session code so that 'killed' and 'cache' marks
> are also ignored.  I also added code to mark any unfetchable articles
> as seen (i.e. read).  However, I don't like the idea of automatically
> marking articles as read so I'm going to take that code out
> tomorrow.

Well, if the article really is gone from the server, you might as
well mark it as read.  It's not going to reappear...

> I'm currently thinking that fetch session should simply display an
> information message that it was unable to fetch all selected
> articles.  Any comments?

It's quite normal for messages to be canceled/superseded/expired, so
you'd basically just display a message saying "I didn't fetch all the
articles" in (just about) every fetching session.  So I don't think
that's necessary.

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

Re: Agent downloading even when there isn't any new articles

[Lars Magne Ingebrigtsen]

kgreiner@xpediantsolutions.com writes:

> The current status of the patch, which has NOT been committed, is that
> I've fixed the fetch session code so that 'killed' and 'cache' marks
> are also ignored.  I also added code to mark any unfetchable articles
> as seen (i.e. read).  However, I don't like the idea of automatically
> marking articles as read so I'm going to take that code out
> tomorrow.

Well, if the article really is gone from the server, you might as
well mark it as read.  It's not going to reappear...

> I'm currently thinking that fetch session should simply display an
> information message that it was unable to fetch all selected
> articles.  Any comments?

It's quite normal for messages to be canceled/superseded/expired, so
you'd basically just display a message saying "I didn't fetch all the
articles" in (just about) every fetching session.  So I don't think
that's necessary.

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

Re: Agent downloading even when there isn't any new articles

[kgreiner]

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

> kgreiner@xpediantsolutions.com writes:
>
>> The current status of the patch, which has NOT been committed, is that
>> I've fixed the fetch session code so that 'killed' and 'cache' marks
>> are also ignored.  I also added code to mark any unfetchable articles
>> as seen (i.e. read).  However, I don't like the idea of automatically
>> marking articles as read so I'm going to take that code out
>> tomorrow.
>
> Well, if the article really is gone from the server, you might as
> well mark it as read.  It's not going to reappear...

Done with the following explanation.  Kai recently asked for a new
SUMMARY command that would fetch articles per the current group's
category and predicate.  If you do the original fetch session in the
group buffer, unfetchable articles are quietly marked as read.  If you
use the new command to fetch the group whose summary you are
examining, the unfetchable articles are not marked as read.

I should mark them as canceled.  Consider that part of my next
check-in...


>> I'm currently thinking that fetch session should simply display an
>> information message that it was unable to fetch all selected
>> articles.  Any comments?
>
> It's quite normal for messages to be canceled/superseded/expired, so
> you'd basically just display a message saying "I didn't fetch all the
> articles" in (just about) every fetching session.  So I don't think
> that's necessary.

Sounds good.