nnslashdot

nnslashdot

[Lars Magne Ingebrigtsen]

Wouldn't it be neat if one could read /. without having to use a web
browser?  nnslashdot could make a group out of each, er, "news" item,
and then all the discussion could be articles in the groups.  And one
could perhaps even follow up to messages in there.

I haven't looked at the /. html at all -- is there sufficient
information in it to do stuff like that?

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

Re: nnslashdot

[David Hedbor]

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

> Wouldn't it be neat if one could read /. without having to use a web
> browser?  nnslashdot could make a group out of each, er, "news" item,
> and then all the discussion could be articles in the groups.  And one
> could perhaps even follow up to messages in there.
> 
> I haven't looked at the /. html at all -- is there sufficient
> information in it to do stuff like that?

I have a program that does this, kinda. It downloads articles and
comments etc. See http://pike-community.org/sites/slashdot/

The actual parsing isn't really all that complicated. I currently
don't have posting inside the program (it spawns a web browser).

-- 
[ Below is a random fortune, which is unrelated to the above message. ]
If two wrongs don't make a right, try three.
		-- Laurence J. Peter

Re: nnslashdot

[Lars Magne Ingebrigtsen]

David Hedbor <david@hedbor.org> writes:

> I have a program that does this, kinda. It downloads articles and
> comments etc. See http://pike-community.org/sites/slashdot/

There is enough data in the topics/messages to uniquely identify them,
then?

... I've now had a look at the HTML, and it looks very
straightforward.  The articles even have IDs that are numerical.  It
should be fairly easy to create an nnslashdot backend.

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

Re: nnslashdot

[Chris Halverson]

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

> ... I've now had a look at the HTML, and it looks very
> straightforward.  The articles even have IDs that are numerical.  It
> should be fairly easy to create an nnslashdot backend.

See:

http://www.slashdot.org/code.shtml

There are a couple of ways to get the headlines at least. Very easily
parseable formats, should be pretty straightforward I would think.

Chris

-- 
Chris D. Halverson                         Complete Internet Solutions
PGP mail accepted, finger for public key   http://www.CompleteIS.com/~cdh/

Re: nnslashdot

[David Hedbor]

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

> David Hedbor <david@hedbor.org> writes:
> 
> > I have a program that does this, kinda. It downloads articles and
> > comments etc. See http://pike-community.org/sites/slashdot/
> 
> There is enough data in the topics/messages to uniquely identify them,
> then?
> 
> ... I've now had a look at the HTML, and it looks very
> straightforward.  The articles even have IDs that are numerical.  It
> should be fairly easy to create an nnslashdot backend.

Yes, as you said each comment has a unique identifier (used when
commenting messages etc). If you base it on the threaded page like I
do you even get a correctly threaded display.

-- 
[ Below is a random fortune, which is unrelated to the above message. ]
Any program which runs right is obsolete.

Re: nnslashdot

[Jon K Hellan]

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

> Wouldn't it be neat if one could read /. without having to use a web
> browser?  nnslashdot could make a group out of each, er, "news" item,
> and then all the discussion could be articles in the groups.  And one
> could perhaps even follow up to messages in there.
> 
> I haven't looked at the /. html at all -- is there sufficient
> information in it to do stuff like that?

Does nnweb work at all? I tried it - with the following results:

altavista - got an article list, but was unable to retrieve the
            articles
dejanews  - failed
reference - failed

I may have done something wrong, of course.

Jon

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Jon K Hellan <Jon.K.Hellan@item.ntnu.no> writes:

> Does nnweb work at all?

I've just gotten a patch which fixes Deja access, so nnweb in 0.94
will probably work.

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

Re: nnslashdot

[Wes Hardaker]

>>>>> On 07 Jul 1999 09:16:28 +0200, Lars Magne Ingebrigtsen <larsi@gnus.org> said:

Lars> Wouldn't it be neat if one could read /. without having to use a
Lars> web browser?  nnslashdot could make a group out of each, er,
Lars> "news" item, and then all the discussion could be articles in
Lars> the groups.  And one could perhaps even follow up to messages in
Lars> there.

1) That would rock.

2) I think it'd be better to have the parent article be the parent of
   a thread, rather than having a separate newsgroup per topic...
   Then you have to do too much subscribing and unsubscribing.
-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: nnslashdot

[David Hedbor]

Wes Hardaker <wjhardaker@ucdavis.edu> writes:

G> >>>>> On 07 Jul 1999 09:16:28 +0200, Lars Magne Ingebrigtsen <larsi@gnus.org> said:
> 
> Lars> Wouldn't it be neat if one could read /. without having to use a
> Lars> web browser?  nnslashdot could make a group out of each, er,
> Lars> "news" item, and then all the discussion could be articles in
> Lars> the groups.  And one could perhaps even follow up to messages in
> Lars> there.
> 
> 1) That would rock.
> 
> 2) I think it'd be better to have the parent article be the parent of
>    a thread, rather than having a separate newsgroup per topic...
>    Then you have to do too much subscribing and unsubscribing.

I agree. Have the headlines in the buffer. Then you can read the
comments by pressing C-g or so on the articles.

-- 
[ Below is a random fortune, which is unrelated to the above message. ]
It is not doing the thing we like to do, but liking the thing we have to do,
that makes life blessed.
		-- Goethe

Re: nnslashdot

[Kai.Grossjohann]

David Hedbor <david@hedbor.org> writes:

> Wes Hardaker <wjhardaker@ucdavis.edu> writes:
> 
> > 2) I think it'd be better to have the parent article be the parent of
> >    a thread, rather than having a separate newsgroup per topic...
> >    Then you have to do too much subscribing and unsubscribing.
> 
> I agree. Have the headlines in the buffer. Then you can read the
> comments by pressing C-g or so on the articles.

Why isn't it sufficient to just have threads?  They can be hidden and
shown, just as the users wish.

kai
-- 
Life is hard and then you die.

Re: nnslashdot

[David Hedbor]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> David Hedbor <david@hedbor.org> writes:
> 
> > Wes Hardaker <wjhardaker@ucdavis.edu> writes:
> > 
> > > 2) I think it'd be better to have the parent article be the parent of
> > >    a thread, rather than having a separate newsgroup per topic...
> > >    Then you have to do too much subscribing and unsubscribing.
> > 
> > I agree. Have the headlines in the buffer. Then you can read the
> > comments by pressing C-g or so on the articles.
> 
> Why isn't it sufficient to just have threads?  They can be hidden and
> shown, just as the users wish.

That would of course work great. I just didn't think about it.

-- 
[ Below is a random fortune, which is unrelated to the above message. ]
BASIC, n.:
	A programming language.  Related to certain social diseases in
	that those who have it will not admit it in polite company.

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Wes Hardaker <wjhardaker@ucdavis.edu> writes:

> 2) I think it'd be better to have the parent article be the parent of
>    a thread, rather than having a separate newsgroup per topic...
>    Then you have to do too much subscribing and unsubscribing.

Hm.  But then nnslashdot would have to download all the articles for
all the topics before letting you read anything.  Which might be
slightly overkill.  :-)

If each topic was a group (darn, those sparrows on the veranda are
awfully distracting), er, uhm, uhm, yes, if each topic was a group,
then `F' would subscribe to the new groups, and entering them would
give you a threaded display of the (unread) messages.  Hm.  But
there's only room for a single line in the group buffer...

Scratch that.  We have a single nnslashdot group.  When you select
that, you get to read the headlines.  And then there's a magic command
that lets you select each topic in a new group, kinda like nndoc.
nnslashtopic groups?  But the groups have to be persistent, since we
want to keep track of what articles we have read.

This is getting rather complicated.  :-)

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

Re: nnslashdot

[Ted Rathkopf]

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

> This is getting rather complicated.  :-)

That's why we pay you the big bucks, Lars.


-- 
Ted Rathkopf

Re: nnslashdot

[Jan Vroonhof]

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

> Hm.  But then nnslashdot would have to download all the articles for
> all the topics before letting you read anything.  Which might be
> slightly overkill.  :-)

Why is that needed. gnus-uu can add pseudo articles to the summary
buffer for the file. Why couldn't nnslashdot add real articles?

Jan

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Jan Vroonhof <vroonhof@math.ethz.ch> writes:

> Why is that needed. gnus-uu can add pseudo articles to the summary
> buffer for the file. Why couldn't nnslashdot add real articles?

It could, but would that be nice?  Hm.  

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

Re: nnslashdot

[Wes Hardaker]

>>>>> On 09 Jul 1999 19:23:29 +0200, Lars Magne Ingebrigtsen <larsi@gnus.org> said:

Lars> Scratch that.  We have a single nnslashdot group.  When you
Lars> select that, you get to read the headlines.  And then there's a
Lars> magic command that lets you select each topic in a new group,
Lars> kinda like nndoc.  nnslashtopic groups?  But the groups have to
Lars> be persistent, since we want to keep track of what articles we
Lars> have read.

That'd be cool, and was sort of what I was thinking...

You know what you need Lars?  New functionality!  I was thinking it
would be really cool to have newsgroups where only the headers from
the first articles of each thread were downloaded, and then when you
click on them it would start downloading the rest.  The backend I'd
really like this for is nnimap, but I doubt any backend but something
like nnslashdot could really make use of it...

Actually, I've been thinking a lot lately about how to make gnus more
asynchronous, and I think its possible to do things like dynamic
summary buffer updates while you're looking at it and its still
pulling in headers.  You'd have to keep a hash of the places to put
things is all based on the sorting preferences (which, um, you may
want to start with just one).  Then, it would have to be an external
program (rather than elisp) doing the fetching since you'd have to
have gnus operate in "do something when data comes in to a buffer"
mode, which I've never looked into much so I'm talking without
knowledge about how truly feasible it would be.

Anyway, that way entering groups would be lighting fast, but the
generation of the summary buffer would happen before your eyes but
would probably be slower (but pertier)...  Now the true feat would to
be able to read articles as things were coming in as well, which
should also be possible I think.

Dynamic group updates would be cool too...  You get a group buffer
listing and the items slowly change from * to 75, etc...

All this requires backends that use external programs I would think
though, which is not necessarily a good thing...

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: nnslashdot

[Simon Josefsson]

Wes Hardaker <wjhardaker@ucdavis.edu> writes:

> Actually, I've been thinking a lot lately about how to make gnus more
> asynchronous, and I think its possible to do things like dynamic
> summary buffer updates while you're looking at it and its still
> pulling in headers.
...
> Dynamic group updates would be cool too...  You get a group buffer
> listing and the items slowly change from * to 75, etc...

Yes, I've been thinking about this too, it would be cool. Every time I
think about things like this I tend to end up cursing emacs for not
having threads though.

> All this requires backends that use external programs I would think
> though

I don't think so, TCP connections are also asynchronous, but you could
also run some code with idle-timer to do things like this. I think
that could work quite well, and the code written to work with that
would require small changes if/when emacs get threads.

Re: nnslashdot

[Wes Hardaker]

>>>>> On 12 Jul 1999 22:13:21 +0200, Simon Josefsson <jas@pdc.kth.se> said:

>> Dynamic group updates would be cool too...  You get a group buffer
>> listing and the items slowly change from * to 75, etc...

Simon> Yes, I've been thinking about this too, it would be cool. Every
Simon> time I think about things like this I tend to end up cursing
Simon> emacs for not having threads though.

>> All this requires backends that use external programs I would think
>> though

Simon> I don't think so, TCP connections are also asynchronous, but
Simon> you could also run some code with idle-timer to do things like
Simon> this.

Cool..  So, ya gonna do it ;-) ?

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: nnslashdot

[Rob Browning]

Simon Josefsson <jas@pdc.kth.se> writes:

> I don't think so, TCP connections are also asynchronous, but you
> could also run some code with idle-timer to do things like this. I
> think that could work quite well, and the code written to work with
> that would require small changes if/when emacs get threads.

The emacs hosted on guile movement is starting to make noise again
(with elisp compatibiliy provided by a translation engine).  If that
were to happen, there'd at least be an underlying threaded base.  Note
that they understand how complicated this job is and expect it to take
on the order of years, but they do have some things working already.
There was a post to the guile list about it a little while ago.
They're looking for interested programmers.

Excerpt from the call for programmers:

Ken Raeburn <raeburn@raeburn.org> writes:

> The longer version, for those who want to help:
> 
> Actually, a fair amount of stuff is working at the moment.  It
> crashes sometimes, and there's no way I'd trust it with my email,
> but it pops up windows, it reads files and colors them in font-lock
> mode, it browses directories, it plays hanoi, it eats up virtual
> memory, lots of basic Emacs stuff like that.
> ...

FWIW

-- 
Rob Browning <rlb@cs.utexas.edu> PGP=E80E0D04F521A094 532B97F5D64E3930

Re: nnslashdot

[Kim-Minh Kaplan]

Simon Josefsson writes:

> Wes Hardaker writes:
> 
> > Actually, I've been thinking a lot lately about how to make gnus more
> > asynchronous, and I think its possible to do things like dynamic
> > summary buffer updates while you're looking at it and its still
> > pulling in headers.
....
> I don't think so, TCP connections are also asynchronous, but you could
> also run some code with idle-timer to do things like this.

Emacs supports the event driven model when it comes to TCP.  So you
would not need to use idle-timers.  W3 already uses this and picons
uses W3's asynchronous mode.  But it is quite messy when you don't
have real closures to do the work...

Kim-Minh.

Re: nnslashdot

[Simon Josefsson]

Kim-Minh Kaplan <kmkaplan@vocatex.fr> writes:

> > > Actually, I've been thinking a lot lately about how to make gnus more
> > > asynchronous, and I think its possible to do things like dynamic
> > > summary buffer updates while you're looking at it and its still
> > > pulling in headers.
> ....
> > I don't think so, TCP connections are also asynchronous, but you could
> > also run some code with idle-timer to do things like this.
> 
> Emacs supports the event driven model when it comes to TCP.  So you
> would not need to use idle-timers.

Well, for nnml and other backends that doesn't use external processes
or TCP I would.

Re: nnslashdot

[Carl R. Witty]

Kim-Minh Kaplan <kmkaplan@vocatex.fr> writes:

> Emacs supports the event driven model when it comes to TCP.  So you
> would not need to use idle-timers.  W3 already uses this and picons
> uses W3's asynchronous mode.  But it is quite messy when you don't
> have real closures to do the work...

Doesn't cl.el have some code to handle lexical scoping and closures?

Carl Witty
cwitty@newtonlabs.com

Re: nnslashdot

[Kim-Minh Kaplan]

Carl R. Witty writes:

> Doesn't cl.el have some code to handle lexical scoping and closures?

Indeed.  I haven't tried it myself but it sounds good !

Kim-Minh.

nnslashdot

[Lars Magne Ingebrigtsen]

I wanted to do some real programming, so I wrote nnslashdot.

It's pretty simple now -- if you put `(nnslashdot "")' in your
gnus-secondary-select-methdods, you'll get a new group for every new
article that appears on Slashdot.  When you enter these groups, you
can read the responses.

And that's pretty much it for now.  The version in CVS doesn't even
download the actual article, but that should be simple to fix, but I
need to sleep now, I think.  Or at least not type for a few minutes.

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

Re: nnslashdot

[Lars Magne Ingebrigtsen]

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

> Or at least not type for a few minutes.

The group names look awful now, though.  But the title of the topics
could be put in the group description, but then it wouldn't be shown
unless you have the %D spec in gnus-group-line-format.  Hm...

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

Re: nnslashdot

[Dan Christensen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> I wanted to do some real programming, so I wrote nnslashdot.
> 
> It's pretty simple now -- if you put `(nnslashdot "")' in your
> gnus-secondary-select-methdods, you'll get a new group for every new
> article that appears on Slashdot.

I'm not getting any nnslashdot groups.  `F' doesn't help.  The server
is open, but when I browse the server I get an empty buffer.  Any
idea what could be going wrong?  

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: nnslashdot

[Shenghuo ZHU]

>>>>> "Dan" == Dan Christensen <jdc@jhu.edu> writes:

Dan> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>> I wanted to do some real programming, so I wrote nnslashdot.
>> 
>> It's pretty simple now -- if you put `(nnslashdot "")' in your
>> gnus-secondary-select-methdods, you'll get a new group for every new
>> article that appears on Slashdot.

Dan> I'm not getting any nnslashdot groups.  `F' doesn't help.  The
Dan> server is open, but when I browse the server I get an empty
Dan> buffer.  Any idea what could be going wrong?

Updating to the latest CVS version would help.

-- 
Shenghuo ZHU

Re: nnslashdot

[Dan Christensen]

Shenghuo ZHU <zsh@cs.rochester.edu> writes:
> >>>>> "Dan" == Dan Christensen <jdc@jhu.edu> writes:
> 
> Dan> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> >> I wanted to do some real programming, so I wrote nnslashdot.
> >> 
> >> It's pretty simple now -- if you put `(nnslashdot "")' in your
> >> gnus-secondary-select-methdods, you'll get a new group for every new
> >> article that appears on Slashdot.
> 
> Dan> I'm not getting any nnslashdot groups.  `F' doesn't help.  The
> Dan> server is open, but when I browse the server I get an empty
> Dan> buffer.  Any idea what could be going wrong?
> 
> Updating to the latest CVS version would help.

No difference.  Any other ideas?

Also, gnus freezes if I start it when not connected to the net.
Is there any way to avoid this, or at least to catch `C-g' and
continue with the other servers?

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> Also, gnus freezes if I start it when not connected to the net.

And slashdot is sometimes down, as well.

> Is there any way to avoid this, or at least to catch `C-g' and
> continue with the other servers?

The version now in CVS now catches C-g's in `gnus-read-active-file'. 

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

Re: nnslashdot

[Matt Pharr]

Dan Christensen <jdc@jhu.edu> writes:
> Shenghuo ZHU <zsh@cs.rochester.edu> writes:
> > >>>>> "Dan" == Dan Christensen <jdc@jhu.edu> writes:
> > 
> > Dan> I'm not getting any nnslashdot groups.  `F' doesn't help.  The
> > Dan> server is open, but when I browse the server I get an empty
> > Dan> buffer.  Any idea what could be going wrong?
> > 
> > Updating to the latest CVS version would help.
> 
> No difference.  Any other ideas?

I've got the latest and it doesn't do anything for me, either.  I have
added it to my secondary select methods.

I thought I was on to something when I changed my gnus-check-new-newsgroups
from nil to (setq gnus-check-new-newsgroups '((nnslashdot ""))), but that
didn't help any either.  I've hit "g", I've hit "F", and nothing nothing
nothing. :-(.

It does indicate that it's contacting slashdot.org, though my 
" *nnslashdot *" buffer is empty.  I've got the following in " *Message-Log*"

Reading active file via nnslashdot...
Contacting slashdot.org:80
Retrieval complete.
Reading active file via nnslashdot...
Reading active file via nnslashdot...done

Anyone?

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Dan Christensen]

Matt Pharr <mmp@graphics.stanford.edu> writes:
> I've got the latest and it doesn't do anything for me, either.  I have
> added it to my secondary select methods.
> 
> It does indicate that it's contacting slashdot.org, though my 
> " *nnslashdot *" buffer is empty.  I've got the following in " *Message-Log*"
> 
> Reading active file via nnslashdot...
> Contacting slashdot.org:80
> Retrieval complete.
> Reading active file via nnslashdot...
> Reading active file via nnslashdot...done

Essentially the same as me.  Here are relevant lines from my *Messages*:

  Reading active file via nnslashdot...
  Opening nnslashdot server...
  Contacting slashdot.org:80
  Request sent, waiting for response...
  Retrieval complete.
  
  Reading active file via nnslashdot...
  Reading active file via nnslashdot...done
  
  Connecting to ...
  Reading active file...
  Contacting slashdot.org:80
  Request sent, waiting for response...
  Retrieval complete.

Is that "Connecting to ..." (with no hostname) a problem?

> To followup to myself, I do have a "slashdot/groups" file in my mail
> directory, which is full of relevant looking stuff.

Hmm, I have that file too, but all it contains is "nil".

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> Essentially the same as me.  Here are relevant lines from my *Messages*:

After updating to the latest CVS, what's in your " *nntpd*" buffer
after saying `B nnslashdot RET RET'?

> Is that "Connecting to ..." (with no hostname) a problem?

Nope.

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

Re: nnslashdot

[Matt Pharr]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Dan Christensen <jdc@jhu.edu> writes:
> 
> > Essentially the same as me.  Here are relevant lines from my *Messages*:
> 
> After updating to the latest CVS, what's in your " *nntpd*" buffer
> after saying `B nnslashdot RET RET'?

Now this is getting weird.  For me, browsing it worked great.  Whoo-hoo!
Then I quit from the foreign server browser buffer, hit G in my normal
group buffer, and got:

Signaling: (error "Selecting deleted or non-existent buffer")
  gnus-read-active-file()
  gnus-group-get-new-news(nil)
  call-interactively(gnus-group-get-new-news)

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Matt Pharr]

[-- Attachment #1: Type: text/plain, Size: 123 bytes --]


To followup to myself, I do have a "slashdot/groups" file in my mail
directory, which is full of relevant looking stuff.


[-- Attachment #2: slashdot/groups file --]
[-- Type: application/octet-stream, Size: 2948 bytes --]

(("Interview: Grill John Vranesevich of AntiOnline (99/11/08/1021204)" 239 "99/11/08/1021204") ("Mainstream Media on Slashdot and Microsoft (99/11/08/1226255)" 214 "99/11/08/1226255") ("Candidates for 1999 GNU Free Software Award (99/11/08/116242)" 116 "99/11/08/116242") ("The Post-Microsoft Era (99/11/07/1750211)" 399 "99/11/07/1750211") ("IBM,Sun and Others to Help Apache Support XML (99/11/08/1133211)" 84 "99/11/08/1133211") ("User Friendly:  The Book (99/11/02/1115243)" 22 "99/11/02/1115243") ("Amazon.com switches to Apache (99/11/05/0947259)" 82 "99/11/05/0947259") ("Legal Actions Against Linux-DVD authors (99/11/05/1158259)" 344 "99/11/05/1158259") ("The Battle That Could Lose Us The War (99/11/05/144251)" 549 "99/11/05/144251") ("Linux on a Magazine Cover? (99/11/03/2329249)" 234 "99/11/03/2329249") ("SGI announces Linux Kernel Crash Dumps (LKCD) (99/11/05/1521233)" 208 "99/11/05/1521233") ("Microsoft == Monopoly says Judge (99/11/05/1841214)" 927 "99/11/05/1841214") (
 "!
!
!
Slashdot's \"Instant\" Legal Analysis of the MS Ruling (99/11/05/2016212)" 457 "99/11/05/2016212") ("Geeks In Space: Easy Listening (99/11/05/2129207)" 48 "99/11/05/2129207") ("The Do-It-All Remote? (99/11/03/2245248)" 236 "99/11/03/2245248") ("Vote in a CNN Poll on the DOJ MS Ruling (99/11/05/2312254)" 214 "99/11/05/2312254") ("Rick Moen on LinuxOne's IPO (99/11/06/1016209)" 120 "99/11/06/1016209") ("Checkpoint Porting Firewall-1 to Linux (99/11/06/1018228)" 132 "99/11/06/1018228") ("Microsoft Adresses World (99/11/06/140237)" 604 "99/11/06/140237") ("RealPlayer Uploads Your ID Too (99/11/06/1347228)" 164 "99/11/06/1347228") ("Communicator Is Losing The War..... (99/11/06/1747232)" 555 "99/11/06/1747232") ("First Journaling FS for Linux (99/11/06/1758233)" 287 "99/11/06/1758233") ("Online Romance - For Good or Evil? (99/11/03/2321243)" 505 "99/11/03/2321243") ("New Genetic Information Web Portal (99/11/07/0833206)" 69 "99/11/07/0833206") ("Lightning On Demand (99/11/07/08402
 0!
!
!
0)" 161 "99/11/07/0840200") ("The JFC Swing Tutorial (99/11/02/1052223)" 93 "99/11/02/1052223") ("Digital Television Transmission Standards (99/11/07/0845212)" 228 "99/11/07/0845212") ("Lotus Domino to ship RSN (99/11/07/127219)" 81 "99/11/07/127219") ("Debian Freeze Rescheduled (99/11/07/1611205)" 231 "99/11/07/1611205") ("David Bowie talks about Technology and Music (99/11/07/1913240)" 149 "99/11/07/1913240") ("Finding an Intellectual Property Patent Lawyer? (99/11/03/2141246)" 73 "99/11/03/2141246") ("Applications Service Providers May Change Your Life (99/11/08/080201)" 78 "99/11/08/080201") ("Does ATi Have a GeForce 256 Killer? (99/11/08/082223)" 94 "99/11/08/082223") ("U.S. Military Grapples With Cyber Warfare Rules (99/11/08/0812254)" 140 "99/11/08/0812254") ("TRUSTe Decides Its Own Fate Today (99/11/05/1021214)" 108 "99/11/05/1021214") ("Coming to a Desktop near you: Tempest Capabilities (99/11/08/093250)" 98 "99/11/08/093250"))

[-- Attachment #3: Type: text/plain, Size: 122 bytes --]


-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> To followup to myself, I do have a "slashdot/groups" file in my mail
> directory, which is full of relevant looking stuff.

If you go to the server buffer and hit `RET' on the nnslashdot server, 
but you get anything?  What's in your " *nntpd*" buffer after doing
that? 

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

Re: nnslashdot

[Matt Pharr]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Matt Pharr <mmp@graphics.stanford.edu> writes:
> 
> > To followup to myself, I do have a "slashdot/groups" file in my mail
> > directory, which is full of relevant looking stuff.
> 
> If you go to the server buffer and hit `RET' on the nnslashdot server, 
> but you get anything?  What's in your " *nntpd*" buffer after doing
> that? 

Yep, that works great, like the browsing stuff.  Same bug out after I exit
though, when I hit G in the group buffer.

Signaling: (error "Selecting deleted or non-existent buffer")
  gnus-read-active-file()
  gnus-group-get-new-news(nil)
  call-interactively(gnus-group-get-new-news)

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> Yep, that works great, like the browsing stuff.  Same bug out after I exit
> though, when I hit G in the group buffer.
> 
> Signaling: (error "Selecting deleted or non-existent buffer")
>   gnus-read-active-file()

Hm.  I've found a bug there.  Check out the CVS, and try again.  :-)

If you get the same bugout again, could you `(load "gnus-start.el")'
to load an uncompiled Gnus?  That will give me a much more detailed
backtrace.

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

Re: nnslashdot

[Matt Pharr]

Some more information.  It turns out that all of the slashdot articles were 
zombified.  Preivously I had gnus-subscribe-newsgroup-method as the default 
gnus-subscribe-zombies; the only other subscribing variable I had set was

(setq gnus-subscribe-hierararchical-interactive t)

I was able to do A z in the group buffer and subscribe to those guys and
successfully read slashdot stuff--wheee!

I've now changed gnus-subscribe-newsgroup-method to
gnus-subscribe-randomly; I suspect that that will fix things and will
report back once slashdot has posted something new and I see if that's the
case.

Now my next question: how to reliably make the original slashdot article be
the first thing in the summary buffer?

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Shenghuo ZHU]

>>>>> "Matt" == Matt Pharr <mmp@graphics.stanford.edu> writes:

Matt> Some more information.  It turns out that all of the slashdot
Matt> articles were zombified.  Preivously I had
Matt> gnus-subscribe-newsgroup-method as the default
Matt> gnus-subscribe-zombies; the only other subscribing variable I
Matt> had set was

Matt> (setq gnus-subscribe-hierararchical-interactive t)

Matt> I was able to do A z in the group buffer and subscribe to those
Matt> guys and successfully read slashdot stuff--wheee!

Matt> I've now changed gnus-subscribe-newsgroup-method to
Matt> gnus-subscribe-randomly; I suspect that that will fix things and
Matt> will report back once slashdot has posted something new and I
Matt> see if that's the case.

Lars did a trick `gnus-subscribe-topics' here. 

I use it as following,
1. (setq gnus-subscribe-newsgroup-method 'gnus-subscribe-topics)
2. (add-hook 'gnus-group-mode-hook 'gnus-topic-mode) ;; be sure using topics
3. `T n' to create a topic named "slashdot"
4. `G p' on the topic
5. Put ((subscribe . "nnslashdot")) as topic parameters

For details, check the info node.

Matt> Now my next question: how to reliably make the original slashdot
Matt> article be the first thing in the summary buffer?

The original article should be article number 1 in the group. If
you've read the article, it may not be the first one in the summary
buffer.

-- 
Shenghuo

Re: nnslashdot

[Matt Pharr]

Shenghuo ZHU <zsh@cs.rochester.edu> writes:
> 
> Lars did a trick `gnus-subscribe-topics' here. 
> 
> I use it as following,
> 1. (setq gnus-subscribe-newsgroup-method 'gnus-subscribe-topics)
> 2. (add-hook 'gnus-group-mode-hook 'gnus-topic-mode) ;; be sure using topics
> 3. `T n' to create a topic named "slashdot"
> 4. `G p' on the topic
> 5. Put ((subscribe . "nnslashdot")) as topic parameters
> 
> For details, check the info node.

Yes, that did it--thanks!  Two questions and another bug now.

Question: I'd like to combine the behavior of gnus-subscribe-randomly and
gnus-subscribe-topics.  That is, now that all the slashdot groups are
appearing in my slashdot topic, I'd like to have the new ones appear at the 
top rather than the bottom as they do now.  Anyone have any ideas?

Question: Now with the addition of the slashdot backend, I have some
backends that I want to check every once in a while and some that I want to 
check every single time I hit 'g'.  Is group levels the way to do this?
Are there other options--what if I wanted to check slashdot no more than
once per 6 hours, say?

Bug: the first time I enter a slashdot group, all is well--I can read
articles, read comments, etc.  But then later on, when gnus reports new
messages, if I try to enter the group it immediately says "No unread news"
and reports no new messages.  It does this quite quickly, so I suspect that 
it isn't even contacting slashdot.org.

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Kai Großjohann]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> Question: Now with the addition of the slashdot backend, I have some
> backends that I want to check every once in a while and some that I want to 
> check every single time I hit 'g'.  Is group levels the way to do this?
> Are there other options--what if I wanted to check slashdot no more than
> once per 6 hours, say?

Yes, group levels are the way to go.  Set gnus-activate-level to 4,
say, then type `g' to update the important groups and `5 g' to update
all groups.  (If your important groups are at level less equal 4 and
the others at level 5.)

Dunno about the `no more often than once every six hours' part,
though.  You could add a demon which did it automatically every six
hours, but that's not what you're asking for.  Or you could train
yourself not to hit `5 g' more often than once per six hours...

kai
-- 
This gubblick contains many nonsklarkish English flutzpahs,
but the overall pluggandisp can be glorked from context. -- David Moser

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> Question: I'd like to combine the behavior of gnus-subscribe-randomly and
> gnus-subscribe-topics.  That is, now that all the slashdot groups are
> appearing in my slashdot topic, I'd like to have the new ones appear at the 
> top rather than the bottom as they do now.  Anyone have any ideas?

Well, you can just alter `gnus-subscribe-topics'...

> Question: Now with the addition of the slashdot backend, I have some
> backends that I want to check every once in a while and some that I want to 
> check every single time I hit 'g'.  Is group levels the way to do this?

Yup.

> Bug: the first time I enter a slashdot group, all is well--I can read
> articles, read comments, etc.  But then later on, when gnus reports new
> messages, if I try to enter the group it immediately says "No unread news"
> and reports no new messages.  It does this quite quickly, so I suspect that 
> it isn't even contacting slashdot.org.

Have you upgraded to the latest CVS?

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

Re: nnslashdot

[Matt Pharr]

> > Bug: the first time I enter a slashdot group, all is well--I can read
> > articles, read comments, etc.  But then later on, when gnus reports new
> > messages, if I try to enter the group it immediately says "No unread news"
> > and reports no new messages.  It does this quite quickly, so I suspect that 
> > it isn't even contacting slashdot.org.
> 
> Have you upgraded to the latest CVS?

Yep, just a few minues ago.  Now I'm getting the following on all slashdot
articles.

Signaling: (args-out-of-range "" 0 3)
  nnslashdot-date-to-date(" Thursday November 11, @11:39AM")
  nnslashdot-retrieve-headers((1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 ...) "Samba 2.06 Released (99/11/11/1220207)" "" some)
  gnus-retrieve-headers((1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 ...) "nnslashdot:Samba 2.06 Released (99/11/11/1220207)" some)
  gnus-cache-retrieve-headers((1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 ...) "nnslashdot:Samba 2.06 Released (99/11/11/1220207)" some)
  gnus-retrieve-headers((1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 ...) "nnslashdot:Samba 2.06 Released (99/11/11/1220207)" some)
  gnus-select-newsgroup("nnslashdot:Samba 2.06 Released (99/11/11/1220207)" nil nil)
  gnus-summary-read-group-1("nnslashdot:Samba 2.06 Released (99/11/11/1220207)" nil nil nil nil nil)
  gnus-summary-read-group("nnslashdot:Samba 2.06 Released (99/11/11/1220207)" nil nil nil nil nil nil)
  gnus-group-read-group(nil nil nil)
  gnus-topic-read-group(nil)
  call-interactively(gnus-topic-read-group)

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> Yep, just a few minues ago.  Now I'm getting the following on all slashdot
> articles.
> 
> Signaling: (args-out-of-range "" 0 3)
>   nnslashdot-date-to-date(" Thursday November 11, @11:39AM")

This is now fixed in CVS.

On the "missing theading due to lacks of cid's" problem -- Malda
didn't want to fix that.  Because that would hurt the banner ads.  So
I offered to include the banner ads in the articles, but didn't hear
back from him after that.  (Does XEmacs do animated gifs, by the way?
:-)

So, there's a new variable in nnslashdot -- nnslashdot-threaded.  If
it's t, nnslashdot will retrieve absolutely all comments each time you 
enter the group, so it'll be slow, but pretty.  

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

Re: nnslashdot

[Norbert Koch]

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

> (Does XEmacs do animated gifs, by the way? :-)

Andy Piper has brought something up in 21.2 ... I've never tested it
though. 

norbert.

Re: nnslashdot

[Hrvoje Niksic]

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

> On the "missing theading due to lacks of cid's" problem -- Malda
> didn't want to fix that.  Because that would hurt the banner ads.

He didn't want to fix a bug for fear of hurting the banner ads?

Tsk tsk tsk.

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@iskon.hr> writes:

> He didn't want to fix a bug for fear of hurting the banner ads?

Well, to be more precise -- he kinda didn't like the non-banner bit
about the, er, project, and apart from that, he didn't think it was a
bug since when you're displaying articles in a non-threaded manner,
they have no parents.

\/\/hatever.

> Tsk tsk tsk.

Well.  I'm not surprised.

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

Re: nnslashdot

[Steinar Bang]

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

> Hrvoje Niksic <hniksic@iskon.hr> writes:
>> He didn't want to fix a bug for fear of hurting the banner ads?

> Well, to be more precise -- he kinda didn't like the non-banner bit
> about the, er, project, and apart from that,

Hm... maybe I should post a feature about pgnus nnslashdot to slashdot 
and see if it is posted...? ;-)

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Steinar Bang <sb@metis.no> writes:

> Hm... maybe I should post a feature about pgnus nnslashdot to slashdot 
> and see if it is posted...? ;-)

:-)

We should perhaps wait until pgnus 0.99 has been released, though.

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

Re: nnslashdot

[Jack Twilley]

>>>>> "ZSH" == Shenghuo ZHU <zsh@cs.rochester.edu> writes:

ZSH> Lars did a trick `gnus-subscribe-topics' here.

[... cookbook instructions on setting up gnus-subscribe-topics ...]

ZSH> For details, check the info node.

Here's the relevant part of the info node:

`gnus-subscribe-topics'
     Put the groups into the topic that has a matching `subscribe' topic
     parameter (*note Topic Parameters::.).  For instance, a `subscribe'
     topic parameter that looks like

          "nnslashdot"

     will mean that all groups that match that regex will be subscribed
     under that topic.

I tried this, and when a new group came up in nnslashdot as seen on
the server menu, it didn't come up subscribed in the Slashdot topic.
Why? 

Jack.
-- 
Jack Twilley
jmt at nycap dot rr dot com
http colon slash slash jmt dot dhs dot org slash tilde jmt slash

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Jack Twilley <jmt+usenet@nycap.rr.com> writes:

> I tried this, and when a new group came up in nnslashdot as seen on
> the server menu, it didn't come up subscribed in the Slashdot topic.

Where did it end up, then?

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

Re: nnslashdot

[Jack Twilley]

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

Jack> I tried this, and when a new group came up in nnslashdot as seen
Jack> on the server menu, it didn't come up subscribed in the Slashdot
Jack> topic.

Lars> Where did it end up, then?

It, umm, didn't end up anywhere.  I was in the Slashdot topic when I
did 'A z', and it came up.  Is this the default behavior?  Is it
possible for me to fix it on my end so mine just appear in the right
place?  

Jack.
-- 
Jack Twilley
jmt at nycap dot rr dot com
http colon slash slash jmt dot dhs dot org slash tilde jmt slash

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Jack Twilley <jmt+usenet@nycap.rr.com> writes:

> It, umm, didn't end up anywhere.  I was in the Slashdot topic when I
> did 'A z', and it came up. 

If you have `gnus-subscribe-topics' as your subscription method, then
it should work.  It doesn't?

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

Re: nnslashdot

[Kevin Ryde]

Dan Christensen <jdc@jhu.edu> writes:
> 
> Also, gnus freezes if I start it when not connected to the net.
> Is there any way to avoid this, or at least to catch `C-g' and
> continue with the other servers?

If that's with FSF emacs it could be open-network-stream retrying
gethostbyname forever, see src/process.c.  I posted bug-gnu-emacs
about this not so long ago, although probably it would've been known
already.  xemacs limits its retries, and has done for a while.

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> I'm not getting any nnslashdot groups.  `F' doesn't help.  The server
> is open, but when I browse the server I get an empty buffer.  Any
> idea what could be going wrong?  

Hm...

It could be a case of `case-fold-search' sloppiness.  If you have
case-fold-search set to nil, I've now fixed this in CVS.

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

Re: nnslashdot

[Matt Pharr]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Dan Christensen <jdc@jhu.edu> writes:
> 
> > I'm not getting any nnslashdot groups.  `F' doesn't help.  The server
> > is open, but when I browse the server I get an empty buffer.  Any
> > idea what could be going wrong?  
> 
> Hm...
> 
> It could be a case of `case-fold-search' sloppiness.  If you have
> case-fold-search set to nil, I've now fixed this in CVS.

I just updated and re-made my elc files and still no luck.  Anything else
to look at?

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> > It could be a case of `case-fold-search' sloppiness.  If you have
> > case-fold-search set to nil, I've now fixed this in CVS.
> 
> I just updated and re-made my elc files and still no luck.  Anything else
> to look at?

Is your case-fold-search nil?

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

Re: nnslashdot

[Matt Pharr]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Matt Pharr <mmp@graphics.stanford.edu> writes:
> > > It could be a case of `case-fold-search' sloppiness.  If you have
> > > case-fold-search set to nil, I've now fixed this in CVS.
> > 
> > I just updated and re-made my elc files and still no luck.  Anything else
> > to look at?
> 
> Is your case-fold-search nil?

Nope; it's t.

-matt
-- 
Matt Pharr                                   mmp@graphics.stanford.edu
<URL:http://graphics.stanford.edu/~mmp>

Re: nnslashdot

[Dan Christensen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Dan Christensen <jdc@jhu.edu> writes:
> 
> > I'm not getting any nnslashdot groups.  `F' doesn't help.  The server
> > is open, but when I browse the server I get an empty buffer.  Any
> > idea what could be going wrong?  
> 
> It could be a case of `case-fold-search' sloppiness.  If you have
> case-fold-search set to nil, I've now fixed this in CVS.

This fixed part of the problem.  Now I get nnslashdot groups created
automatically, and I can enter them and see summary lines.  However,
if I select an article I get things like:

Search failed: "^<p>.*A href=http://slashdot.org/article.pl"
Search failed: "<a name=\"4\">"
Search failed: "<a name=\"31\">"
Search failed: "<a name=\"6\">"

The " *nnslashdot*" buffer contains capital A's in places of the small
a's.

case-fold-search's value is 
nil
Local in buffer  *nnslashdot *; global value is 
nil

I just updated to the latest cvs version a couple of minutes ago.

Dan

Re: nnslashdot

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> The " *nnslashdot*" buffer contains capital A's in places of the small
> a's.

Ok; this should be fixed in CVS now.

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

Re: nnslashdot

[Dan Christensen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Dan Christensen <jdc@jhu.edu> writes:
> 
> > The " *nnslashdot*" buffer contains capital A's in places of the small
> > a's.
> 
> Ok; this should be fixed in CVS now.

Yup, thanks.

nnslashdot

[Lars Magne Ingebrigtsen]

Hm.  nnimap has the same problem, I think -- there are some natural
group names that can't be done under the current Gnus group name
regime, most notably group names with spaces in them.

But if we just allow *-request-list to answer with lines that look
like

"This is a really weird group name" 9 1 m

then everything will work, as if by magic.  So I've done that, and
made nnslashdot use then (uniquified) article title as the group name, 
which means that the group buffer looks kinda like:

     81: Amazon.com switches to Apache (99/11/05/0947259)
    141: Caldera vs. Microsoft Goes to Jury Trial (99/11/05/0814243)
    120: Checkpoint Porting Firewall-1 to Linux (99/11/06/1018228)
     79: Cobalt IPO Opens...High (99/11/05/1113248)

It's in CVS.

The more serious problem with nnslashdot is that when requesting the
comments strictly sorted by age (which is nice, since that means not
having to fetch the entire list of comments when entering the group),
all the threading info is gone.  (That is, the "Parent" link is
"cid"-less.)  Looks kinda like a bug in Slashdot to me.  

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

Re: nnslashdot

[Shenghuo ZHU]

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

LMI> Hm.  nnimap has the same problem, I think -- there are some natural
LMI> group names that can't be done under the current Gnus group name
LMI> regime, most notably group names with spaces in them.

LMI> But if we just allow *-request-list to answer with lines that look
LMI> like

LMI> "This is a really weird group name" 9 1 m

LMI> then everything will work, as if by magic.  So I've done that, and
LMI> made nnslashdot use then (uniquified) article title as the group name, 
LMI> which means that the group buffer looks kinda like:

LMI>      81: Amazon.com switches to Apache (99/11/05/0947259)
LMI>     141: Caldera vs. Microsoft Goes to Jury Trial (99/11/05/0814243)
LMI>     120: Checkpoint Porting Firewall-1 to Linux (99/11/06/1018228)
LMI>      79: Cobalt IPO Opens...High (99/11/05/1113248)

LMI> It's in CVS.


[...]

Now it is supported by agent, cache, browser-foreign-server. It's also
in CVS.

-- 
Shenghuo ZHU

nnslashdot

[Lars Magne Ingebrigtsen]

Argh.  Slashdot has kinda completely changed quite a lot these past
days.  Would anyone take a whack at updating nnslashdot to use the new 
format?

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

Re: nnslashdot

[BrYan P. Johnson]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Argh.  Slashdot has kinda completely changed quite a lot these past
> days.  Would anyone take a whack at updating nnslashdot to use the new 
> format?

How about an nnrdf backend. If you look at
http://www.slashdot.org/slashdot.rdf that format doesn't change nearly
as often as the html.

It's xml, and quite a few news log sites use it, hence the suggestion
for a more generic backend.

BrYan
-- 
Old people are not food.

Re: nnslashdot

[BrYan P. Johnson]

Silly me. I just realized that while this may help with generating the
summary, it won't help with displaying the article.

Feel free to ignore me now. :-)

BrYan

BrYan P. Johnson <bilko@onebabyzebra.com> writes:
> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> > Argh.  Slashdot has kinda completely changed quite a lot these past
> > days.  Would anyone take a whack at updating nnslashdot to use the new 
> > format?
> 
> How about an nnrdf backend. If you look at
> http://www.slashdot.org/slashdot.rdf that format doesn't change nearly
> as often as the html.
> 
> It's xml, and quite a few news log sites use it, hence the suggestion
> for a more generic backend.

Re: nnslashdot

[dme]

"bpj" == BrYan P Johnson <bilko@onebabyzebra.com> writes:
bpj> How about an nnrdf backend. If you look at
bpj> http://www.slashdot.org/slashdot.rdf that format doesn't change
bpj> nearly as often as the html.

That would be fine, except that it would still be necessary to
retrieve the actual articles (and comments, if that's interesting).
They don't appear to be available other than through parsing the
HTML.

I have a small Perl program (which also uses wget and lynx) that grabs
the .rdf file, gets the articles and injects them into my local news
server, so the change to the HTML format bit me as well :(

dme.

Re: nnslashdot

[Kai Großjohann]

I wonder if somebody tried to talk to them to get a different
interface than HTTP.  I mean, they're not advertising, so there's no
need for banner ads, and thus they could just as easily use NNTP, say.
kai
-- 
Beware of flying birch trees.

Re: nnslashdot

[Alan Shutko]

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

> I wonder if somebody tried to talk to them to get a different
> interface than HTTP.  I mean, they're not advertising,
[...]

Yes, they are advertising.  There are banners at the top of every
page, and they've been there for years.  We are still talking about
slashdot, right?

(Actually, right now, they're not advertising or doing anything else,
because they look down, from here at least.)

-- 
Alan Shutko <ats@acm.org> - In a variety of flavors!
182 days, 23 hours, 9 minutes, 32 seconds till we run away.
Siskel and Ebert gave this tagline two thumbs up.

Re: nnslashdot

[Kai Großjohann]

Alan Shutko <ats@acm.org> writes:

> Yes, they are advertising.  There are banners at the top of every
> page, and they've been there for years.  We are still talking about
> slashdot, right?

Argh.  Why did I confuse them with Advogato?

*hangs head in shame*

Sorry.

kai
-- 
Beware of flying birch trees.