gmane.network.leafnode

checkgroups usage

Thufir — 2013-05-19T21:18:46

I've removed the gmane group from the config, but gmane keeps showing up 
in interesting groups.  I think this can be fixed with checkgroups?

Am I not doing this correctly?


root< at >dur:~# 
root< at >dur:~# head -n 3 /var/spool/news/leaf.node/groupinfo 
0 1 2 1368645998 0 0 
0.akita-inu 1 2 1368645998 0.akita-inu
0.alaskan-malamute 1 2 1368645998 0.alaskan-malamute
root< at >dur:~# 
root< at >dur:~# checkgroups /var/spool/news/leaf.node/groupinfo 
cannot read configuration.
root< at >dur:~# 



thanks,

Thufir

Running trn with Leafnode?

Hart Larry — 2013-04-06T01:14:47

Well, in asking this in blind Linux lists-and-elsewhere, I installed Leafnode 
here in Debian, but my binary news-reader of choice is trn.  Cannn some1 please 
inform on steps I can have these 2 work together?  From what I understand 
Leafnode can open a portion of a group, which would be great as trn crashes on 
larger groups-and-increasing article numbers.
As of now I am running trn with an rn script, like this:
#!/bin/sh
# run trn
/usr/bin/trn -J5 -G -x0sml "-X1XD" -Ots -e -M +m -t -h +hfrom +Hnewsgr "$< at >"
# move any clarinet.* groups to .newsrc.clari
#mv-clari
# re-sort the .newsrc file
sortrc
Back again live, I already edited a config with giganews  related items, but 
have no idea if any changes need to go in my .newsrc or .trnrc, ETC.
Thanks so very much in advance
Hart

Leafnode-List administrivia "Mail forwarding loop"

Matthias Andree — 2012-11-24T10:23:19

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Greetings,

Some of you may recently have received "Undelivered Mail returned to
send" with a "Mail forwarding loop" complaint from
MAILER-DAEMON-DcxqERLcJkSjd9KveT9UGqzufkE7zoXZyCaAmVia5Yk< at >public.gmane.org recently.

If you did not, skip this message.

If you did: THIS IS NOT YOUR FAULT, and you NEED NOT RESEND your
messages -- your list posts have been processed properly.  The only
thing is that one of the subscribers had - through misconfiguration or
by accident or other misfortune - caused /your/ messages to be re-sent
to the list, and this caused half a dozen complaints for me.

I have set the "nomail" flag for the offending list subscriber address,
and sent him a separate heads-up message in what I believe to be his/her
native language, explaining the problem and asking him/her to resolve it.

Again, except for the list subscriber who was notified off-list,
YOU NEED NOT take any action.  Your subscription to the list, and the
processing of your messages sent to the list were unaffected.

Sorry again and best regards
Matthias
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAlCwoBIACgkQvmGDOQUufZW87wCgpFmGsqAjH9oh2MCwGuYAWi7L
CxwAn2FHi8ESY8iVL9qHt6LVsi87m2UE
=Fp+j
-----END PGP SIGNATURE-----

illegal headers

thufir — 2012-11-07T06:28:17

when leafnode finds illegal headers:


root< at >dur:~# article /var/spool/news/comp/os/linux/advocacy/6401 contained 
illegal headers: non-printable characters in headers (relaxed check 
allows for iso-8859*)



does it remove that article from the spool?



thanks,

Thufir

leafnode-2.0.0.alpha20121101a snapshot available

Matthias Andree — 2012-11-01T14:23:58

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Leafnode 2.0.0.alpha20121101a is available from
http://home.pages.de/~mandree/leafnode/beta/

NEWS:
KNOWN BUGS:
* The code sometimes uses timeout_client where it should use a server timeout.
|
2.0.0.alpha20121101a: Changes since 20110807a:
+ Bugfix: if reading descriptions of new newsgroups fails, don't update
  "active last updated" time stamp for that server. To solve a problem
  reported by Antoine Levitt.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.18 (GNU/Linux)

iEYEARECAAYFAlCShf4ACgkQvmGDOQUufZXcawCgrDPyhzAgh9xHni0GEhW6YQv0
09wAniM/9vjpklgotUqMqYUC0+WzLIv5
=8wRj
-----END PGP SIGNATURE-----

leafnode-2.0.0.alpha20121101a snapshot available

Matthias Andree — 2012-11-01T14:18:44

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Leafnode 2.0.0.alpha20121101a is available from
http://home.pages.de/~mandree/leafnode/beta/

NEWS:
KNOWN BUGS:
* The code sometimes uses timeout_client where it should use a server timeout.
|
2.0.0.alpha20121101a: Changes since 20110807a:
+ Bugfix: if reading descriptions of new newsgroups fails, don't update
  "active last updated" time stamp for that server. To solve a problem
  reported by Antoine Levitt.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAlCShMQACgkQvmGDOQUufZUBgwCdF/e3X5lB9LdDuLv2jBxq8AM6
AGYAn2vBo4INt0aHZjpFfTPWUfyYHWh9
=omz5
-----END PGP SIGNATURE-----

HEADS UP Clemens Fischer, lua authentication bugs, and your mail is non-functional

Matthias Andree — 2012-11-01T13:54:49

Clemens,

sorry to contact you this way, and I hope you are seeing this through
GMANE - your dnsalias.org address is dysfunctional, so my mail bounced.

In de.comm.software.newsserver are complains - and I could verify them -
that authentication is broken in the lua version of leafnode.  I suppose
that this is an artifact of an incomplete groupauth implementation.

The detailed complaint is that the luascript version of leafnode will
behave as though no authentication is required, i. e. lets users see all
groups, post everywhere, without authentication.


Can you please see to that:

- with --disable-lua, traditional leafnode behaviour is reinstated, i.
e. global authentication required for mostly everything that affects
groups or articles

- in the absense of Lua scripts, traditional leafnode behaviour is
reinstated, same conditions as above


I have, in the meanwhile, fixed some paths for the scripting, but I seem
to be unable to get groupauthentication running. There is a want_*
variable in scripthooks.lua, but it does not appear to be used anywhere
else.

Reading the C code, I do not see where leafnode would fall back to
traditional authentication behaviour for SCRIPT_UNAVAILABLE or
--disable-lua code.


Please contact me with a valid sender address that is not based on
dynamic DNS off-list.

Thanks.


In the meanwhile, I will prepare a non-scripting release of leafnode so
that people can get authentication (PAM or built-in) back.

Best regards
Matthias

backup or migration strategy

Thufir — 2012-09-17T21:18:25

Looking at a thread on backups:

http://thread.gmane.org/gmane.network.leafnode/3287

So that when the hard drive fails or fills up, you can just tar the spool?

Or, perhaps it would make sense configure leafnode to work off a 
secondary hard drive, although perhaps that would require configuring /
var with symlinks?


thanks,

Thufir

XNA

Thufir — 2012-09-10T05:54:51

I'm looking at the wikipedia entry on XNA:  

http://en.wikipedia.org/wiki/X-No-Archive

and looked over rfc's 1036 and 977 and see that there's an expiry header 
but don't see mention of this header.  How does XNA play with texpire?

I see that man texpire says:

       Texpire ignores "Expires" headers.

but makes no explicit mention of X-No-Archive which I see.




thanks,

Thufir

mail2news

Thufir — 2012-09-01T20:34:45

I believe that mail2news is part of mailman, but looking at:

http://thread.gmane.org/gmane.network.leafnode/3405

Maybe you don't need all of mailman?


The script is at:
http://pyg.sourcearchive.com/documentation/0.9.7build2/
news2mail_8py_source.html

thanks,

Thufir

re-fetch

Thufir — 2012-08-29T16:18:40

I set maxage to 0 after:

tin.users: skipping articles 1-1944 inclusive (initial limit)
tin.users: considering articles 1945 - 1994
tin.users: killed 1945 (<E1SJTed-00057J-NJ-KwcHiF2ZJ3sU7xepRLgDsBeDI10EO7PJ< at >public.gmane.org
karlsruhe.de>), too old (136 > 10) days

It's an initial fetch.  My expire settings:

root< at >dur:~# 
root< at >dur:~# grep expire /etc/news/leafnode/config
## Articles that haven't been read for <expire> days get expired and
## removed by texpire(8), unless you define special expire times.
## Note that texpire arguments can change this to relate to the date of
## arrival. See the texpire manual page.
expire = 999
groupexpire gwene.ca.craigslist.vancouver.labour.general = 3
groupexpire gwene.com.economist = 3
## Non-standard expire times (glob(7) wildcard constructs possible)
# groupexpire comp.os.linux.* = 5
# groupexpire any.local.newsgroup = 100
# archive this group (fetchnews uses the global expire to figure the
# maxage, but texpire will skip it):
# groupexpire my.archived.group = -1
## WARNING: maxage can be overridden by expire! See the clamp_maxage
## By default, maxage is limited to groupexpire (if applicable) or
## expire (otherwise). On some rare occasions, this limit may be
root< at >dur:~# 


to refetch, just remove from interesting.groups and then resubscribe?



thanks, 

Thufir

e-mail

Thufir — 2012-08-26T20:36:22

does Leafnode have a feature to send or forward messages to e-mail?  If 
not, any suggestions?

http://serverfault.com/questions/421414/nntp-to-e-mail-gateway

thanks,

Thufir

collision between number newsgroup names andmessage-ids

Joey Hess — 2012-08-25T19:29:57

Hi, I'm using leafnode to run the 30 year delayed usenet server olduse.net.
I recently ran into an interesting problem with /var/spool/news which
may or may not be specific to leafnode (been a long time since I used
other news servers!).

There was a newsgroup named net.micro.432. After net.micro got 431
posts, I noticed that my code was wanting to use the same number for a
post as was used for the newsgroup directory. While I've made my code
skip to the next post when this happens, it'd probably still break if a
new.micro.433 newsgroup appeared.

I'd like to avoid the ambiguity if I can, rather than do complicated
workarounds. Is there any alternative /var/spool/news layout I could use
to avoid it, that's supported by leafnode?

leafnode --> RDBMS

Thufir — 2012-08-14T11:28:18

Am I the only to ever try to squeeze NNTP into a database?

I really like the GNU Java API for NNTP, but it's too slow with large 
newsgroups.  It has to open a whole Folder and then search each message 
for the string to get the article number -- not practical.  However, it's 
light years ahead of the Apache API.

In any event, the GNU API uses javax.mail.Message, which is what e-mails 
use!  So, I was thinking of just using something like http://
www.dbmail.org/ and just e-mail all the articles to myself.

On a scale of one to ten, how crazy is that?



thanks,

Thufir

groupexpire syntax

Thufir — 2012-07-24T01:31:30

Just want ensure that I have the syntax correct:

expire = 999
groupexpire = gwene.com.economist = 35

will expire that specified group after 35 days while keeping the default 
of 999 days for other groups.



thanks,

Thufir

running texpire weekly instead of daily?

Adam Funk — 2012-07-12T11:06:23

I have a fairly long retention period in my leafnode configuration,
and the cron.daily/leafnode job takes about 1 hour to run.  (I'm
running leafnode-2.0.0.alpha20110807b.luascript, and the cron script
might have been copied from an old Debian/Ubuntu package.)

Is the texpire running time related more to the size of the spool or
the number of articles that it decides to delete?

Would it do any harm to run it weekly instead of daily?

Thanks,
Adam

large volume (binary) groups

Thufir — 2012-05-24T14:13:58

Leafnode doesn't handle binary groups well?

"Leafnode was not made for binary newsgroups. It is a "private" news-
_server_ and not a offline reader (though you can use it to fetch news 
online, store and read them offline)."

https://groups.google.com/group/alt.linux/msg/6877c648f696d94c?
dmode=source&output=gplain&noredirect



Could it be possible to have two different leafnode servers running, one 
for binary groups?  Or, is INN the better tool for grabbing binaries?  I 
have a slower broadband connection, and don't mind large download so long 
as they're in the background and not, for instance, holding up leafnode 
getting gmane text mailing list groups or something.



thanks,

Thufir

last message in a group

Thufir — 2012-05-24T13:08:32

I kinda get the interface, that you can use NEXT and LAST to iterate 
through the messages.  How do you jump to the last message, however? 

thufir< at >dur:~$ 
thufir< at >dur:~$ telnet localhost 119
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
200 Leafnode NNTP Daemon, version 1.11.8 running at localhost (my fqdn: 
dur.bounceme.net)
GROUP gwene.com.economist
211 257 3 259 gwene.com.economist group selected
LAST
422 There is no previous article
NEXT
223 4 <x1-GJW4YkN1lAh+j3VIpeOm7TsfRmc-DJntScIPk0PYtjvyW6yDsg< at >public.gmane.org> article retrieved
LAST
223 3 <x1-7lUSDcNIkWCalgmmCdb9BeZRH4g-DJntScIPk0PYtjvyW6yDsg< at >public.gmane.org> article retrieved
LAST
422 There is no previous article
NEXT
223 4 <x1-GJW4YkN1lAh+j3VIpeOm7TsfRmc-DJntScIPk0PYtjvyW6yDsg< at >public.gmane.org> article retrieved
QUIT
205 Always happy to serve!
Connection closed by foreign host.
thufir< at >dur:~$

add server

Thufir — 2012-05-22T05:24:37

I'm looking at:

http://leafnode.sourceforge.net/doc_en/FAQ.html

and don't see where to add a news server..?


I definitely do have groups and servers in
/var/spool/news/leaf.node# 

but I don't seem to have a 
/etc/leafnode/configure file.  I'm running Ubuntu 11.10




thanks,

Thufir

message index

Thufir — 2012-05-10T10:52:10

I'm using the following Java to interface with leafnode:

getMessages

public Message[] getMessages(int start,
                              int end)
             throws MessagingException

     Returns the messages in the given range (inclusive).

     Parameters:
         start - the number of the first message
         end - the number of the last message

http://www.gnu.org/software/classpathx/javamail/javadoc/javax/mail/
Folder.html#getMessages%28int,int%29

but just wanted to clarify how the nntp protocol and leafnode work a bit.

Is it correct that messages in a folder are numbered 1..n?  Is it, unless 
maybe in the middle of an operation, continuous, or can there be 
discontinuities?

I ask because I suppose it's possible to work at a lower level then GNU 
package, or retrieve messages individually, but it's much easier to a 
range of messages, as above, provided there are no missing messages or 
gaps.



thanks,

Thufir

received bogus greeting (502): Access Denied yet can telnet to port 119 fine

Arthur Marsh — 2012-04-08T03:23:04

Hi, I've been getting errors of the form:

error: news.internode.on.net: received bogus greeting (502): 502 
news.astraweb.com: Access Denied

from fetchnews, but I can do:

$ telnet news.internode.on.net 119
  DNS Lookup...  Trying 216.151.153.138...  Reverse DNS Lookup... (OK)
200 news.astraweb.com NNRP Service Ready (posting ok) (yEnc enabled).
quit
205 Transferred 71 bytes in 0 articles, 0 groups.  Disconnecting.

Any suggestions on how to debug this further?

Arthur.