noffle-users Mailing List for Noffle - news server (Page 2)
Brought to you by:
bears
Menu
▾
▴
noffle-users
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(2) |
Jul
(2) |
Aug
(4) |
Sep
|
Oct
(1) |
Nov
(1) |
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(4) |
Feb
|
Mar
(2) |
Apr
(2) |
May
|
Jun
(2) |
Jul
|
Aug
|
Sep
(1) |
Oct
(1) |
Nov
|
Dec
|
| 2002 |
Jan
|
Feb
|
Mar
(1) |
Apr
(6) |
May
|
Jun
(3) |
Jul
(1) |
Aug
(2) |
Sep
(3) |
Oct
(1) |
Nov
|
Dec
(1) |
| 2003 |
Jan
(17) |
Feb
(24) |
Mar
(3) |
Apr
(1) |
May
(27) |
Jun
(5) |
Jul
(7) |
Aug
(9) |
Sep
(8) |
Oct
(13) |
Nov
(9) |
Dec
(5) |
| 2004 |
Jan
|
Feb
|
Mar
(24) |
Apr
(5) |
May
(5) |
Jun
(2) |
Jul
(10) |
Aug
(8) |
Sep
(8) |
Oct
(18) |
Nov
(5) |
Dec
(1) |
| 2005 |
Jan
|
Feb
|
Mar
(1) |
Apr
(1) |
May
(2) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
| 2006 |
Jan
(3) |
Feb
(3) |
Mar
(1) |
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
(48) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2009 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2023 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
|
From: Dan J. <ji...@ji...> - 2004-10-17 23:30:03
|
J> How many groups do you have? I just tried your function on my news host (a J> mighty K6/2 350MHz) which has 38k groups registered, and it took 3 seconds. 34k. Sure it is quick once the file has been cached into memory or whatever. I bet it isn't that quick the first time done after reboot. M> echo "list active *shortwave*" | noffle --server Hrumf, no speed improvement over noffle -g|awk. P.S., python gave: gdbm.error: (11, 'Resource temporarily unavailable') |
|
From: Dan J. <ji...@ji...> - 2004-10-17 18:43:31
|
Guess what I caught that nasty noffle doing?
Whenever you read an article it causes a request for all its parents.
Indeed,
J> They are further marked INTERESTING when read or an article that
J> references them is read
J> INTERESTING
J> - This message should be downloaded on the next fetch, if it has
J> not been downloaded already.
and this was in overview mode too. I've never used thread mode.
It turn out that (except for perhaps noffle -a) there's just no
concept of reading a single article. Oh no, plenty of background
material for we avid readers. No jumping to conclusions without
dragging all the postulates down our modem to boot.
Indeed, young Dan was mislead by the man page:
over fetch only article overviews by default. Opening an article
marks it for download next time online,
thread like over, but download articles full if an article of the same
thread already has been downloaded.
Note how "over" doesn't mention the free Encyclopedia Britannica you
unknowing also sign up for each fateful press of the spacebar. "I made
a date with the daughter, only to find her parents along for the ride too."
|
|
From: Mirko L. <mir...@we...> - 2004-10-15 10:57:21
|
Dan Jacobson:
> Gentlemen, today we shall discuss searching for group names. E.g.,
> let's ask noffle what is the full name of that shortwave group.
echo "list active *shortwave*" | noffle --server
> I see I have written a bash function to do this:
> noffle_search_group ()
> {
> noffle -g | awk -F\\t '$1~'/$@/'{print $6,$1}' | sort -k 2
> }
> e.g.,
> $ noffle_search_group shortwave
> y rec.radio.shortwave
What about reading directly from the groupinfo database? I'd rather
backup the database before doing anything fancy, though. Here's a
simple example:
#!/usr/bin/env python
# 0755 noffle_active_re.py
# Tested at least once. Works for me but your mileage may vary.
# Requires Python incl. gdbm module
# Mirko Liss 2003-08-10
import gdbm, re, string, struct
import sys
groupdb = "/var/spool/noffle/data/groupinfo.gdbm"
def active_re( pattern ):
searcher = re.compile( pattern, re.IGNORECASE ).search
db = gdbm.open( groupdb, "r" )
matchlist = filter( searcher, db.keys() )
matchlist.sort()
for a in matchlist:
entry = db[a]
(first, last, remotelast, createdTime,
lastAccessTime) = struct.unpack( "iiiii", entry[:20] )
(server, description) = string.split( entry[20:-6], "\0" )
(postAllow, lastPostTime) = struct.unpack( "=ci", entry[-5:] )
# print groupname without "\0"
print a[:-1], last, first, postAllow
return
if len( sys.argv ) > 1:
active_re( sys.argv[1] )
else:
print "Usage: %s pattern" % sys.argv[0]
# End-Of-File
|
|
From: Jim H. <jim...@ac...> - 2004-10-15 09:24:14
|
On 14-Oct-2004 Dan Jacobson wrote:
> Gentlemen, today we shall discuss searching for group names. E.g.,
> let's ask noffle what is the full name of that shortwave group.
>
> I see I have written a bash function to do this:
> noffle_search_group ()
> {
> noffle -g | awk -F\\t '$1~'/$@/'{print $6,$1}' | sort -k 2
> }
> e.g.,
> $ noffle_search_group shortwave
> y rec.radio.shortwave
>
> The problem is it is computationally expensive, judged by how long it
> takes, at least on initial run.
How many groups do you have? I just tried your function on my news host (a
mighty K6/2 350MHz) which has 38k groups registered, and it took 3 seconds.
That being said, I do think there is something to be said for the output from
'noffle -g' being by default
group-name group-description
and the current
group-name group-server first-article-no last-article-no upstream-next-article
post-status create-date last-access-date group-description
being reserved for 'noffle -g -v' or 'noffle -g verbose'.
The first option would reduce or remove the requirement for awk here.
--
Jim Hague - jim...@ac... Never trust a computer you can't lift.
|
|
From: Dan J. <ji...@ji...> - 2004-10-15 00:01:05
|
Gentlemen, today we shall discuss searching for group names. E.g.,
let's ask noffle what is the full name of that shortwave group.
I see I have written a bash function to do this:
noffle_search_group ()
{
noffle -g | awk -F\\t '$1~'/$@/'{print $6,$1}' | sort -k 2
}
e.g.,
$ noffle_search_group shortwave
y rec.radio.shortwave
The problem is it is computationally expensive, judged by how long it
takes, at least on initial run.
Therefore, one wonders if instead inside noffle itself, with its vast
computational capabilities, such searching could be done, as a new
feature?
|
|
From: Dan J. <ji...@ji...> - 2004-10-04 22:03:50
|
Jim> How does this sound - add a new option, Jim> -U, --uninteresting <msgid> Yes, good to have at least this way to unrequest articles. |
|
From: Jim H. <jim...@ac...> - 2004-09-30 22:36:30
|
On 30-Sep-2004 Dan Jacobson wrote:
> Well still sometimes there are articles with empty X-NOFFLE-Status's.
> I saw some for some filter action=over matching articles.
You're quite right. I just did a 'noffle -a all' and got several recent
articles with empty X-NOFFLE-Status, so I've missed a bit of analysis somewhere.
Oh, of course. If a group is subscribed in 'full', articles will be downloaded
regardless of their INTERESTING status. So, since I subscribe to everything in
'full', only the articles I've actually read should have an INTERESTING status.
I'll update the FAQ text.
> By the way, if a filter matches, maybe a header about it with details
> could be added to the message. SpamAssassin adds them, matches or
> not, so noffle could perhaps add them for matches.
>
> Jim> INTERESTING
> Jim> - This message should be downloaded on the next fetch, if it has
> Jim> not been downloaded already.
> Any antidotes in case we accidentally cause "Garage Sale Last Thursday
> in Fergus Falls" to be marked INTERESTING? Does tampering with
> var/spool/noffle/outgoing/* affect it?
Not right now, I'm afraid. Fiddling with outgoing will stop it being downloaded
on the next fetch (I think), but it will remain marked as INTERESTING and get
added to the queue again.
I really should do something about that. How does this sound - add a new
option,
-U, --uninteresting <msgid>
to allow you to remove the interesting tag from an article and remove it from
the current requested queue, if on it.
I suppose
-b, --boring <msgid>
could be a synonym :-)
--
Jim Hague - jim...@ac... Never trust a computer you can't lift.
|
|
From: Dan J. <ji...@ji...> - 2004-09-30 04:04:20
|
Well still sometimes there are articles with empty X-NOFFLE-Status's. I saw some for some filter action=over matching articles. Also noffle -a on articles fetched by early noffle versions seem to have empty X-NOFFLE-Status. By the way, if a filter matches, maybe a header about it with details could be added to the message. SpamAssassin adds them, matches or not, so noffle could perhaps add them for matches. Jim> INTERESTING Jim> - This message should be downloaded on the next fetch, if it has Jim> not been downloaded already. Any antidotes in case we accidentally cause "Garage Sale Last Thursday in Fergus Falls" to be marked INTERESTING? Does tampering with var/spool/noffle/outgoing/* affect it? P.S. seeing X-NOFFLE-Status now in the gnus summary buffer has added new colour to my life. A new air of command has swept over me. As Aerosmith said: back in the saddle again. |
|
From: Jim H. <jim...@ac...> - 2004-09-29 22:15:08
|
On 28-Sep-2004 Dan Jacobson wrote:
> To the documentation please add a table of all the possible
> X-NOFFLE-Status headers' meanings and how they might be found combined
> on one X-NOFFLE-Status line, and CC it here.
Will this do? As I type this mail, it occurs to me that it doesn't cover the
expected combinations. All articles are marked NOT_DOWNLOADED when initially
created. They are further marked INTERESTING when read or an article that
references them is read, and both articles are known at the time of reading,
and once marked INTERESTING stay so marked. NOT_DOWNLOADED is removed after a
successful download. RETRIEVING_FAILED I think should only appear in
conjunction with both the others, and again is removed after a successful
download.
Q: What are the possible values of the X-NOFFLE-Status: header that
Noffle adds to messages?
A: The X-NOFFLE-Status: header can contain one, two or all of the
following, separated by spaces:
INTERESTING
- This message should be downloaded on the next fetch, if it has
not been downloaded already.
NOT_DOWNLOADED
- The article has not yet been downloaded.
RETRIEVING_FAILED
- An attempt was made to retrieve the article from upstream, but
the attempt failed.
--
Jim Hague - jim...@ac... Never trust a computer you can't lift.
|
|
From: Dan J. <ji...@ji...> - 2004-09-29 06:00:54
|
The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.
Done.
(setq gnus-agent nil) ;we use noffle
(setq nntp-nov-is-evil t);get X-NOFFLE-Status always
(setq gnus-extra-headers
'(To Newsgroups X-NOFFLE-Status))
(setq gnus-summary-line-format
(concat
"%U%R%z%I%(%t|%~(cut-left 3)d%~(max-right 2)d%4L|%-20,20f%)"
"%~(max-right 1)~(form(gnus-extra-header 'X-NOFFLE-Status))@"
;I want to use s/([A-Z])[A-Z]+/\1/g; s/ //g; above to get the first
;letter of all the items on the X-NOFFLE-Status line, but don't know
;how in gnus, so we just get the first letter.
" %s"
"\n"))
(setq nnmail-extra-headers gnus-extra-headers)
|
|
From: Dan J. <ji...@ji...> - 2004-09-28 23:50:53
|
To the documentation please add a table of all the possible X-NOFFLE-Status headers' meanings and how they might be found combined on one X-NOFFLE-Status line, and CC it here. |
|
From: Dan J. <ji...@ji...> - 2004-09-28 23:49:48
|
The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.
Odd, one day I had a modem hiccup and this started working,
(setq gnus-extra-headers
'(To Newsgroups X-NOFFLE-Status))
(setq gnus-summary-line-format
;"%U%R%z%I%(%t|%~(cut-left 3)d%~(max-right 2)d%4L|%-20,20f%):%s\n")
(concat "%U%R%z%I%(%t|%~(cut-left 3)d%~(max-right 2)d%4L|%-20,20f%):"
"%~(max-right 1)~(form(gnus-extra-header 'X-NOFFLE-Status))@"
" %s"
"\n"))
But only for some of the articles from some of the groups... I will
continue observations. Note I have not arrived at the idea
X-NOFFLE-Status slice yet. I am first observing if I get any string at all.
|
|
From: Dan J. <ji...@ji...> - 2004-09-09 22:37:18
|
The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.
Y> Y==Yair Friedman (ya...@Ma...)
Y> If I understood you you want to change the summary according to the
Y> presence/absence of X-NOFFLE-Status header. I did something similar for
Y> Importance MS headers my boss uses :)
Y> The basic steps for doing it are:
Y> * adding X-NOFFLE-Status to nnmail-extra-headers
Y> * defining gnus-user-format function that will search this header using
Y> mail-header-extra and return mark needed.
Y> * adding %u spec in gnus-summary-line-format.
From what I could piece together from the Gnus manual,
(setq gnus-extra-headers
'(To Newsgroups X-NOFFLE-Status))
(setq gnus-summary-line-format
(concat "%U%R%z%I%(%t|%~(cut-left 3)d%~(max-right 2)d%4L|%-20,20f%):%s"
"%~(form(gnus-extra-header 'X-NOFFLE-Status))@\n"))
(setq nnmail-extra-headers gnus-extra-headers)
should have appended X-NOFFLE-Status to each summary line.
But it doesn't.
Yes X-NOFFLE-Status is in the articles.
|
|
From: Carles A. <ll...@ar...> - 2004-09-03 16:53:02
|
I use to modiffy by hand the /var/spool/noffle/requested/* files and never noticed this problem using different news clients on both Linux and Windows. Maybe your news client has some "feature" causing those unwanted requests. Regards, Carles Arjona |
|
From: Dan J. <ji...@ji...> - 2004-08-31 18:10:05
|
Gentlemen, I clearly remove the unwanted requests, see: # ed /var/spool/noffle/requested/news.individual.net 131 ,d wq 0 but half an hour later, there they are back in that file, even though I long ago stopped browsing that group. Something is making real sure those requests stay in that file. I have to delete them twice before I connect to make sure I don't slurp them down. |
|
From: Dan J. <ji...@ji...> - 2004-08-17 20:17:18
|
The following message is a courtesy copy of an article that has been posted to gnu.emacs.gnus as well. It seems us noffle gnus users do indeed have a chance to know if a message has been downloaded or not when merely looking at gnus' *Summary* buffer. The X-NOFFLE-Status header could be examined by gnus, and then some change to the article's summary line could be made, much like the number of lines is extracted and put on the article's summary line. Within moments of my posting, the formula will surely be posted. If one could tell which articles haven't been downloaded (due to one's noffle filters, etc.) in the summary buffer, one could avoid requesting them if one didn't want to read them that bad, etc. I hate that land mine feeling: I thought you were on disk already, that's why I bothered to try to read you. But you aren't, and now I'm making a special request. Blast. Now I must _try_ to clean up the mess with http://jidanni.org/comp/ppp/noffle-wipe-last-request |
|
From: Jim H. <jim...@ac...> - 2004-08-07 06:59:54
|
On 05-Aug-2004 Dan Jacobson wrote: > Jim> What exactly are you wanting to see? > > Ah, I know, something like procmail's log: >>From fet...@ji... Fri Aug 06 03:46:27 2004 > Subject: Re: [Noffle-users] msgid filter didn't work > procmail: No match on "^From: .*jidanni" > procmail: No match on "^From: .*jidanni@" > ... > procmail: Match on "^From: .*jidanni@" > procmail: No match on "^X-Mailer: reportbug" > procmail: Match on "^To:.*\<()jidanni\>" > procmail: Match on ! "(^(Mailing-List:|Precedence:.*(junk|bulk|list)... > > Or maybe just print the match lines. I'm working on it. There'll be a couple of weeks delay while I'm away. -- Jim Hague - jim...@ac... Never trust a computer you can't lift. |
|
From: Dan J. <ji...@ji...> - 2004-08-05 23:31:00
|
Jim> What exactly are you wanting to see? Ah, I know, something like procmail's log: >From fet...@ji... Fri Aug 06 03:46:27 2004 Subject: Re: [Noffle-users] msgid filter didn't work procmail: No match on "^From: .*jidanni" procmail: No match on "^From: .*jidanni@" ... procmail: Match on "^From: .*jidanni@" procmail: No match on "^X-Mailer: reportbug" procmail: Match on "^To:.*\<()jidanni\>" procmail: Match on ! "(^(Mailing-List:|Precedence:.*(junk|bulk|list)... Or maybe just print the match lines. |
|
From: Dan J. <ji...@ji...> - 2004-08-05 23:30:53
|
Jim> Just a swift guess, but don't forget the msgid parameter is a regular Jim> expression. So I wonder if some escaping is needed in the pattern, e.g. Jim> msgid="ce....\$...\$...@netnews.hinet.net" Well, hurmf, we all know that $ is only special at the end of regxps. I would test your theory if I didn't have to wait for a real message to come along to do it. I.e., too bad one can't do $ noffle --test-rcfile [rcfile] < test-message which would be great. Of course this should write to stdout and not disturb the database. |
|
From: Jim H. <jim...@ac...> - 2004-08-05 14:07:06
|
On 03-Aug-2004 Dan Jacobson wrote: > filter group=tw.bbs.rec.radio.amateur > msgid="ce....$...$...@netnews.hinet.net" action=over > didn't work on Message-ID: > <cem6eh$9n3$18...@ne...> Just a swift guess, but don't forget the msgid parameter is a regular expression. So I wonder if some escaping is needed in the pattern, e.g. msgid="ce....\$...\$...@netnews.hinet.net" -- Jim Hague - jim...@ac... Never trust a computer you can't lift. |
|
From: Jim H. <jim...@ac...> - 2004-08-05 13:58:34
|
Hi Dan, (I'm in and out a bit at present. Holiday time...) On 30-Jul-2004 Dan Jacobson wrote: > It is very important that noffle tell more than just > > Filter 16 fired on message <cedirn$1ui$1...@ne...> > > in order for us to be sure that we are not killing off good messages. > We need to be able to see Subject and From, etc. if we want. What exactly are you wanting to see? The Subject: and From: of the message killed? That plus a log of which rules fired? E.g. how about Number of references rule: 1 < 3 Filter 3 fired on message <something@some.where> or do you want Number of references rule: 1 < 3 Filter 3 fired on message <something@some.where> Subject: Kill this message From: ign...@ho... -- Jim Hague - jim...@ac... Never trust a computer you can't lift. |
|
From: Dan J. <ji...@ji...> - 2004-08-03 22:32:46
|
filter group=tw.bbs.rec.radio.amateur msgid="ce....$...$...@netnews.hinet.net" action=over didn't work on Message-ID: <cem6eh$9n3$18...@ne...> |
|
From: Dan J. <ji...@ji...> - 2004-07-30 23:09:01
|
It is very important that noffle tell more than just Filter 16 fired on message <cedirn$1ui$1...@ne...> in order for us to be sure that we are not killing off good messages. We need to be able to see Subject and From, etc. if we want. (True I bet one could still turn on maximum debugging or something.) |
|
From: Balas M. <ba...@tt...> - 2004-07-05 15:56:07
|
Hi! Setting the file permissions did solve the problem; thank you very much for your help! Mark |
|
From: Jim H. <jim...@ac...> - 2004-07-03 13:16:58
|
On 02-Jul-2004 Balas Mark wrote: > There is a problem here: > > trychydts:~# telnet localhost nntp > Trying 127.0.0.1... > Connected to trychydts. > Escape character is '^]'. > 200 NNTP server NOFFLE 1.0.1 > list active uk.comp.os.* > 503 Cannot init server > 215 Groups > Connection closed by foreign host. OK. I think I know what's up. It's a file permissions problem. Run as a server, Noffle can't open its database file. Assuming your inetd line looks like this: nntp stream tcp nowait news /usr/sbin/tcpd /usr/bin/noffle -r Noffle is run as user 'news'. /var/spool/noffle should be owned by news, and the noffle executable suid news. Try doing the following: chown news:news /usr/bin/noffle chmod 06755 /usr/bin/noffle chown -R news:news /var/spool/noffle and see if it works. -- Jim Hague - jim...@ac... Never trust a computer you can't lift. |
60 messages has been excluded from this view by a project administrator.
