I seem to have some problems with nget and the news
provider 'ultrafeed'. If I retrieve headers or files
from a group from which I already have the headers,
nget tells me the group does not exist, although the
group DOES exist. If I remove the headers files, I do
not get the error.
I tested this with 0.24.1, 0.24.2, 0.25 and the latest
CVS. All with the same result. However, I don't have
any problems with other newsservers.
OS: RedHat Linux 9
% rm $HOME/.nget4/alt.binaries.multimedia.startrek
% ./nget -h ultrafeed -g alt.binaries.multimedia.startrek
make_connection(1,news.ultrafeed.com,nntp,0xbfffe560,512)
Connecting to 66.150.105.205:119
u >> 200 Welcome xxx.xxx.xxx.xxx cnt 2
u << GROUP alt.binaries.multimedia.startrek
u >> 480 Username/password required
u << AUTHINFO USER
u >> 381 More authentication information required
u << AUTHINFO PASS *
u >> 281 Authentication accepted: cnt: 4
u << GROUP alt.binaries.multimedia.startrek
u >> 211 19276 2409766 2428946
alt.binaries.multimedia.startrek selected
Retrieving headers 2409766-2428946 : 18003/18003/19181
100% 48342B/s 1m34s
saving cache: 18003 parts, 299 files.. done. (18003 sa)
u << QUIT
OK: 1 group
% ./nget -h ultrafeed -g alt.binaries.multimedia.startrek
make_connection(1,news.ultrafeed.com,nntp,0xbfffde60,512)
Connecting to 66.150.105.205:119
u >> 200 Welcome xxx.xxx.xxx.xxx cnt 2
u << GROUP alt.binaries.multimedia.startrek
u >> 480 Username/password required
u << AUTHINFO USER *
u >> 381 More authentication information required
u << AUTHINFO PASS *
u >> 281 Authentication accepted: cnt: 4
u << GROUP alt.binaries.multimedia.startrek
u >> 211 19276 2409827 2428946
alt.binaries.multimedia.startrek selected
fatal error, won't try ultrafeed again (bad reply 412:
412 no such newsgroup)
no servers queried successfully
freeing cache: 18003 parts, 299 files..
u << QUIT
ERRORS: 1 group
=====
Dennis
Logged In: YES
user_id=65253
Can you try setting debug=2 in your .ngetrc and see what it
says then?
Logged In: NO
=====
goodgetcwd 48 /export/home/dennis/.nget4(0xbffff07c-0x80eacb8)
init_term_stuff: using tputs_clr_bol
goodgetcwd 48 /export/home/dennis/.nget4(0xbffff088-0x80d9358)
attempting to flock
/export/home/dennis/.nget4/alt.binaries.multimedia.startrek,midinfo
read 18053 parts (18053 sa) 300 files
nntp_group: serv(ultrafeed) 0.000000>=1.000000
aquiring connection to ultrafeed
make_connection(1,news.ultrafeed.com,nntp,0xbfffeae0,512)
Connecting to 66.150.105.205:119
aquired connection to ultrafeed (0x8112368)
u >> 200 Welcome xxx.xxx.xxx.xxx cnt 2
u << MODE READER
u >> 200 Are you happy now?
u << GROUP alt.binaries.multimedia.startrek
u >> 480 Username/password required
u << AUTHINFO USER *
u >> 381 More authentication information required
u << AUTHINFO PASS *
u >> 281 Authentication accepted: cnt: 2
u << GROUP alt.binaries.multimedia.startrek
u >> 211 18011 2415153 2434351
alt.binaries.multimedia.startrek selected
u << LISTGROUP
u >> 412 no such newsgroup
releasing connection to (0x8112368)
penalized ultrafeed: count 1, last 1071653062, time 0
fatal error, won't try ultrafeed again (log.cc:38:caught
exception prot_nntp.cc:95:ProtocolExFatal: bad reply 412:
412 no such newsgroup)
no servers queried successfully
freeing cache: 18053 parts, 300 files..
u << QUIT
ERRORS: 1 group
=====
It looks like the 'listgroup' command is given, without any
argument. IIRC, the listgroup command requires the group as
argument... ???
Dennis
Logged In: YES
user_id=65253
Both rfc2980 and draft-ietf-nntpext-base-20 state that the
argument is optional. (I'd like to avoid adding workarounds
for broken servers, so I'd be happy if you could contact
them and see if they can fix it..) As a temporary
workaround, you can use fullxover=1 for that server.
Logged In: NO
I contacted the ultrafeed guys, and I got a very short reply:
-----
There are reasons why we do not accept certain commands.
Bill
------
I then asked for those reasons (and explained it again), but
I didn't get an answer as of yet....
I tried the fullxover option. It works, but it gets terribly
slow (100B/s), which is far too slow....
Is there any possibility of a 'broken-listgroup'
configuration option? Or perhaps something at compile-time..??
Thanks
Dennis
Logged In: YES
user_id=65253
If fullxover=1 is too slow, you could always try fullxover=0..
Logged In: NO
fullxover=0 seems to be a good workaround....
Thanks