1# Options on a per-server basis, defaults shown
2
3
4# IP address of server if not same as server identifier (ie the tail of the
5# newsrc filename).
6# server news.cis.dfn.de
7
8# NNTP port.
9port 119
10
11# If no username or password are given, no authentication is performed.
12
13# User name.
14# user username
15
16# Password.
17# password xxxxxx
18
19# Whether to use SSL instead of a clear text socket for this server
20ssl no
21
22# Server needs MODE READER command.
23mode_reader yes
24
25# Some servers don't allow articles to be fetched even by Message-ID unless
26# a group has been entered. This option makes sure that is the case.
27enter_group no
28
29# Maximum number of threads for this server, overrides main.cf's
30# threads_per_server for this server.
31max_threads 3
32
33# Weighting used to cause this server to be favoured over others, the larger
34# the number, the heavier the bias. Typically you'd use a value in the
35# thousands, but anything over 10,000 will tend to prevent any unweighted
36# servers from ever being connected to. Avoid negative numbers unless you want
37# to temporarily prevent a server being used altogether, in which case use at
38# least 20,000 (ie mathematically < -20,000).
39server_weight 0
40
41# This overrides the option in main.cf on a per-server basis.
42pipeline yes
43
44# Whether server should be fed with outgoing articles from a feed with the
45# same name as the server's identifier
46feed yes
47
48# Leafname of an extra feed directory as well as or instead of the one enabled
49# by the feed option above. Usually this would be named after another server's
50# identifier.
51#extra_feed
52
53# If you often download very few articles at a time, the overhead of connecting
54# extra threads to some servers, only to find there are no new articles to
55# download, may actually slow the overall process down, so this setting causes
56# newsstar not to start additional threads until a server has at least
57# min_worth_getting articles available. Previous versions had this fixed at 25.
58min_worth_getting 0
59
60# Usually newsstar uses XHDR to receive a list of Message-IDs of articles to be
61# potentially fetched. XHDR may cause a performance hit on some servers as they
62# look up the body of the article, so it may be more efficient to use XOVER to
63# extract the information from the overview, even though it uses more bandwidth
64# by also providing extra header information we don't need - although future
65# versions of newsstar will use this for killfiling. The xover option enables
66# use of XOVER.
67xover no
68
69# Default score for killfile
70default_score 100
71
72# If kill_score < 0 and an article scores between 0 and kill_score, newsstar
73# will generate a dummy article for it from its XOVER headers. Articles scoring
74# any lower will be killed altogether (as earlier behaviour)
75kill_score 0
76
77# Due to clustering, some servers create articles out of order aka backfilling,
78# ie article numbers are allocated without the article necessarily becoming
79# available until after later article numbers have been allocated to articles
80# that are available. This makes it more difficult for fetchers to keep track
81# of what they've downloaded. Supernews.com is one widely used server reputed
82# to do this. Newsstar can be configured to keep track of such articles for a
83# certain number of days given by backfill_expiry, or until they can be
84# downloaded. Use a value of 0 to disable the feature. Note this used to
85# default to 7 days, now changed to 0 (disabled).
86backfill_expiry 0
87
88# One server I use rejects duplicate uploads with a message of the form:
89# 441 436 <Message-ID>
90# Enabling this option causes such messages to be ignored if duplicate
91# reporting is not enabled.
92ignore_436 false
93
94# Entry to add to path header of downloaded articles. This makes it easier to
95# configure INN not to attempt to post incoming articles back out whence they
96# came. May also be necessary to make it accept header-only pseudo-articles.
97# path server.
98