|
|
1.1 root 1: .TH MKGRDATES 8C "26 October 1985"
2: .UC 4
3: .SH NAME
4: mkgrdates \- Make the newsgroup creation date file
5: for the NNTP server
6: .SH SYNOPSIS
7: .B /usr/spool/news/lib/mkgrdates
8: .SH DESCRIPTION
9: .I Mkgrdates
10: maintains the newsgroup creation date file for the
11: Network News Transfer Protocol server (which uses
12: the newsgroup creation date file for the NEWGROUPS
13: command). When executed,
14: .I mkgrdates
15: goes through the active file
16: (usually
17: .IR /usr/spool/news/lib/active )
18: and creates a file
19: (usually
20: .IR /usr/spool/news/lib/groupdates )
21: containing the creation dates and times of the
22: newsgroups in the active file. The
23: .I groupdates
24: file has one line for each newsgroup, in the form
25: ``date newsgroup'' where ``date'' is the newsgroup's
26: creation time, in seconds since midnight, Jan. 1,
27: 1970, GMT,
28: and ``newsgroup'' is the name of the newsgroup in question.
29: .PP
30: Since generating the
31: .I groupdates
32: file is expensive, when
33: .I mkgrdates
34: is run it creates a file (usuaully
35: .IR /usr/spool/news/lib/mkgdstats )
36: containing the length, in bytes, of the active file.
37: The next time
38: .I mkgrdates
39: is invoked, it compares the contents of the
40: .I mkgdstats
41: file against the length of the active file, and
42: exits if these lengths are the same. If the
43: lengths are not the same, it rebuilds the
44: .I groupdates
45: file, and sets the
46: .I mkgdstats
47: file to the appropriate value.
48: This scheme avoids a fair amount of what would
49: essentially be wasted computing.
50: .PP
51: .I Mkgrdates
52: should be run once or twice a day by an entry in
53: .I /usr/lib/crontab
54: .SH BUGS
55: It would be good to get rid of this and have the
56: news server do it, but nobody seems to be able
57: to come up with an easy, inexpensive way.
58: I'm certainly open to suggestions.
59: .SH AUTHOR
60: Phil Lapsley
61: .SH SEE ALSO
62: cron(1),
63: nntpd(8C)
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.