lad@lad.sdl.UUCP (Lawrence Deleski) (09/16/89)
I'm having a problem doing a rebuild with the patchlevel 18 expire. Using
expire -r -I -e 5 -v6, expire runs for quite some time before spouting out the
message:
Bogus local distribution rejected
Expire then dies.
The code from header.c where the error occurs is:
case DISTRIBUTION:
getfield(hp->distribution, sizeof(hp->distribution));
if (STRCMP(hp->distribution, "net") == 0
|| STRCMP(hp->distribution, "world") == 0)
hp->distribution[0] = '\0';
else if (STRCMP(hp->distribution, "local") == 0 &&
hp->path[0] != '\0' &&
#ifdef GENERICPATH
strncmp(hp->path, PATHSYSNAME,
strlen(PATHSYSNAME)) != 0) {
#else /* !GENERICPATH */
strncmp(hp->path, LOCALPATHSYSNAME,
strlen(LOCALPATHSYSNAME)) != 0) {
#endif /* !GENERICPATH */
hp->distribution[0] = '\0';
xerror("Bogus local distribution
rejected");
It seems that it looks in the /usr/lib/news/distributions file for the
standard distributions net, world, and local. I edited mine to have just
those three. Expire did the same thing.
BTW, I can do the standard expire (no rebuild) but I cannot rebuild. Any
ideas?
--
Lawrence A. Deleski | Silicon Compiler Systems
uunet!sdl!lad | 15 Independence Blvd.
Cash-We-Serve 76127,104 | Warren, NJ 07060
MABELL: (201) 580-0102 | Ext. 216tbetz@lilink.UUCP (Tom Betz) (09/17/89)
Quoth lad@lad.sdl.UUCP (Lawrence Deleski) in <53@lad.sdl.UUCP>: | |I'm having a problem doing a rebuild with the patchlevel 18 expire. Using |expire -r -I -e 5 -v6, expire runs for quite some time before spouting out the |message: | I had all sorts of troubles, (on upaya - which is why I am posting from lilink) and am in the process of trying to set them aright. First, all the new articles tha made it through were created with numbers having 64 in them (ie. 64,264,6464,26464,646464,2646464....) which really threw off everything, and made rn look for a long, long time from one article to the next. It's now reporting "2576768 unread articles in misc.jobs.resumes" when I have done an 'expire -i0 -I21 -r' to clean up the message base and start over, and there are exactly 0 articles there.... but still rn is having to seek forever to find this out. Is there a way to invoke rn so that it will catch up all newsgroups? -- "There are no magicians. There are no hidden | rutgers!lilink!upaya!tbetz people or gurus somewhere in secret. You have | Tom Betz, Greyston Foundation already been given the teaching by the greatest | 114 Woodworth Ave. illuminated sages born." - Gopi Krishna | Yonkers, NY 10701-2509
ambar@bloom-beacon.mit.edu (Jean Marie Diaz) (09/18/89)
From: tbetz@lilink.UUCP (Tom Betz) Date: 16 Sep 89 21:50:32 GMT It's now reporting "2576768 unread articles in misc.jobs.resumes" when I have done an 'expire -i0 -I21 -r' to clean up the message base and start over, and there are exactly 0 articles there.... but still rn is having to seek forever to find this out. Is there a way to invoke rn so that it will catch up all newsgroups? This isn't rn's problem, this is expire's problem. If you read the code, you will note that expire reads the max field from the active file, and unless min > max or an-article-number > max, it never gets changed. What we did here was to turn off news; run a emacs macro over the active file to convert all the min and max fields to 0000000, and then run expire -u, which updated the active file correctly. AMBAR