Wayne,
I managed to telnet into the news server shortly after hitting
my conditional breakpoint in bits.c and tried some NNTP commands manually...
First the gdb info:
Breakpoint 1, find_existing_articles () at bits.c:288
288 ap = article_ptr(an);
(gdb) print *ngptr
$1 = {prev = 0x80f0620, next = 0x80f0670, rc = 0x80cd948,
rcline = 0x80f24e8 "comp.os.linux", abs1st = 13947, ngmax = 14077,
toread = 36, num = 619, numoffset = 14, subscribechar = 58 ':',
flags = -112 '\220'}
(gdb) print absfirst
$2 = 13947
(gdb) print an
$3 = 13946
(gdb) print ser_line
$4 = "13946\000\n\000cle Numbers Follow\000\n\000ows\000\n\000 y\000\n\000097 y\000\n\000\000\000\000140 y\000\n", '\000' <repeats 453 times>
(gdb)
Same problem as before. Now the telnet and manual NNTP commands:
group comp.os.linux
211 129 13946 14078 comp.os.linux
listgroup comp.os.linux
211 Article Numbers Follow
13947
13949
13950
....
article 13945
220 13945 <qcuqma.il8.ln@cc49395-b.wodhvn01.mi.comcast.net>
Path: rwcrnsc53!attbi_slave52!attbi_master51!attbi_feed4!attbi.com!12.120.28.17!attla2!attla1!ip.att.net!bos-service1.ext.raytheon.com!cyclone.swbell.net!newsfeed1.easynews.com!easynews.com!easynews!nntp2.aus1.giganews.com!nntp.giganews.com!nntp3.aus1.giganews.com!bin3.nnrp.aus1.giganews.com.POSTED!not-for-mail
From: Stuart Fuller <stufuller@usa.net>
You might ask why I asked for article 13945. Well on at least one
attempt, the 'listgroup' command told me it existed. It appears that
this server plays fast and loose with what 'group' and 'listgroup' say
about the lowest number article that is available. Sometimes there are
even articles available lower then either command returns.
Clearly trn shouldn't try to probe for all article numbers starting
from 1. My local newserver seems to arbitrarily return a lower value
for the lowest article for no apparent reason. So far the biggest
drop I've seen is 2. Under the circumstances your patch to bits.c
reasonable. I'll run trn with your bits.c patch for a while, but with a
conditional breakpoint as well to see if I ever see a discrepancy of
more then 2. If I do, I'll let you know.
Bill Bogstad
bogstad@pobox.com
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
