From rsalz@uunet.uu.net Mon Nov 28 14:45:23 EST 1994 Article: 7063 of news.software.nntp Path: news.pop.psu.edu!news.cac.psu.edu!howland.reston.ans.net!news.sprintlink.net!EU.net!uunet!not-for-mail From: rsalz@uunet.uu.net (Rich Salz) Newsgroups: news.software.nntp Subject: Re: Yet Again (was Re: Innxmit bug again) Date: 28 Nov 1994 13:41:37 -0500 Organization: UUNET Technologies Inc, Falls Church, VA, USA Lines: 7 Distribution: inet Message-ID: <3bd891$oqs@rodan.UU.NET> References: <39nqti$kvu@horos.kbfi.ee> <3a8pf7$rbv@papaioea.manawatu.planet.co.nz> <3bb891$kbr@sol.sun.csd.unb.ca> <3bbs90$9t5@slapshot.Stanford.EDU> NNTP-Posting-Host: rodan.uu.net In <3bbs90$9t5@slapshot.Stanford.EDU> schemers@leland.stanford.edu (Roland Schemers) writes: >Why doen't innxmit just punt on lines that are too long? My code has a bug. Sorry. In backends/innxmit.c, make sure the if() for QIOtoolong occurs before the if() for QIOerror. There are a couple of places to fix this. /r$ From rsalz@uunet.uu.net Fri Dec 2 09:31:48 EST 1994 Article: 7154 of news.software.nntp Path: news.pop.psu.edu!news.cac.psu.edu!howland.reston.ans.net!europa.eng.gtefsd.com!gatech!swrinde!pipex!uunet!not-for-mail From: rsalz@uunet.uu.net (Rich Salz) Newsgroups: news.software.nntp Subject: Re: Problem w/ innxmit in a master/slave configuration Date: 2 Dec 1994 01:34:54 -0500 Organization: UUNET Technologies Inc, Falls Church, VA, USA Lines: 10 Distribution: na Message-ID: <3bmf6e$5si@rodan.UU.NET> References: <3bfkrl$f22@purfleet.advantis.com> NNTP-Posting-Host: rodan.uu.net In <3bfkrl$f22@purfleet.advantis.com> crystal@purfleet.advantis.com (Darren Crystal) writes: >When innxmit (on the master) encounters a news article >that is cross-posted to LOTS of other news groups (ie. >a VERY long line) it segmentation faults...and WON'T >get past the offending line. This causes the news flow >to stop until I can manually fix it... It's a bug. See my note about reversing QIOerror/QIOtoolong tests... Also in one place the QIO test is called with the wrong QIO variable. /r$ Return-Path: cudep@csv.warwick.ac.uk Received: from spatula.csv.warwick.ac.uk (cudep@spatula.csv.warwick.ac.uk [137.205.193.200]) by bosnia.pop.psu.edu (8.6.9/8.6.9) with ESMTP id NAA20069 for ; Wed, 14 Dec 1994 13:48:56 -0500 Date: Wed, 14 Dec 1994 18:48:43 GMT From: Ian Dickinson Message-Id: <6600.199412141848@spatula.csv.warwick.ac.uk> Received: by spatula.csv.warwick.ac.uk id SAA06600; Wed, 14 Dec 1994 18:48:43 GMT In-Reply-To: David Barr "Re: INN patch for the repository" (Dec 14, 1:38pm) X-Mailer: Mail User's Shell (7.2.5 10/14/92) To: David Barr Subject: Re: INN patch for the repository MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Here's the patch for innxmit-long-lines. I don't honestly know is this is correct, since I've never encountered the problem so far... *** backends/innxmit.c.ORIG Thu Mar 18 21:03:28 1993 --- backends/innxmit.c Wed Dec 14 18:41:54 1994 *************** *** 415,425 **** for ( ; ; ) { if ((p = QIOread(BATCHqp)) == NULL) { - if (QIOerror(BATCHqp)) { - (void)fprintf(stderr, "Can't read \"%s\", %s\n", - BATCHname, strerror(errno)); - ExitWithStats(1); - } if (QIOtoolong(BATCHqp)) { (void)fprintf(stderr, "Skipping long line in \"%s\".\n", BATCHname); --- 415,420 ---- *************** *** 426,431 **** --- 421,431 ---- (void)QIOread(BATCHqp); continue; } + if (QIOerror(BATCHqp)) { + (void)fprintf(stderr, "Can't read \"%s\", %s\n", + BATCHname, strerror(errno)); + ExitWithStats(1); + } /* Normal EOF. */ break; *************** *** 569,584 **** for (InHeaders = TRUE; ; ) { if ((p = QIOread(qp)) == NULL) { if (QIOerror(qp)) { (void)fprintf(stderr, "Can't read \"%s\", %s\n", Article, strerror(errno)); return FALSE; } - if (QIOtoolong(qp)) { - (void)fprintf(stderr, "Line too long in \"%s\"\n", Article); - (void)QIOread(BATCHqp); - continue; - } /* Normal EOF. */ break; --- 569,584 ---- for (InHeaders = TRUE; ; ) { if ((p = QIOread(qp)) == NULL) { + if (QIOtoolong(qp)) { + (void)fprintf(stderr, "Line too long in \"%s\"\n", Article); + (void)QIOread(qp); + continue; + } if (QIOerror(qp)) { (void)fprintf(stderr, "Can't read \"%s\", %s\n", Article, strerror(errno)); return FALSE; } /* Normal EOF. */ break; *************** *** 982,997 **** Interrupted(Article, MessageID); if ((Article = QIOread(BATCHqp)) == NULL) { - if (QIOerror(BATCHqp)) { - (void)fprintf(stderr, "Can't read \"%s\", %s\n", - BATCHname, strerror(errno)); - ExitWithStats(1); - } if (QIOtoolong(BATCHqp)) { (void)fprintf(stderr, "Skipping long line in \"%s\"\n", BATCHname); (void)QIOread(BATCHqp); continue; } /* Normal EOF -- we're done. */ --- 982,997 ---- Interrupted(Article, MessageID); if ((Article = QIOread(BATCHqp)) == NULL) { if (QIOtoolong(BATCHqp)) { (void)fprintf(stderr, "Skipping long line in \"%s\"\n", BATCHname); (void)QIOread(BATCHqp); continue; + } + if (QIOerror(BATCHqp)) { + (void)fprintf(stderr, "Can't read \"%s\", %s\n", + BATCHname, strerror(errno)); + ExitWithStats(1); } /* Normal EOF -- we're done. */ -- Ian 'Vato' Dickinson [ID17] Kibo bait :-) cudep@csv.warwick.ac.uk ...!uknet!warwick!cudep vato@spuddy.uucp MIME mail welcome - don't send me no steenkin' X.400 Click here.