From nobody@FreeBSD.org  Tue Oct 30 19:58:56 2001
Return-Path: <nobody@FreeBSD.org>
Received: from freefall.freebsd.org (freefall.FreeBSD.org [216.136.204.21])
	by hub.freebsd.org (Postfix) with ESMTP id CA41D37B406
	for <freebsd-gnats-submit@FreeBSD.org>; Tue, 30 Oct 2001 19:58:55 -0800 (PST)
Received: (from nobody@localhost)
	by freefall.freebsd.org (8.11.4/8.11.4) id f9V3wtA62468;
	Tue, 30 Oct 2001 19:58:55 -0800 (PST)
	(envelope-from nobody)
Message-Id: <200110310358.f9V3wtA62468@freefall.freebsd.org>
Date: Tue, 30 Oct 2001 19:58:55 -0800 (PST)
From: Lyndon Griffin <lyndon@bsd4us.org>
To: freebsd-gnats-submit@FreeBSD.org
Subject: man(1) dumps core with groff/troff seg fault
X-Send-Pr-Version: www-1.0

>Number:         31650
>Category:       gnu
>Synopsis:       man(1) dumps core with groff/troff seg fault
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue Oct 30 20:00:00 PST 2001
>Closed-Date:    Thu Nov 1 01:09:36 PST 2001
>Last-Modified:  Thu Nov 01 01:13:56 PST 2001
>Originator:     Lyndon Griffin
>Release:        4.4-RELEASE
>Organization:
bsd4us.org
>Environment:
FreeBSD gemini.bsd4us.org 4.4-RELEASE FreeBSD 4.4-RELEASE #3: Sat Oct 20 15:53:44 EDT 2001     root@gemini.bsd4us.org:/usr/src/sys/compile/GEMINI  i386
>Description:
man(1) dumps core on some manpages (I believe these are not preformatted pages).  groff(1) also dumps core with any (or no) options.  The specific error is a segmentation fault from groff: troff.  Backtrace shows in main()->free()->free().  When attempting to recompile groff with symbols, the problem disappears.  Subsequent builds did not restore the problem.
>How-To-Repeat:
Install clean from a 4.4-RELEASE cdrom.  Run man(1) on a manpage for which a preformatted version does not exist on disk.
>Fix:
Compile and install groff from /usr/src/gnu/usr.bin/groff.  It appears that the source code on the 4.4-RELEASE cdrom does not contain the bug.
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->closed 
State-Changed-By: johan 
State-Changed-When: Wed Oct 31 02:46:48 PST 2001 
State-Changed-Why:  
As you say the problem does not appear in  
the src after release. We can not do anything about 
the release after the release. Hence problem solved. 

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=31650 

From: Ruslan Ermilov <ru@FreeBSD.ORG>
To: Lyndon Griffin <lyndon@bsd4us.org>
Cc: freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: gnu/31650: man(1) dumps core with groff/troff seg fault
Date: Wed, 31 Oct 2001 14:43:11 +0200

 On Tue, Oct 30, 2001 at 07:58:55PM -0800, Lyndon Griffin wrote:
 > 
 > man(1) dumps core on some manpages (I believe these are not preformatted pages).
 > groff(1) also dumps core with any (or no) options.
 > The specific error is a segmentation fault from groff: troff.
 > Backtrace shows in main()->free()->free().
 > When attempting to recompile groff with symbols, the problem disappears.
 > Subsequent builds did not restore the problem.
 > 
 What are the sizes and MD5 checksums of /usr/bin/groff and /usr/bin/troff?
 
 
 Cheers,
 -- 
 Ruslan Ermilov		Oracle Developer/DBA,
 ru@sunbay.com		Sunbay Software AG,
 ru@FreeBSD.org		FreeBSD committer,
 +380.652.512.251	Simferopol, Ukraine
 
 http://www.FreeBSD.org	The Power To Serve
 http://www.oracle.com	Enabling The Information Age
State-Changed-From-To: closed->open 
State-Changed-By: johan 
State-Changed-When: Wed Oct 31 07:24:29 PST 2001 
State-Changed-Why:  
Ok, there seems to be a real problem which should be mentioned in  
the ERRATA. 

=== 
From: "Lyndon Griffin" <lyndon@bsd4us.org> 
To: Ruslan Ermilov <ru@FreeBSD.ORG> 
Cc: johan@FreeBSD.ORG 
Subject: Re: gnu/31650: man(1) dumps core with groff/troff seg fault 
Date: Wed, 31 Oct 2001 14:52:20 GMT 

OK - here's a problem.  The bin.bq on the 4.4-RELEASE ISO does not pass an 
md5 test:  CHECKSUM.MD5 contains 
... 
MD5 (bin.bp) = d095859893d4f170b748f6253a04eb8f 
MD5 (bin.bq) = 8994c5cf77a2f9bcf34dbb02516784c1 
MD5 (bin.br) = b48c94797fa6fe4885945e0f4c1c4cc8 
... 

while the real deal is: 
betelgeuse:/cdrom/bin/$ md5 bin.bp 
MD5 (bin.bp) = d095859893d4f170b748f6253a04eb8f 
betelgeuse:/cdrom/bin/$ md5 bin.bq 
MD5 (bin.bq) = 9f6098cc2f5505d6e5c62e67d0c252cf 
betelgeuse:/cdrom/bin/$ md5 bin.br 
MD5 (bin.br) = b48c94797fa6fe4885945e0f4c1c4cc8 

=== 

On ftp.se.freebsd.org 'bin.bq' passes a md5 checksum test 
hence something seems to be wrong with at least some CDs. 

Leaving this open until this or something else has been  
added to the ERRATA. 


http://www.FreeBSD.org/cgi/query-pr.cgi?pr=31650 

From: Johan Karlsson <k@numeri.campus.luth.se>
To: Lyndon Griffin <lyndon@bsd4us.org>
Cc: freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: gnu/31650: man(1) dumps core with groff/troff seg fault
Date: Wed, 31 Oct 2001 19:06:42 +0100

 Yes, it does seem like a bigger problem.
 Where did you get your CD set from,
 did you download the ISO images and burned your own CDs
 or did you buy them from some place?
 
 If you downloaded them can you please verify that the md5
 checksum is correct for the images.
 According to CHECKSUM.MD5 on ftp.se.freebsd.org:
 MD5 (4.4-disc2.iso) = 87c51fae8aa5e448243fd53a9ed70627
 MD5 (4.4-disc3.iso) = bef91f95beeae551412e1ea99ffc9fd1
 MD5 (4.4-disc4.iso) = f7e1ee195735d185d91682147e2e3ffd
 MD5 (4.4-mini.iso) = 33f1812cd11a38fab1f42d9399b30b60
 MD5 (4.4-install.iso) = 7dc237940080379760f9deac29f08384
 
 
 /Johan K
 
 Ps please cc freebsd-gnats-submit@FreeBSD.ORG any reply, so our PR database
 will record your mails. Be sure to use the same subject.
 
 
 On Wed, Oct 31, 2001 at 16:58 (+0000) +0000, Lyndon Griffin wrote:
 > Final output:
 > 
 > betelgeuse:/cdrom/$ md5check.pl
 > file bin/bin.bq does not match
 > file packages/All/./gal-0.11.2.tgz does not match
 > file src/scontrib.cg does not match
 > file XF86336/PC98-Servers does not match
 > file XF86336/Servers does not match
 > file XF86336/Xlkit.tgz does not match 
 > 
 > This looks like more than a GNU/troff problem... 
 > 
 > Lyndon Griffin writes: 
 > 
 > > johan@FreeBSD.org writes: 
 > > 
 > >> Synopsis: man(1) dumps core with groff/troff seg fault  
 > >> 
 > >> State-Changed-From-To: closed->open
 > >> State-Changed-By: johan
 > >> State-Changed-When: Wed Oct 31 07:24:29 PST 2001
 > >> State-Changed-Why:  Ok, there seems to be a real problem which should be 
 > >> mentioned in  the ERRATA.  
 > >> 
 > > 
 > > I threw together a PERL script to check all the published MD5's against 
 > > their real values.  So far, I've had one other hit besides bin.bq, which 
 > > I've verified is incorrect.  The hit is packages/All/gal-0.11.2.tgz:
 > > published:  MD5 (./gal-0.11.2.tgz) = 6247063a176ff2d5ca640957758a8779
 > > actual:  MD5 (./gal-0.11.2.tgz) = 58f75ee41f30999d3730ce4256fd3062  
 > > 
 
 -- 
 Johan Karlsson		mailto:k@numeri.campus.luth.se

From: "Lyndon Griffin" <lyndon@bsd4us.org>
To: Johan Karlsson <k@numeri.campus.luth.se>
Subject: Re: gnu/31650: man(1) dumps core with groff/troff seg fault
Date: Wed, 31 Oct 2001 20:33:31 GMT

OK - now I'm feeling stupid.  Those checksums don't match any of my
images...  I pulled the ISO images from ftp5.freebsd.org about 3 days after
the release announcement.  Here's what I get

MD5 (4.4-disc2.iso) = 8a6265ed197df46f530058fda98e97c8
MD5 (4.4-disc3.iso) = f58d6436222a217fe90cc7d39267e7df
MD5 (4.4-disc4.iso) = f7e1ee195735d185d91682147e2e3ffd
MD5 (4.4-install.iso) = 98db423a291297a7ebcdcdcf67fe7413
MD5 (4.4-mini.iso) = aac83f38474d7ba2326b198a9acbf03a

gemini:/mnt/$ ll *.iso
 -rwxr-xr-x  1 root  wheel  627048448 Sep 22 00:56 4.4-disc2.iso*
 -rwxr-xr-x  1 root  wheel  658374656 Sep 23 01:14 4.4-disc3.iso*
 -rwxr-xr-x  1 root  wheel  667090944 Sep 22 10:34 4.4-disc4.iso*
 -rwxr-xr-x  1 root  wheel  664305664 Sep 21 13:49 4.4-install.iso*
 -rwxr-xr-x  1 root  wheel  184549376 Sep 23 01:15 4.4-mini.iso*
 
The file sizes on ftp5.freebsd.org do not match mine, either.
 -rw-r--r--  1 mirror  mirror  627048448 Sep 19 07:49 4.4-disc2.iso
 -rw-r--r--  1 mirror  mirror  658374656 Sep 19 03:58 4.4-disc3.iso
 -rw-r--r--  1 mirror  mirror  667090944 Sep 18 20:05 4.4-disc4.iso
 -rw-r--r--  1 mirror  mirror  667222016 Sep 22 02:36 4.4-install.iso
 -rw-r--r--  1 mirror  mirror  187072512 Sep 22 03:17 4.4-mini.iso

I watched the file transfers...  I'm quite certain that they transfered OK.
It's very, very interesting to me that the file sizes on discs 2, 3, and 4
are the same, while disc 1 and mini (the two that would presumably have this
problem) have changed.  It's also interesting that in both of those cases,
the new file is 3mb larger.  Also note that my timestamp on the install disc
is OLDER than the timestamp on the current release.

Well, looks to me like I grabbed a bad ISO image.  I'm pulling a new one
down, now.  I would recommend closing this PR out, unless you think
significant numbers of other people got hold of a bad ISO, too.

Thanks for your help.


State-Changed-From-To: open->closed 
State-Changed-By: johan 
State-Changed-When: Thu Nov 1 01:09:36 PST 2001 
State-Changed-Why:  
Submitter reports that his images do not 
pass a md5 checksum test. And concludes  
the PR should be closed. 


http://www.FreeBSD.org/cgi/query-pr.cgi?pr=31650 
>Unformatted:
