Conference smurf::buildhelp

154.0. "bsubmit core dump & bad header.msg?" by SMURF::FILTER (Automatic Posting Software - mail to flume::puck) Thu Jun 17 1993 19:18

Date Of Receipt: 	17-JUN-1993 17:54:30.15
From: 	WASTED::jmf "Joshua M. Friedman ULTRIX SDE  17-Jun-1993 1741"
To: 	unxa::menner
CC: 	odehelp@wasted:zko.dec, rtm@wasted:zko.dec, walker@wasted:zko.dec
Subj: 	bsubmit core dump & bad header.msg?

Ron, when you bcreated this file, somehow it got a comment leader of "$"
instead of "$ ".  (Do a blog on the file and you'll see:)
	comment leader: "$"

This has confused bci/bsubmit, since it expects to find an exact match between
the header of the file and what it expects.  

You can change the comment leader using
	bcs -c"$ " filename

and just try the submit again.  My understanding is that it should NOT
core dump; this bug was fixed and we have the patches, but that's for
decwest to think about.  If you're doing this from new jersey, Rich may
not have the latest bug fixes which would have prevented this core dump,
but still required that the "$" be changed to "$ ", with a message like:

mossey.zso.dec.com> bci -m"test" -all

[ ./bin/main.c ]
[ ' *' is the RCS comment leader set for ./bin/main.c ]
[ ' * ' is the comment leader found in the file ./bin/main.c ]
[ Comment leaders MUST be an EXACT MATCH between RCS and the working file. ]
[ Use bcs -c' * ' to set the comment leader in RCS ]


-josh


------- Forwarded Messages

Return-Path: [email protected]
Received: by abyss.zk3.dec.com (5.65/DEC-USSG-ZK3-ULTRIX-09/27/91);
	id AA27247; Thu, 17 Jun 1993 16:59:13 -0400
Date: Thu, 17 Jun 1993 16:59:13 -0400
From: [email protected] (Ron Menner OSG/NJ)
Message-Id: <[email protected]>
To: [email protected]
Subject: found the culprit file...
Cc: [email protected]

Hi-


	The culprit file is usr/bin/eucset/eucset.msg (the same file i
was having problems with the comment leader).  i successfully submitted
all my kernel files and usr/bin/eucset/eucset.c (which shouldn't harm
the build 'cuz no Makefiles were submitted).  Guess the thing to do is
save and outdate eucset.msg then try to bcreate it.  For the sake of
doco, here's the bsubmit panic once again:

$ bsubmit -defect agosminor-329-rtm eucset.msg

*** No user interaction will be required during the merge step. ***

/usr/sde/ode2.0/tools/alpha_ace/bin/bsubmit[198]: 1612 Segmentation
fault(coredump)

	i won't do anything until i hear from all you ode gurus.


			thx...
				/ron
 
------- Forwarded Messages

Return-Path: [email protected]
Received: by abyss.zk3.dec.com (5.65/DEC-USSG-ZK3-ULTRIX-09/27/91);
	id AA27538; Tue, 15 Jun 1993 10:20:04 -0400
Date: Tue, 15 Jun 1993 10:19:57 -0400
Message-Id: <[email protected]>
From: [email protected] (Pseudo Sick)
To: [email protected]
Cc: [email protected]
Subject: bad header.msg?

Hi again-

	i'm trying to check in a new msg file (usr/bin/eucset/eucset.msg).
i've grapped the header.msg from /usr/sde/alpha/headers.  But when i try to
check the file in, i get:

>>Warning in bci:
>>Improper header line between HISTORY marker and $LOG marker in eucset.msg
>>Line:6


Here's what Line 6 contains:

$ $Log:$

And this is the "stock" header.msg w/o any history comments between HISTORY
and $Log.  The sandbox source lives at:
 /usr/projects/netode3/rtm/sb/check_in/src/usr/bin/eucset/eucset.msg

The setname is Ronald_Menner_check_in

		as always any help much appreicated...

				/ron

btw: i'm in NJ this week.

------- Message 2

Return-Path: [email protected]
Received: by abyss.zk3.dec.com (5.65/DEC-USSG-ZK3-ULTRIX-09/27/91);
	id AA19100; Thu, 17 Jun 1993 15:32:18 -0400
Date: Thu, 17 Jun 1993 15:32:18 -0400
From: [email protected] (Ron Menner OSG/NJ)
Message-Id: <[email protected]>
To: [email protected]
Subject: bsubmit seg faults.....
Cc: [email protected]

 
	bsubmit seg faulted right smack in the middle of my submit.
Here's a log of what happened:


$ bsubmit -defect agosminor-329-rtm -all
NFS server bigbld not responding still trying
NFS server bigbld ok

*** No user interaction will be required during the merge step. ***

/usr/sde/ode2.0/tools/alpha_ace/bin/bsubmit[198]: 1155 Segmentation
fault(coredump)


	i don't think the source pool was corrupted but i will verify.  
Does this have anything to do with removal of the au* backing trees.
(the eucset stuff i'm submitting originally came from OSF R1.2 and was 
rcs'd in the au backing trees.  After verifying that there is no source
corruption should i resubmit?

			/ron_who_is_running_scared!



------- End of Forwarded Messages

Date Of Receipt: 	17-JUN-1993 21:03:08.63
From: 	ABYSS::unxa::menner
To: 	UNXA::ABYSS::jmf
CC: 	ABYSS::ODEHELP, abyss::MENNER
Subj: 	RE: bsubmit core dump & bad header.msg?

Josh et al,

	Changing the bcs header from "$" to "$ " did the trick - mucho
thanks.  For the record, i am working in NJ but the bsubmit panic happened 
on a ZKO machine (which supposely has the patches).

			thanks again...
					/ron

Date Of Receipt: 	20-JUL-1993 10:13:51.30
From: 	LOCORE::"[email protected]"
To: 	[email protected], [email protected]
CC: 	[email protected]
Subj: 	Bsubmit core dump
Someone has reported that bsubmit aborted with the error: 	

		Segmentation fault(core dump)

on a system running Maintenance SSB. Has anyone else reported
a similar problem?


Thanks,

Rich
----------------------------------------------------------------
Rich Larsen, M/S: UNX			TCP/IP: [email protected]
USSG/User Env. & Std. Group		DECnet: UNXA::LARSEN
Digital Equipment Corporation		FAX:	908-577-6003
200 Route 9 North			Voice:	908-577-6083	
Manalapan, New Jersey 07726		DTN:	462 

Date Of Receipt: 	22-JUL-1993 17:11:02.62
From: 	WASTED::jmf "Joshua M. Friedman ULTRIX SDE  22-Jul-1993 1709"
To: 	"Rich Larsen (908) 577-6083 DTN 462-6083 [email protected]" <[email protected]>
CC: 	[email protected], [email protected], [email protected],
	jmf@wasted:zko.dec
Subj: 	Re: Bsubmit core dump

Rich, please report the details of the core dump via "gen-ptt", run,
if possible, in a workon session on the actual sandbox on the system where 
the error occurred.  Provide as much information about what was being
done in the submit, especially anything out of the normal "hands off"
submit.  Put the core file in a place that is accessible or at least
protected from accidental deletion, and include this location in the
ptt report.

thanks...		-josh

(there was a bmerge core dump problem which was recently found & fixed, but
I doubt this is related)

thanks...	-josh