[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference smurf::buildhelp

Title:USG buildhelp questions/answers
Moderator:SMURF::FILTER
Created:Mon Apr 26 1993
Last Modified:Mon Jan 20 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2763
Total number of notes:5802

1573.0. "Re: markrcs problem in QAQC pool" by AOSG::FILTER (Automatic Posting Software - mail to flume::puck) Fri Jun 23 1995 20:04

Date Of Receipt: 	21-JUN-1995 18:29:06.66
From: 	SMURF::ALPHA::jmf "Joshua M. Friedman OSF/UNIX SDE  21-Jun-1995 1827"
To: 	cnoel@DEC:.zko.alpha
CC: 	odehelp@DEC:.zko.alpha, hantman@DEC:.zko.alpha
Subj: 	Re: markrcs problem in QAQC pool

Chuck, does the log have any successful entries or is the only output
like you've shown?  The markrcs was updated on Jun 19 17:24, and I have
used it without any problems.  Does this coincide with when you started
failing?

The only differences were in command line parsing, and the retry loop
in the mark.  The old markrcs is under /usr/sde/tools/common/BACKUP/markrcs;
you can try that to see if the problem goes away.

What exactly are the arguments on your markrcs command line?  You can
add -debug (this will turn on set -x in the script); and capture the
output in a script we can examine; that might help identify the problem.

-josh

------- Forwarded Message

Return-Path: cnoel 
Delivery-Date: Wed, 21 Jun 95 11:34:39 -0400
Return-Path: cnoel
Received: from alpha.zk3.dec.com by flume.zk3.dec.com; (5.65v3.0/1.1.8.2/16Jan95-0946AM)
	id AA01038; Wed, 21 Jun 1995 11:34:37 -0400
Received: by alpha.zk3.dec.com; (5.65v3.0/1.1.8.2/20May95-1022AM)
	id AA22352; Wed, 21 Jun 1995 11:34:37 -0400
Date: Wed, 21 Jun 1995 11:34:37 -0400
From: Chuck Noel USG <cnoel>
Message-Id: <[email protected]>
To: jmf
Subject: markrcs problem



 Josh,

    Paul told me that you installed a new version of markrcs. I have
  encountered the following problem on the last two nightly builds of
  QAQC.

/usr/sde/tools/alpha_OSF1/bin/markrcs: ./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v: cannot execute
/usr/sde/tools/alpha_OSF1/bin/markrcs: ./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v: cannot execute
/usr/sde/tools/alpha_OSF1/bin/markrcs: ./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v: cannot execute
/usr/sde/tools/alpha_OSF1/bin/markrcs: ./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v: cannot execute
/usr/sde/tools/alpha_OSF1/bin/markrcs: ./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v: cannot execute


      It just goes into a loop and then stops. Could this be related to
  your changes or do you have any ideas what could be causing the problem.

- - Chuck Noel


------- End of Forwarded Message


T.RTitleUserPersonal
Name
DateLines
1573.1Re: markrcs problem in QAQC poolAOSG::FILTERAutomatic Posting Software - mail to flume::puckFri Jun 23 1995 20:1521
Date Of Receipt: 	22-JUN-1995 08:59:13.28
From: 	SMURF::TLAZA::hantman "Paul C. Hantman OSG Test Tools  22-Jun-1995 0834"
To: 	"Joshua Friedman, OSF/UNIX SDE 381-1548" <jmf@DEC:.zko.tlaza>
CC: 	odehelp@DEC:.zko.tlaza, cnoel@DEC:.zko.tlaza, hantman@DEC:.zko.tlaza
Subj: 	Re: markrcs problem in QAQC pool

The only file that markrcs had a problem with was 
./cde/comp_suites/dthelp/shared/TestVolumes/Vol1/C/DC12.xwd,v

I checked the rcs tree, the submit tree and the backing tree and there appeared 
to be nothing special about this file. The permissions were the same as all the 
other files and the owner and group were also consistant.

The first time we noted a problem was when markrcs was run on June 19th 
at 02:01. On th 19th markrcs (the old version) was able to recover, but on the 
run from  June 20th and beyond markrcs got into a loop and did NOT recover.

I will re-run my script this morning using the old markrcs and see what happens.

-Paul

1573.2Re: markrcs problem in QAQC poolAOSG::FILTERAutomatic Posting Software - mail to flume::puckFri Jun 23 1995 20:51179
Date Of Receipt: 	22-JUN-1995 15:49:21.09
From: 	SMURF::FLUME::jmf "Joshua M. Friedman OSF/UNIX SDE  22-Jun-1995 1547"
To: 	hantman@DEC:.zko.flume, cnoel@DEC:.zko.flume
CC: 	odehelp@DEC:.zko.flume
Subj: 	Re: markrcs problem in QAQC pool

Paul/Chuck, here's the PTT report I entered on what I found; I've fixed
the copy of markrcs here, and you should fix your SNAPSHOT file entries
that have spaces instead of a tab, (though the script will handle that
now, other ode commands may still break with this "malformed" SNAPSHOT).
These were probably cases of hand-edited SNAPSHOT entries.  If not, there
some other bug somewhere.

-josh


------- Forwarded Message

Return-Path: [email protected] 
Delivery-Date: Thu, 22 Jun 95 15:39:45 -0400
Return-Path: [email protected]
Received: from abelia.zk3.dec.com by flume.zk3.dec.com; (5.65/1.1.8.2/16Jan95-0946AM)
	id AA24807; Thu, 22 Jun 1995 15:39:43 -0400
Received: from rust.zso.dec.com by wasted.zk3.dec.com; (5.65v3.0/1.1.8.2/18Feb95-1123AM)
	id AA01248; Thu, 22 Jun 1995 15:39:42 -0400
Received: from DECWET.DECnet MAIL11D_V3 by rust.zso.dec.com (5.65/DECwest-CLUSTRIX-cjf-30Jun93)
	id AA25649; Thu, 22 Jun 1995 12:39:40 -0700
Date: Thu, 22 Jun 1995 12:39:40 -0700
Message-Id: <[email protected]>
From: [email protected] (22-Jun-1995 1241 -0700)
To: joshua friedman
Subject: Problem entry confirmation

Your problem report has been entered into the PTT_ODE database as follows:



[This update was sent from JOSHUA FRIEDMAN on 22-JUN-1995 12:41]
[ODE responsible for next action, Problem is undefined]                        

Mail this report to:         [email protected]

Note: ALL CAPS=Mandatory field
Problem Submission Form  Thu Jun 22 14:51:43 EDT 1995
Required Information
====================
YOUR NAME:                   Joshua Friedman
Your Email address:          [email protected]
YOUR PHONE NUMBER:           dtn 381-1548
Your FAX number:             DTN 881-2257

Your company name:           DEC
Your mailstop:               ZKO3-3
YOUR ORGANIZATION'S NAME:    UEG-ZK
YOUR COST CENTER:            335
Your street address:         Digital Equipment Corporation
                             OSF/1 Engineering
                             110 Spit Brook Rd.
                             Nashua, NH 03062 USA
- --------------------------------------
CUSTOMER'S COMPANY NAME:     DEC
CUSTOMER COUNTRY:            USA

- --------------------------------------
PRODUCT:                     ODE
Product version (optional):  V3.0
OS HARDWARE:                 UNIX/ALPHA
OS Hardware version:         347
Workstation soft version (optional): V3.0
COMPONENT:                   SHELL_SCRIPTS
Component version (optional):
Other (optional):
OTHER Version (optional):
SYSTEM WHERE PROBLEM OCCURRED: longbow.zk3.dec.com

CPU type:                    alpha DEC3000_500
Display type:                1 display of type=14
Memory size:
System disk:
Other hardware configuration information:

Other software configuration information:
DEC OSF/1 V3.0 (Rev. 347); Wed Feb  1 11:54:43 EST 1995
DEC OSF/1 V3.0 Worksystem Software (Rev. 347)

- --------------------------------------
PROBLEM ABSTRACT (80 characters or less):
markrcs exits if file locked; temp$$ not rm'd; bad SNAP entries caused sed errs

PROBLEM PRIORITY (1-5, 1 being most critical): 1
CAN THE CUSTOMER REPRODUCE THE PROBLEM [Y/N]:  Y
CAN YOU REPRODUCE THE PROBLEM [Y/N]:           Y


PROBLEM STATEMENT:
=================
There are really three problems here.

    1) The retry while loop for the "rcs -N" command quits out with an
    "exit 1", which causes the whole markrcs to terminate.  This should
    quit out of the loop with a "break" instead; this has been tested
    (see below).

    2) The temporary file "temp$$" produced in the src directory of
    the rcs pool is never removed; these could consume a lot of space
    over time and should be removed, however they were useful for
    debugging the next problem.

    3) The QAQC pool discovered that some of its files get "cannot
    execute" errors.  I've cut out some of the lines which cause this
    error, and the problem is that the 'sed' failes to convert the
    SNAPSHOT entry to the "rcs -N" command.  Below is a test case.
    This is not new with the new markrcs, but it was noticed because
    of the abort out (problem 1).

    New information - these lines were "malformed" in the SNAPSHOT -
    they had spaces after ,v instead of a plain <tab>.  "GIGO".
    I don't know if the sed should be clever enough to allow this,
    but I proposed a syntax.


Instructions to reproduce the problem:
======================================

Problem 1)
- ----------
Here's a test SNAPSHOT file:

    devrcs> cat TEST
    ./foobar,v      1.1
    ./Makefile,v    1.1

Note that in the current markrcs, (with RETRY_LIMIT=2), ./Makefile
never gets marked:

    devrcs> /tmp/markrcs -snap TEST TEST
    markrcs starting marking at this time Thu Jun 22 14:56:27 EDT 1995
    rcs error: RCS/foobar,v: No such file or directory
    RCS error. Going to retry for the 1 time.
    rcs error: RCS/foobar,v: No such file or directory
    error: markrcs failed. File seems to have a permanent lock.
    markrcs completed at Thu Jun 22 14:56:33 EDT 1995

Now changing "exit 1" to "break":

    devrcs> /tmp/markrcs -snap TEST TEST
    markrcs starting marking at this time Thu Jun 22 14:58:38 EDT 1995
    rcs error: RCS/foobar,v: No such file or directory
    RCS error. Going to retry for the 1 time.
    rcs error: RCS/foobar,v: No such file or directory
    error: markrcs failed. File seems to have a permanent lock.
    RCS file: Makefile,v
    done
    markrcs completed at Thu Jun 22 14:58:43 EDT 1995


Problem 2)
    n/a

Problem 3)
    # cat test
    ./foo,v 1.1.2.2  	/* tab */
    ./bar,v 1.1.2.2	/* space */

    # sed s;^\./\(.*\),v	\(.*\);rcs -NTEST:\2 \1; test
                           ^-<tab>
    rcs -NTEST:1.1.2.2 foo
    ./bar,v 1.1.2.2

Solution, if you choose to use it:
    sed  's;^\./\(.*\),v[	 ]*\(.*\);rcs -N'${tag}':\2 \1;' test
                         ^-<tab><space>


- --------------------------------------

------- End of Forwarded Message