Usenet 1994 January/usenetsourcesnewsgroupsinfomagicjanuary1994.iso/sources/misc/volume26/QBATCH/patch03a

Text File | 1991-11-29 | 13.0 KB | 340 lines

Newsgroups: comp.sources.misc
From: alan@tharr (Alan Saunders)
Subject:  v26i085:  QBATCH - a queued batch processing system for UNIX, Patch03a/2
Message-ID: <csm-v26i085=[e-mail]>
X-Md4-Signature: 10a79651645b139542b22099da7f4e79
Date: Sat, 30 Nov 1991 03:20:05 GMT
Approved: [e-mail]

Submitted-by: alan@tharr (Alan Saunders)
Posting-number: Volume 26, Issue 85
Archive-name: QBATCH/patch03a
Environment: UNIX
Patch-To: QBATCH: Volume 25, Issue 20-25

Well, the mail seems to have tailed off a fair amount now, so I'm assuming
that all the glaring holes have been found.

The only outstanding problems I am aware of are with a user who doesn't have
news or mail access, and I can't get the patches to him 'til next week.  It
may be that the latest version solves his problems, but I'll state them
briefly here:

HP 9000/800 HP/UX 8.00 After a system reboot (with a job in the queue) the
job restarts and completes but qp hangs without apparently recognising that
the child has terminated.  Seems to clear up ok when another job is
submitted.

Interactive 386/IX  Version 2.2 seems ok but systems in 2.01 just seem to
hang.

If anyone has ported QBATCH to either of these environments or can give any
help, please contact me.

This is qbatch patch 03. Issued 22 Nov 1991.

Change directory to the root of the qbatch hierarchy (where you originally
copied and unpacked the shar files and the patches.)  Unshar BOTH parts of
this submission and apply the patches as:

patch -p0 < patches/qbatch.p3

(You may want to read the notes in src/config.ARCH before making).

The following points are addressed in this patch.

change process priority (nice) to be BEFORE seutid.  (non root uid's can't
increase priority (reduce nice value)!!)

This was reported in comp.sources.bugs.  If you manually changed qp.c, you
will need to re-instate the patch 2 source before applying this patch.

Changed occurrences of cc to $(CC) to allow definition of compiler.
(compiles with gcc vn 1.39)

Added qbatch.h dependency to logging.o and time.o to force recompilation
after changing Makefile.

Pointed out to me by Michael Hamilton <[e-mail]> that POSIX
conforming cuserid() returns the user name relating to the EFFECTIVE user id
and not the logged in user id!  Since js is setuid root, the entry.qe_uname
displayed in ql, and used by jobdone (and used to send mailed monitors to) is
always root on a POSIX conformant system.

added 3 new defines to config.h, and related code to config.c to define the
source of the logged in user name.

Tidied up sources a bit so that lint doesn't complain so much.

Added facility for using predefined configurations defined in a new file
config.ARCH.  If users send me their configurations when they have QBATCH
working, I can include them in this file and we can forestall much hacking of
config.h for future users on the same architecture.

Fixed a 'tail chasing dog' in config.h.  If a system needs ???_t defining, it
needs it first of all in qbatch.h.  qbatch.h is included before config.h
since the latter needs some of the defines in qbatch.h???

Since I want to keep all architecture variables in the same file (config.h)
I've added an #ifdef for __QBATCH_H_ around the typedefs and #included
config.h in qbatch.h.  .. Sounds convoluted but it does what I want!

Implemented specifying monitors as pipes.

Found bug! js was only checking enviroment for MONITOR if queue default
monitor was NONE.  Changed this and made rejection message more informative.

Added leading zero to patchlevel .. at the rate this is going, I'll need it!

Regards ..Alan
#! /bin/sh
# This is a shell archive.  Remove anything before this line, then unpack
# it by saving it into a file and typing "sh file".  To overwrite existing
# files, type "sh file -c".  You can also feed this as standard input via
# unshar, or by typing "sh <file", e.g..  If this archive is complete, you
# will see the following message at the end:
#        "End of archive 1 (of 2)."
# Contents:  PLEASE src/config.ARCH
# Wrapped by root@vfib_d on Fri Nov 22 12:15:09 1991
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test -f 'PLEASE' -a "${1}" != "-c" ; then
  echo shar: Will not clobber existing file \"'PLEASE'\"
else
echo shar: Extracting \"'PLEASE'\" \(3761 characters\)
sed "s/^X//" >'PLEASE' <<'END_OF_FILE'
X[
X  I started to write this to add to the header of qbatch patch 3.  It struck me
X  whilst writing it that it actually applies to any package that has been
X  posted to the net.  Personally, I have been pleasantly surprised and
X  encouraged by the response I have received, but there are cases where
X  users have sent me queries, then nothing further.  I don't know in which
X  of the following categories they now belong.  I am sure that other software
X  authors have the same experience.
X
X  I'm therefore posting it as a separate message in comp.sources.d.  As it
X  stands it is related to QBATCH.  Just substitute 'product/program/package'
X  for 'QBATCH' and 'author(s)' for 'me/i/' and see if it applies to you.
X
X  Alan .. 22 Nov 1991
X]
In the absence of evidence to the contrary, (and assuming that you have any
interest at all in QBATCH) I can only assume one of the following
possibilities:
X
X    1- You got it all working by just changing definitions in config.h.
X
X        2- You had to make changes to the released version and patches to
X       get it working.
X
X    3- You got it working and have in some way altered it's functionality
X       (by adding facilities enhancements etc.)
X
X    4- You can't get it working at all.
X
X    5- You haven't got time to work on it now.
X
X1- If you haven't already done so please send me details of your configuration.
X   The easiest way is :
X   grep '^#define' config.h >file
X   grep '^typedef' config.h >>file
X   If your machine supports uname then :
X   uname -a >> file
X   Edit the file if necessary to provide full information on system type,
X   operating system and level, and compiler (if not native) and compiler
X   switches.
X   Then please mail it to me.
X
X2- Please mail me the diffs between your sources and mine (at  the latest
X   patch level) If the changes are necessary to get it working, others need to
X   know.
X
X3- If you've enhanced the functionality of QBATCH, others may want to share.
X   As of patch level 3 I've implemented all suggestions received (except a
X   request for multiple concurrent jobs in a queue which would have belied
X   the nature of a queued batch system).
X
X   If the enhancement is one or more standalone programs, then why not post
X   them to comp.sources.misc (but also mail me a copy so that I can include
X   them in the distribution).
X   If it involves changes to the existing sources, then please mail me the
X   diffs so that I can incorporate it at source distribution level.  I can then
X   issue a patch to bring all users into line.
X
X   It is to your own advantage to do this, as patches and enhancements
X   issued by by me will ALWAYS be based on the latest patch level as issued.
X   If your sources differ through local changes, then you will have to patch
X   by hand to take advantage of them.
X
X4- Contact me!  Someone else may have got it working on the same platform,
X   and has let me know.  I may have the info you want.
X   Please include ALL the information that you can regarding the problem:
X
X   The system you are trying to build on
X   The operating system, version and level.
X   The compiler, make, etc you are using to build (if not system standard)
X   error messages produced by 'make'.
X
X   If I can help I will do so, if I can't I'll let you know, and you can
X   post a query onto the net (someone will have the answer).
X
X5- If you feel the QBATCH be useful, try to find the time.
X   One sees messages on the net of the form:
X   "I'm trying to contact the author of so-and-so', I've got problems."
X   If it's got to this stage, it could be too late.
X   If you find problems, then the sooner they're sorted out the better.
X   If not for yourself, then for the person less experienced than yourself
X   who may have less work to do when you send me your configuration details.
END_OF_FILE
if test 3761 -ne `wc -c <'PLEASE'`; then
    echo shar: \"'PLEASE'\" unpacked with wrong size!
fi
# end of 'PLEASE'
fi
if test -f 'src/config.ARCH' -a "${1}" != "-c" ; then
  echo shar: Will not clobber existing file \"'src/config.ARCH'\"
else
echo shar: Extracting \"'src/config.ARCH'\" \(3846 characters\)
sed "s/^X//" >'src/config.ARCH' <<'END_OF_FILE'
X/************************************************************************
X*
X* This file is designed to help new users of QBATCH to get it compiled
X* on their system.  It comprises the #defines in config.h claimed to
X* give a working QBATCH on various systems.
X*
X* If you get QBATCH working on your system, please send details to me
X* in the following format, so that it can be included in the release.
X*
X* The easiest way to do this is:
X* uname -a           >    <filename>
X* grep '^#define' config.h >>    <filename>
X* grep '^typedef' config.h >>    <filename>
X*
X* edit the resulting <filename> to enclose the architecture info as a quote
X* and to bracket the entire group with an #ifdef/#endif on a name preferrably
X* not in the following list.  Then send it to me.
X*
X* Also included are any comments relating to oddballs in configuration.
X*
X* Current known working configurations as at 18 Nov 1991:
X*
X* To use one of these predefined configurations, #define one of the following
X* architectures in config.h on the line BEFORE this file is #included.
X*
X* #define        Source and Platform.
X*
X* SUNSPARC.411            Me.
X*            SunOS 4.1 4.1.1 on Sun4 Sun4C (Sparc)
X*
X* DEC3100.U40        Bryan        <bryan%sphinx.uucp.uunet.uu.net>
X*            Ultrix 4.0 on DECstation 3100
X*
X* RS6000.AIX315        Michael Gray <[e-mail]>
X*            AIX 3.1.5 on IBM RS6000
X*           and  Michael Hamilton <[e-mail]>
X*            (os version unstated) on IBM RS6000
X*
X************************************************************************/
X
X#ifdef SUNSPARC.411
X/* SunOS 4.1.1 version 2, Architecture sun4, sun4c (sparc) */
X#ifndef __QBATCH_H_
X#define ARCH_DEFINED
X#define XPG3PGP        /* setpgid() and kill() with negative pid    XPG3 */
X#define FCNTL        /* use fcntl() calls for file locking (XPG3) */
X#define SIGACTION    /* XPG3 */
X#define WAIT_XPG3    /* XPG3 */
X#define QCLOCK_TICK sysconf(_SC_CLK_TCK);
typedef void HANDLER_TYPE;
typedef int WAIT_TYPE;     /* XPG3 */
X#define Q_UNAME_PWD
X#else /* __QBATCH_H_ */
X/* include any of the ???_t typedefs here */
X#endif /* __QBATCH_H_ */
X/*
X* Comments:
X* This is the original release platform.  If you get a config.h containing
X* different definitions, it'll be because I have been testing.  Several
X* combinations will actually work on this platform, but the above are the
X* recommended ones.
X*/
X#endif /* SUNPARC.411 */
X
X#ifdef DEC3100.U40
X/* DECstation 3100 running Ultrix 4.0 */
X#ifndef __QBATCH_H_
X#define ARCH_DEFINED
X#define XPG3PGP        /* setpgid() and kill() with negative pid    XPG3 */
X#define FCNTL        /* use fcntl() calls for file locking (XPG3) */
X#define SIGPAUSE
X#define SIGNAL
X#define WAIT_XPG3    /* XPG3 */
X#define QCLOCK_TICK sysconf(_SC_CLK_TCK);
X#define Q_UNAME_PWD
X#else /* __QBATCH_H_ */
X/* include any of the ???_t typedefs here */
X#endif /* __QBATCH_H_ */
X/*
X* Comments:
X* Apparently requires a change to src/rcgen line 2:
X* From: if  [ ${ech:-\c} = "A-n" ]
X* to:   if  [ ${ech--\c} = "A-n" ]
X*                  ^^
X*/
X#endif /*ifdef DEC3100.U40 */
X
X#ifdef RS6000.AIX315
X/* IBM rs6000 running AIX version 3.1.5 */
X#ifndef __QBATCH_H_
X#define ARCH_DEFINED
X#define XPG3PGP        /* setpgid() and kill() with negative pid    XPG3 */
X#define FCNTL        /* use fcntl() calls for file locking (XPG3) */
X#define SIGACTION    /* XPG3 */
X#define WAIT_XPG3    /* XPG3 */
X#define QCLOCK_TICK sysconf(_SC_CLK_TCK);
typedef void HANDLER_TYPE;
typedef int WAIT_TYPE;     /* XPG3 */
X#define Q_UNAME_PWD
X#else /* __QBATCH_H_ */
X/* include any of the ???_t typedefs here */
X#endif /* __QBATCH_H_ */
X/*
X* Comments:
X* Michael found a bug in qp.c which his version ifdef'ed round. This will be
X* fixed in patch 3.
X* Also I hope to resolve a query relating to the need for '&' when structures
X* are passed as parameters to functions.
X* Given that This can be resolved, should be ok without any extra defines.
X* resolved 21 Nov 91 changes incorporated in p3
X*/
X#endif /* RS6000.AIX315 */
END_OF_FILE
if test 3846 -ne `wc -c <'src/config.ARCH'`; then
    echo shar: \"'src/config.ARCH'\" unpacked with wrong size!
fi
# end of 'src/config.ARCH'
fi
echo shar: End of archive 1 \(of 2\).
cp /dev/null ark1isdone
MISSING=""
for I in 1 2 ; do
    if test ! -f ark${I}isdone ; then
    MISSING="${MISSING} ${I}"
    fi
done
if test "${MISSING}" = "" ; then
    echo You have unpacked both archives.
    rm -f ark[1-9]isdone
else
    echo You still need to unpack the following archives:
    echo "        " ${MISSING}
fi
##  End of shell archive.
exit 0

exit 0 # Just in case...
--
Kent Landfield                   INTERNET: [e-mail]
Sterling Software, IMD           UUCP:     uunet!sparky!kent
Phone:    [phone]         FAX:      [phone]
Please send comp.sources.misc-related mail to [e-mail].