ALLOC Technique
Mosley, George
I'm wondering if anyone out there has a technique for handling the following
file allocation task in one step:
If the file exists, I want to allocate as OLD, i.e. I want to write over any
existing content. If the file doesn't exist, I want to allocate as NEW.
I currently do it this way:
msgstat = MSG('off')
"DELETE ('"BTS_session"')"
msgrset = MSG(msgstat)
"ALLOC F(BTSOUT) DA('"BTS_Session"') NEW ,
UNIT(SYSDA) TRACKS SP(60,50) ,
BUFNO(15) LRECL(133) ,
RECFM(F B A) DSORG(PS)"
Is there some way I can drop the DELETE?
Thanks,
George
----------------------------------------------------------------------
For TSO-REXX subscribe / signoff / archive access instructions,
send email to LIST...@VM.MARIST.EDU with the message: INFO TSO-REXX
Valone, Toren W.
existing data with a SHR if the file does not exist, allocate it and write
to it. Don't think you can do it in one step.
Frank Clarke
wrote:
<E478EF9AB22CD3119B44...@ex01mb299.icbc.net>
>I currently do it this way:
>
>msgstat = MSG('off')
>"DELETE ('"BTS_session"')"
>msgrset = MSG(msgstat)
>
>"ALLOC F(BTSOUT) DA('"BTS_Session"') NEW ,
> UNIT(SYSDA) TRACKS SP(60,50) ,
> BUFNO(15) LRECL(133) ,
> RECFM(F B A) DSORG(PS)"
>
>Is there some way I can drop the DELETE?
fb133.0 = "NEW CATALOG UNIT(SYSDA) SPACE(60 50) TRACKS",
"RECFM(F B A) LRECL(133) BLKSIZE(0)"
fb133.1 = "SHR" /* if it already exists... */
tempstat = Sysdsn(outdsn) = "OK" /* 1=exists, 0=missing */
"ALLOC FI(BTSOUT) DA("outdsn") REU" fb133.tempstat
Jeremy C B Nicoll
Frank Clarke <ni...@MINDSPRING.COM> wrote:
> tempstat = Sysdsn(outdsn) = "OK" /* 1=exists, 0=missing */
In some cases testing to see if: sysdsn(dsn) equals "OK" is a bad idea,
because sysdsn() can return a bunch of other values admittedly in
strange circumstances. I think for example there's a no-access to
dataset message. It's sometimes safer to test for sysdsn() not being
equal to "DATASET NOT FOUND".
--
Jeremy C B Nicoll - my opinions are my own.
Frank Clarke
Nicoll) wrote:
<4c01ab41...@omba.demon.co.uk>
>In article <kv3ievsomufoijn15...@4ax.com>,
> Frank Clarke <ni...@MINDSPRING.COM> wrote:
>
>> tempstat = Sysdsn(outdsn) = "OK" /* 1=exists, 0=missing */
>
>In some cases testing to see if: sysdsn(dsn) equals "OK" is a bad idea,
>because sysdsn() can return a bunch of other values admittedly in
>strange circumstances. I think for example there's a no-access to
>dataset message. It's sometimes safer to test for sysdsn() not being
>equal to "DATASET NOT FOUND".
Won't there be unpleasant results regardless? If the response is "NO
ACCESS", I'll try to allocate it NEW, you'll try to allocate it SHR.
Neither will succeed.
Which reminds me... Don, Chris, JD... the bet's off.
Jeremy C B Nicoll
Frank Clarke <ni...@MINDSPRING.COM> wrote:
> On 12 Jun 2003 18:15:24 -0700, Jer...@OMBA.DEMON.CO.UK (Jeremy C B
> Nicoll) wrote:
> <4c01ab41...@omba.demon.co.uk>
> >In article <kv3ievsomufoijn15...@4ax.com>,
> > Frank Clarke <ni...@MINDSPRING.COM> wrote:
> >
> >> tempstat = Sysdsn(outdsn) = "OK" /* 1=exists, 0=missing */
> >
> >In some cases testing to see if: sysdsn(dsn) equals "OK" is a bad
> >idea, because sysdsn() can return a bunch of other values admittedly
> >in strange circumstances. I think for example there's a no-access
> >to dataset message. It's sometimes safer to test for sysdsn() not
> >being equal to "DATASET NOT FOUND".
> Won't there be unpleasant results regardless? If the response is "NO
> ACCESS", I'll try to allocate it NEW, you'll try to allocate it SHR.
> Neither will succeed.
Yes in this case, but code that checks whether a dataset (or member of
a dataset) exists doesn't always then do something that will run under
that userid - it might a generate a batch job, or set someting up to be
run by some other id. And, getting an access-denied message when you
expect just "OK" or "MEMBER NOT FOUND" and only testing for one
possible message is potentially dangerous. I only know this because
I've had code fail as a consequence. Most people use sysdsn() as if it
will *always* tell them that the dataset does or does not exist, or
that a member does or does not exist in a pds, and unfortunately the
routine can essentially return "maybe" answers too - that's all I'm
saying.
Diehl, Gary
and the reallocate it using SHR?
That way, if it doesn't exist, MOD will create it. If it does exist, MOD
will have no effect. And on the second allocation SHR will put the file
pointer to the beginning of file for output.
Best Regards,
Gary
-----Original Message-----
From: Mosley, George [mailto:George...@ICBC.COM]
Sent: Thursday, June 12, 2003 5:27 PM
To: TSO-...@VM.MARIST.EDU
Subject: ALLOC Technique
Hello List.
I'm wondering if anyone out there has a technique for handling the following
file allocation task in one step:
If the file exists, I want to allocate as OLD, i.e. I want to write over any
existing content. If the file doesn't exist, I want to allocate as NEW.
I currently do it this way:
msgstat = MSG('off')
"DELETE ('"BTS_session"')"
msgrset = MSG(msgstat)
"ALLOC F(BTSOUT) DA('"BTS_Session"') NEW ,
UNIT(SYSDA) TRACKS SP(60,50) ,
BUFNO(15) LRECL(133) ,
RECFM(F B A) DSORG(PS)"
Is there some way I can drop the DELETE?
Thanks,
George
Mosley, George
elegance of Frank's solution. Plus, in my instance, there's no risk if the
rare exception does occur.
George
Ryerse, Robin
returns a logical value (1 or 0) indicating the existence of the the named
dataset. The assembler routine does a (catalog) LOCATE and OBTAIN (if disk
dataset). Apart from its primary function, the routine will optionally
populate variable names with the value of the VOLSER and the absolute value
of a relative GDG.
I have found the function quite useful in production jobstreams with
conditional JCL.
The assembler code is ~ 300 lines long. If anyone is interested I can
probably post it in file520 at CBT.
Fogg, George C
dataset is catalog, the function will return your access level (none,
read, update, control, alter) to the dataset for RACF and ACF2 systems.
I haven't tested for a Top Secret security system but I think it will
work. If the dataset is not cataloged or not on the volume that can be
optionally defined then you will receive error messages instead of
return codes. Let me know and I will send you a copy.
George Fogg
> -----Original Message-----
> From: Ryerse, Robin [mailto:robin....@EDS.COM]
> Sent: Friday, June 13, 2003 9:42 AM
> To: TSO-...@VM.MARIST.EDU
> Subject: Re: ALLOC Technique
>
Dave Salt
If what you have can stop me getting ACF2 violations then I'd be very
interested in getting a copy!Thanks in advance,
Dave Salt
http://www.soft-center.com
_________________________________________________________________
The new MSN 8: advanced junk mail protection and 2 months FREE*
http://join.msn.com/?page=features/junkmail
George, William , DHS-ITSD
Is this the RACFACL program?
If not, how does it differ?
Bill
Fogg, George C
Dave Salt
send me a copy. Thanks George,
Dave Salt
http://www.soft-center.com
_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*
http://join.msn.com/?page=features/virus
Fogg, George C
query on what access level you have, not for example, like trying to
open a dataset for update and you only have read or none. I'll send a
copy to those that have requested a copy next Tuesday because I'm out of
the office today and Monday.
John Kalinich
As I discovered, you need to add a SAFDEF in order to successfully use George's rexx function in an ACF2 environment. See doc below.
Regards,
John Kalinich
CSC, St. Louis, MO
To maintain system integrity, CA-ACF2 requires that a user be
authorized to access security definitions; however, some products that
use STATUS=ACCESS are not authorized when they issue the request. The
result is that CA-ACF2 abends the task with a S047 from ACF9C000.
To accommodate these products, CA-ACF2 lets the security administrator
define the specific calls for which the authorization check for
STATUS=ACCESS will be bypassed. This is done with the NOAPFCHK
keyword on a SAFDEF record that describes the specific environment
from which this call is made. For example:
INSERT SAFDEF.apf PROGRAM(pgmname) RB(pgmname) NOAPFCHK
RACROUTE(REQUEST=AUTH,CLASS=DATASET,STATUS=ACCESS)
Dave Salt
in any case was looking for a genric, run it anywhere utility. George,
please disregard my earlier request. Thanks John,
Dave Salt
http://www.soft-center.com
_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail
Craddock, Chris
message when you ask the question unless you are APF auth
and specify the <mumble> keyword. The theory behind issuing
the message is that somebody could just troll for access
until they found something that was "accessible" and go
at it. The messages alert the installation that you're
potentially up to no good.
There's a way you can define a rule that obviates this
requirement for a specific program call but that's maybe
more difficult than you were hoping (and I can't remember
what it is anyway <g>) ACF2 isn't something I dabble in...
Chris
> -----Original Message-----
> From: Fogg, George
> You shouldn't see violation messages because the function is
> basically a
> query on what access level you have, not for example, like trying to
> open a dataset for update and you only have read or none.
Fogg, George C
the function. It's allowed in the RACF world to issue the SAF AUTH
request with the STATUS=ACCESS keyword without any special authority,
however, ACF2 requires you to go the extra mile and define the SAFDEF
statement in order for the problem state user to use STATUS=ACCESS. As
Cris mentioned below, perhaps ACF2 developers with pressure from
auditors required this to stop a user to "troll" for access to resources
or some other reasons unknown to me.
George Fogg
> -----Original Message-----
> From: Craddock, Chris [mailto:Chris_C...@BMC.COM]
> Sent: Monday, June 16, 2003 6:43 AM
> To: TSO-...@VM.MARIST.EDU
> Subject: Re: ALLOC Technique
>