MVSFORUMS.com

BigDaddy

I have a batch cobol program that needs to bump through a table that is really a assembler pgm. I can get the address of the beginning of the table buy using a procedure pointer and doing the following.

SET called-pgm-pointer TO ENTRY ws-called-pgm

What I need to be able to do is bump through the table/pgm and stop when I get to end but I have no way of knowing the length of the table/pgm. Does anybody have any ideas on getting the length of the called table/pgm?

TIA.

kolusu · Posted: Thu Apr 12, 2007 1:23 pm Post subject:

BigDaddy,

Can you use the LENGTH OF special register? It is explained here

http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/IGY3LR10/1.2.8.4?SHELF=&DT=20020920180651&CASE=
_________________
Kolusu
www.linkedin.com/in/kolusu

dbzTHEdinosauer · Posted: Thu Apr 12, 2007 1:39 pm Post subject:

way back when (1960's, 70's, 80's) when we created a table with assembler, the first item was always an S9(9) comp field that equated to the length of the table (area) - 4 (first item) divided by the length of an item (second - thru the last-field). That way, each time the assembler prgm was recompiled (new item) the counter was automatically
increased (or decreased) without having to rely on the programmer to adjust the counter.
_________________
Dick Brenholtz
American living in Varel, Germany

BigDaddy · Posted: Thu Apr 12, 2007 2:10 pm Post subject:

Thanks dinosauer, that is a good idea. Wish I had thought of that in the beginning when I created the assembler tables. My stuff is already in our production so I was trying to get around have to make that change. I have done that in the past but I didn't even think about it, I guess I am getting old.

Kolusu, so you are saying I can do.... LENGTH OF ws-called-pgm? Wouldn't that just give me back the length of the ws field (8)? Am I missing something?

dbzTHEdinosauer · Posted: Thu Apr 12, 2007 3:31 pm Post subject:

BigDaddy,

Enterprise Cobol is near Assember-COBOL compared to VS COBOL. The usage of the SET statement is greatly expanded. I had not bothered to read the EntCOBOL manual. thx to your question, I will read it with care.

Unfortunately I don't think LENGTH of is going to work. If you could adequately describe your table (ASM Module) you would not have the problem that you do now.

I assume you want to make as few changes as possible --- to be able to effortlessly and flawlessly bump thru your table.

Is there no recognizable delimiter within an ASM-Load module? has the object NO length entry before the ENTRY Point?

Way I see it, unless you are willing to redo all your tables, you are going to have to code something in your COBOL module to 'read' the ASM load module for which you receive an ENTRY ADDRESS. I find it hard to believe there is no delimiter for an ASM Object generated by the compiler or that there is no Length - VLI within the ASM Object.

Anyway, interesting problem.

How do you address your COBOL table definition to the procedure-pointer? Do you cheat and redefine and move/or use the value as a regular pointer?
_________________
Dick Brenholtz
American living in Varel, Germany

kolusu · Posted: Fri Apr 13, 2007 6:07 am Post subject:

bigdaddy,

My apolozies. I did not understand the question correctly

Kolusu
_________________
Kolusu
www.linkedin.com/in/kolusu

BigDaddy · Posted: Fri Apr 13, 2007 9:17 am Post subject:

Dino, I had to breakdown and change the table. I went I little futher and added 3 fullwords to the beginning of the table, the first one has the addr of the beginning of the table, the second has the length of a entry and the third has the addr of the end of the table. And to answer you question I do cheat and redefine the procedure pointer as a pointer. I can post what I did if you guys think it would be beneficial.

FYI, I did try to just bump though the table till I hit low values but I got the great S0C4.

dbzTHEdinosauer · Posted: Fri Apr 13, 2007 9:36 am Post subject:

BigDaddy,

as a matter of fact, would like to see the code used to address your table def to the 'loaded module' via procedure-ptr.

actually, I wish to copy and keep it as reference material. my only question is, why does IBM NOT provide this in the SET formats?. That I will research.
_________________
Dick Brenholtz
American living in Varel, Germany

BigDaddy · Posted: Fri Apr 13, 2007 10:31 am Post subject:

here is the assembler table...

BigDaddy · Posted: Fri Apr 13, 2007 10:47 am Post subject:

here is the cobol code...sorry about the crude format, I was in a hurry.

work storage...

dbzTHEdinosauer · Posted: Fri Apr 13, 2007 11:04 am Post subject:

BigDaddy,

thx. i am going to follow-up and try to determine why IBM ENT COBOL does NOT provide a valid SET format.
_________________
Dick Brenholtz
American living in Varel, Germany

dbzTHEdinosauer · Posted: Mon Apr 16, 2007 8:31 am Post subject:

BigDaddy,

though it is a 'non-subject' now, the binder APIs to return information - if you have read access to the data set where the module resides.

The other advantage of the Binder APIs is they work for program objects in PDSE data sets as well as load modules.
_________________
Dick Brenholtz
American living in Varel, Germany

semigeezer · Posted: Mon Apr 16, 2007 11:46 pm Post subject:

For masochists, you could also run the cde control blocks. Somewhere in there is the address and length of the loaded module. usually done in assembler, but can be done in COBOL with a bit of pain.

BigDaddy · Posted: Tue Apr 17, 2007 8:48 am Post subject:

Dino, I do have read access to the loadlib. I might give that a try. This is what happens when you take a old mainframe assembler guy and make him do cobol. Smile

Thanks all for the info

semigeezer · Posted: Tue Apr 17, 2007 1:15 pm Post subject:

I should have mentioned, my post assumed that the called pgm is a separate load module (which it sounds like it may be) and if so, then you could call an assembler stub to use CSVQUERY to get this information. Or call AMBLIST dynamically maybe.