0077 (double BIND) in ADS/O?!!

[Anon Anon]

PMOND017 links to PTELD860, all retrieval, each has unique subschema

After return to PMOND017 (in second pass through), the first DML statement (OBTAIN CALC) is successful but the dialog issues a BIND record upon execution of the 2nd DML statement causing the 0077 abend. 

THe attached doc info was created using TRACE=ALL commands, the first pass through is successful but the second is not....

THe second pass through shows a 'CURRENCIES SAVED FOR PMOND017' afer execution of the first DML statement...

Attachment(s)

ADSO - 0077 abend.txt   9 KB 1 version

[Gary Cherlet]

Hi Alex - I have taken a look at the log - and I have a few questions - first let me see if I understand scenario and log contents:

PMOND017 - does some work -  we do not see if it has done any DML or if a run unit is bound

PMOND017 - LINK to PTELD860 - are there any "options" on the lINK command?
         - we do not see if there is a "Currency Save for PMOND017 at Link Level 000000?
-------------- 1st Pass through we start to see output from the log at this point
PTELD860 - Obtain
PTELD860 - Obtain
PTELD860 - FRMTPART - Goback
PTELD860 - RETURN
PMOND017 - Finish
PMOND017 - Bind/Ready sequence
PMOND017 - Currency Restore + Release performed by ADS (preceding commmands in support of 1st DML statement)
PMOND017 - OBTAIN calc PM-PART-REC
PMOND017 - "other" commands
PMOND017 - OBTAIN calc PT-VENDOR-REC
---------------- Log output missing
PMOND017 - does some more work - presumably
PMOND017 - LINK to PTELD860 - are there any "options" on the lINK command?
         - we do not see if there is a Currency Save for PMOND017
---------------- 2nd pass through - we start to see output from the log at this point
PTELD860 - Obtain
PTELD860 - Obtain
PTELD860 - FRMTPART - Goback
PTELD860 - RETURN
PMOND017 - Finish
PMOND017 - Bind/Ready sequence
PMOND017 - Currency Restore + Release performed by ADS (preceding commmands in support of 1st DML statement)
PMOND017 - OBTAIN calc PM-PART-REC
PMOND017 - Currencies SAVED for dialog PMOND017 - I don't understand why this is here! Is there some DML command not shown? A "link" we don't see?
PMOND017 - BIND REC PARTS-CONTROL - 0077 - Run Unit not bound or bound twice ---->

Well - there's something mysterious happening here! Do you have any SAVE / NOSAVE other options on any LINK commands? Are there any other dialogs involved in the thread that use the same subschema as PMOND017? Do they reference different records or the same records as PMOND017?

I have seen a problem in the past where there is a record that is not referenced in code not being bound and then that causes a problem - EVENTUALL - perhaps put a "Move" or "Initialise" for the PARTS-CONTROL record in amongst the first statements that PMOND017 executes BEFORE the LINK to PTELD860 - so that ADS knows that you will be referencing the record later on!

Not much help I know - but maybe there is something useful in the questions or suggestions?

Take care - cheers - GaryC

[Anon Anon]

Hello Gary,

  There are no options with th LINK commands.

   Just prior to first pass-through, PMOND017 does do an OBTAIN CALC PARTS-CONTROL before LINK to PTELD860 and currencies are saved 

  Just prior to second pass-through, PMOND017 does not repeat the OBT CALC PARTS-CONTROL nor does it do any other DML statements before linking to PTELD860.    

  Yes, the 'currencies saved' message for PMOND017 after the OBT CALC PM-PART-REC implies that there is a link or a display or something telling ADS/O to anticipate performing another set of BINDs upon execution of the next DML statement in PMOND017 when it doesn't have to...

 

[Anon Anon]

Can you please post the actual source code from both programs.  Are they both ADS?  or is the linked to dialog DC Cobol or assembler?

 

[Anon Anon]

Hello Linda,

  At this point, would rather not post my client's proprietory code but can tell you that both programs are ADS dialogs.  The error is occuring in our primary unit testing environment but not in our QA test environment (which is pretty much a copy of production).

[Anon Anon]

Because the changed dialogs worked in our QA test environment, I did implement to prod some time ago with no subsequent issues. The quality of our data in our primary test environment is somewhat wonky so I have to attribute the 0077 abends to that (somehow).

At any rate, just wanted to share access to a classic old school IDMS support site that may be of interest: 

http://www.idms.ws/