Re: QWVRCSTK api returns value of N in Information-status field

Yikes!!! Not working. I changed my code to replace the QWVRCSTK api
with combination call to QMHSNDPM and QMHRCVPM api and getting the
CPF2479 (Call stack entry not found) error. I then tried changing the
CREATE FUNCTION to include NOT FENCED and still getting the same error.
Now, tell me why some people keep thinking that SQL is more superior
than RPG now :)?

Any suggest to get around this problem.



"Lim Hock-Chai" <Lim.Hock-Chai@xxxxxxxxxxxxxxx> wrote in message
news:<mailman.1141.1232311740.21608.rpg400-l@xxxxxxxxxxxx>...
Thanks Chuck, Your explanation fits what I have experienced. Now, the
question is how can I get around this issue with a more solid solution
(since you said NOT FENCED is not a absolute solution). My procedure
calls the QWVRCSTK api for one reason only. It is to make sure the
program-name, being passed in as a parameter, that the caller passed in
does exist in the call stack. Another method that I know of that will
allows me to determine if a program exists on the call stack is to do a
combination call to the QMHSNDPM and QMHRCVPM api. Do you know if this
solution will be a more solid solution to my problem?

Thanks.




"CRPence" <CRPbottle@xxxxxxxxx> wrote in message
news:<mailman.1053.1232152022.21608.rpg400-l@xxxxxxxxxxxx>...

Lim Hock-Chai wrote:

I've an export procedure that will call the QWVRCSTK api to
get the call stack info of the current job. It works fine
when calling from a RPG program. However, it failed (Returns
value of N in Information-status field) when it is call from
a sql UDF. Any idea why?

I believe that is the result of the SQL UDF running in its own
thread. There is a clause on the CREATE FUNCTION which can minimize

the

chance the function will run in its own thread; IIRC it is the NOT
FENCED clause which will encourage, but not require, that the SQL

should

run the function in the primary thread. The function runs in\from the

LIC, thus no stack information will be available. If the primary

thread

of the job is being retrieved instead [of the thread dedicated to the
function], then a similar origin [per threading] may be the cause, but

it would most likely be that the primary thread is initially busy down

in the LIC [e.g. starting new threads] such that repeated attempts

might

be required to obtain the information.

Regards, Chuck

----------