Re: Problem with PASCAL. (245 Views)
Reply
Valued Contributor
Cleber Zorzi
Posts: 193
Registered: ‎11-20-2005
Message 1 of 6 (245 Views)

Problem with PASCAL.

Hello,

I have a program written in PASCHAL, in the compiler I qualified the option float=d_float (it functioned well of the VAX for ALPHA), this program reads an archive base with products (opens a time) and for each product the archive calls n times packings, in the ALPHA functions well but in the INTEGRITY "WORKING SET" of the process is continuously being developed. In a similar program it gives error when the 1242 arrive and in another one it executed but it arrived the 3541.

Can help me?

Thanks,

Cleber
Honored Contributor
Hoff
Posts: 4,958
Registered: ‎01-29-2006
Message 2 of 6 (245 Views)

Re: Problem with PASCAL.

I'll assume you are aware that Integrity does not implement D_float, it converts it. D_float was native on VAX, D_float and the S and T_float IEEE are native on Alpha, and IEEE S and T_float are native on Integrity.

As for your problem, there's apparently a leak somewhere.

You will want to try to create a small reproducer and verify return status and such, and then contact HP support. This could be something latent in the code, or something in the Pascal compiler, or it could be lurking within OpenVMS I64.

You'll want to apply ECO kits for whichever release of OpenVMS I64 is involved, as well.

Stephen Hoffman
HoffmanLabs
HP Pro
Ian Miller.
Posts: 4,371
Registered: ‎06-03-2003
Message 3 of 6 (245 Views)

Re: Problem with PASCAL.

just to restate your question to see if I understand it.
You have a program written in PASCAL running on OpenVMS I64 which was compiled with the /FLOAT=D_FLOAT option and the working set continuously grows. The program does not do this on VAX and Alpha.

- correct so far?

Which version of OpenVMS I64 and which version of the PASCAL compiler are you using?

I did not understand this part
"this program reads an archive base with products (opens a time) and for each product the archive calls n times packings"
can you reword that part?
____________________
Purely Personal Opinion
Honored Contributor
Hein van den Heuvel
Posts: 6,588
Registered: ‎05-19-2003
Message 4 of 6 (245 Views)

Re: Problem with PASCAL.

>> In a similar program it gives error when the 1242 arrive and in another one it executed but it arrived the 3541.

It would be useful to know the actual error message(s), as exactly as possible (cut & paste).

A few weeks back Cor Mom reported an inssue which turned out to be Pascal on Integrety no auto-closing files. Does that 'packings' function open files? While the program is running, check with $SHOW DEV/FILE or $ANALYZE/SYS... SHOW PROC/CHAN "problem program"

Check out:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1087750

hth,
Hein van den Heuvel
Valued Contributor
Cleber Zorzi
Posts: 193
Registered: ‎11-20-2005
Message 5 of 6 (245 Views)

Re: Problem with PASCAL.

Ian,

The program was carried of the Vax for Alpha and now for Integrity. The version OpenVMS IA64 is 8.3 and the Pascal I do not know.

The program read one file with products (open once) and to each product read "n" times the file ( open and close "n" times the file).

Thanks,

Cleber
Honored Contributor
Hein van den Heuvel
Posts: 6,588
Registered: ‎05-19-2003
Message 6 of 6 (245 Views)

Re: Problem with PASCAL.

And the actual error message is... ?

>> The program read one file with products (open once) and to each product read "n" times the file ( open and close "n" times the file).

So if there is a problem with the close, then that might have gone unnoticed on Alpha, but would run into the auto-close issue I refered to on I64

Anyway you made sure no collection of open files is build up during the run right?

fwiw... if it is the same file being opened and closed over and over, then that seems hardly optimal. Any chance to just keep it open?

Regards,
Hein.





The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.