| Title: | -={ H A C K E R S }=- |
| Notice: | Write locked - see NOTED::HACKERS |
| Moderator: | DIEHRD::MORRIS |
| Created: | Thu Feb 20 1986 |
| Last Modified: | Mon Aug 03 1992 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 680 |
| Total number of notes: | 5456 |
I have a customer who has an application written in COBOL and DCL.
The application processes data and produces reports. After a user
run the program for a while, he gets an error saying that he has
run out of symbol table space. He tried raising CLISYMTBL to 500,
but that just seem to postpone the occurence of the problem. On
possible offender in his opinion is the PRINT/DELETE DCL command,
I'm not so sure. Does anyone out there have some ideas as to what
I should look for?
Thanks much,
Rob
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 492.1 | SHOW SYMBOL/ALL | USHS01::BLANDO | Reality, what a concept! | Wed Jun 10 1987 01:03 | 6 |
PRINT/DELETE does not create any symbols (Unless PRINT is actually
not print but redefined to do something else). When you get the
error do a SHOW SYMBOL/ALL and see what it shows. It should help
you track the problem.
FJBlando
| |||||
| 492.2 | What about other things | AUNTB::BRILEY | Wed Jun 10 1987 12:47 | 7 | |
That has been done. There do not appear to be a lot a symbols.
However, according to some sources, symbols aren't the only thing
that is put in the symbol tables. There is alos some scratch stuff
stored in this area.
Thanks,
Rob
| |||||
| 492.3 | how I solve memory depletion bugs | VIDEO::OSMAN | type video::user$7:[osman]eric.six | Wed Jun 10 1987 20:49 | 20 |
What I've done to solve memory-depletion problems of this sort, is something like this, in order: o Make program or .COM file to drive the system in a loop, such that you can reproducibly show the problem o Remove about half of the "guts" of the sysem, and see if problem still exists. o If so, remove half of what's left. o If removing half makes problem go away, put it back, and remove the other half, and verify that problem still exists. Now you've cut entire thing in half, so you repeat. In this fashion, you quickly narrow down to what is causing the problem. If this doesn't make sense, I'll be glad to explain more, just ask. /Eric | |||||
| 492.4 | Understood but.... | AUNTB::BRILEY | Thu Jun 11 1987 13:06 | 6 | |
I understand the technique. Thats no problem. I guess my question
is, what is the other stuff that gets put in symbol table space?
I understand that process symbols and global symbols get mapped
there, but what are the other things that get put there?
Rob
| |||||
| 492.5 | CTLPAGES ? | FNYFS::REZLAN | Thu Jun 11 1987 18:44 | 6 | |
Hi,
Did you try to increase the sysgen parameter CTLPAGES ?`
Alain.
| |||||
| 492.6 | Labels, too | MDVAX3::COAR | A wretched hive of bugs and flamers. | Mon Dec 14 1987 18:57 | 9 |
The only other thing I can think of at the moment that goes into
the `local symbol' table is the internal representation of all labels
in the current command procedure. Try reducing the number of GOTOs
and unused labels in the .COM file. At any rate, raising CLISYMTBL
should correct this - eventually. Changing it online will NOT affect
existing processes - he'd have to log off and back on again to get
the benefit of the new value.
#ken :-)}
| |||||