T.R | Title | User | Personal Name | Date | Lines |
---|
3057.1 | No clues | UTRTSC::SMEETS | Workgroup support | Thu Mar 20 1997 10:12 | 9 |
| Hi,
Doesn't anybody has an idea about this ?
Could the customer's code help to isolate the problem ?
Thanks,
Martin
|
3057.2 | DECEDI_FREE_ITEM_LIST ?? | UTRTSC::SMEETS | Workgroup support | Thu Mar 20 1997 14:41 | 18 |
| Hi,
some more information on this one....
Performing exactly the same command from the CLI doesn't show the behaviour.
The customer builds the itemlist among others with the following identifiers
DECEDI_ITM_IO_DEBUG and DECEDI_ITM_ERROR_LOG.
After the trade fetch command has succesfully completed, one of the last actions
in the code is a DECEDI_FREE_ITEM_LIST.
The next time the customer builds a new itemlist without specifying a
DECEDI_ITM_IO_DEBUG. the IO_DEBUG file is still open by the OBB_<pid> process.
Doesn't the DECEDI_FREE_ITEM_LIST free all items ?
Martin
|
3057.3 | Confused | SYSTEM::HELLIAR | http://samedi.reo.dec.com/ | Fri Mar 21 1997 14:39 | 15 |
| MArtin,
FREE_LIST frees the list and all memory associated with it. It does not
free any file handles associated with any files on the list as there
should not be any!
The fact that the CLI works but your program does not implies that it
is your program that has the handle to the file (as the CLI would run
up the image each time and on image exit all file handles are closed
by the operating system).
To confirm this, run your program and then do a SHOW DEVICE/FILES
for the device holding the problem file.
Graham
|
3057.4 | Two different fetches | UTRTSC::SMEETS | Workgroup support | Fri Mar 21 1997 15:25 | 12 |
| Hi Graham,
It looks like it's bit different than we first thought. After I asked the
customer what the program is doing exactly, he told me the following.
1. First the customer builds an itemlist which includes a debugfile.
2. After a successfull fetch, the itemlist is freed up
3. customer builds a new itemlist which is used to fetch all other documents
This all happens in the same programm.
Martin
|
3057.5 | Little update on this | UTRTSC::SMEETS | Workgroup support | Wed Apr 02 1997 16:21 | 13 |
| Hi,
A little update on this.
I wrote an program which performed the same tasks as the customer's program, and
I wasn't able to reproduce the problem.
Now at this moment it seems that the problem doesn't show up any more at the
customers site after raising some DECnet Phase IV related resources.
Regards,
Martin
|