Does anyone have the source for Universe's LISTFILES program? I need to
determine which files are local and which are remote. The LISTFL verb points
to *LISTFILES which I can't find the source to.

Alternatively, If someone can explain how LISTFILES determines local/remote
status, that would be just as good.

Thanks,

Ed

Ed Sheehan wrote:
In both cases TYPE = "F".

Local has selection criteria F2 NOT.MATCHING ".../..." AND F2
NOT.MATCHING "...\..."

Remote has selection criteria F2 MATCHING ".../..." OR F2 MATCHING
"...\..."

Regards, Keith

Thanks Keith,

I suspected it had something to do with slashes. I've noticed that many
local files can/do have slashes as well (pointing to their own account), so
I'll need to do some further processing, but this will start the ball
rolling.

Thanks again,

Ed

"Pickie" <keith.johnson (AT) datacom (DOT) co.nz> wrote

On 2007-01-22 12:43:39 -0800, "Ed Sheehan" <NOedsSPAM (AT) xmission (DOT) com> said:

You can also open the file to a filevar, use the STATUS statement to
get a dynamic array of information about the file. Field 27 is the full
pathname to the file. Compare that (or portions of it) to the pathname
to the directory you are in.


--

Regards,

Clif

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
W. Clifton Oliver, CCP
CLIFTON OLIVER & ASSOCIATES
Tel: +1 619 460 5678 Web: www.oliver.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Clifton Oliver wrote:

!!! Note: status STATEMENT, not function. !!!

As a matter of fact, I keep a set of I-descriptors in DICT VOC (I share
one D_VOC across all accounts.) that extracts most of the STATUS info
and makes it generally available to SELECT & LIST. Each I-desc
displays a separate field from the array generated by STATUS.

FILEINFO function, too.

cds

FWIW, I'm reading the F-item from each account's VOC. If there are no
slashes in attribute 2, I set a variable to the root dir (everything starts
out in /prod), then the account name, then the file name, so the end result
for a native file of ACCOUNT.PERIODS in the PR.BACKUP account would be
/prod/PR.BACKUP/ACCOUNT.PERIODS. I do the same for attribute 3, since both
dict and data may or may not have pathing embedded independently of each
other.

If there are already slashes, I leave the pathing intact. I write each new
F-pointer to my own VOC's REMOTE.QFILE item, then open the qfile and
READNEXT once to see if it's populated. Then on to the next F-pointer, then
the next account until all accounts are parsed.

This is a very old legacy system, and over the years, many programmers (some
were former engine mechanics, no joke) would stick their own F-pointers in
various VOCs in the system. This all needs to be data-warehoused, hence the
need to find them all from one location. There are many duplicate file
paths, so the last step is to delete all but one path (in my own workspace
of course). I should end up with one path per location to grab.

Am I correct in observing that, if the F-pointer has embedded pathing, the
STATUS Field 27 is always the same as the contents of the F-pointer?

Thanks for the info!

Ed

"Chuck" <charles.stevenson (AT) getronics (DOT) com> wrote

On 2007-01-23 13:15:28 -0800, "Ed Sheehan" <NOedsSPAM (AT) xmission (DOT) com> said:

Should be. But I've never tested it to see how it reacts to symbolic links.

- Clif

re,

the source code is stored into uvhome/BP
regards.

Manu
"Ed Sheehan" <NOedsSPAM (AT) xmission (DOT) com> a écrit dans le message de news:
ep2sg3$qu0$1 (AT) news (DOT) xmission.com...