snip
http://www-128.ibm.com/developerworks/edu/dm-dw-dm-0512baldridge-
i.html?S_TACT=105AGX11S_CMP=FP
A little tip for posting long URLs. As some mail clients may wrap this
URL making it unclickable, surrounding the entire hyperlink in and
should helplike this:
Susan,
We are on an AIX platform and we used NFS / CP to move from an IBM F50
to F80 recently. Additionally I use a CRON with find -mtime - 24 and cp
to copy anything that has changed in the last 24 hours so that old F50
becomes a standby system. This usually copies within 3 hours over a
100baseT
Original Message
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Kevin King
Sent: Tuesday, December 27, 2005 12:56 PM
To: u2-users@listserver.u2ug.org
Subject: RE: [U2] UD: Named common
And if you're using an include to bring in the named common, what's
the big deal
another alternative is to join the lines and go to the url
Wednesday, December 28, 2005, 2:57:09 AM, you wrote:
AP snip
http://www-128.ibm.com/developerworks/edu/dm-dw-dm-0512baldridge-
i.html?S_TACT=105AGX11S_CMP=FP
AP A little tip for posting long URLs. As some mail clients may wrap this
The following code works if the item is on file but fails if item is not
on file. Is this correct, shouldn't it just return an empty string ?
If this behavior is the intended behavior then do we just wrap in a try
catch block.
Thanks for replies
george
// open customer file
UniFile fl =
Bill,
real programmers don't compile in PICK mode .. or use PICK .. they
just scale back the expectations G
Tuesday, December 27, 2005, 8:19:42 AM, you wrote:
BH Of course, programs won't compile in PICK mode if the label exceeds 7
BH characters. :-)
-Original Message-
From: [EMAIL
Keith,
Wednesday, December 28, 2005, 7:24:34 AM, you wrote:
And if you're using an include to bring in the named common, what's
the big deal anyway? Okay, so the named common name can't be larger
than 7 characters. Frankly, the 7 is a little weird (3 bits?) but
considering we have so few
Will there ever be a similar offering for Unidata? I struggle with
performance issues every day and would look forward any advice possible.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Clifton Oliver
Sent: Tuesday, December 27, 2005 11:45 PM
To:
Pooh! A stupid (and antediluvian) restriction is still stupid (and
antediluvian) even if it can be worked around. :)
From Google:define:antidiluvian: A term meaning or relating to a
timeless age wherein historical events before the deluge or before
the Biblical Flood took place. Because the
Good point, Andy. I didn't know about www.notlong.com.
www.tinyurl.com provides the same service. I wonder if, fearing a
phishing attempt, there would be trepidation in clicking on a notlong
or tinyurl link by someone who was not aware of these redirector
services. But you've given me the
Rod wrote:
Will there ever be a similar offering for Unidata? I struggle with
performance issues every day and would look forward any advice possible.
Well, I have recently published another article that may be of interest to
you (Rod regarding RFS and deciphering sm.log messages. There is a link
Since we don't have Semaphore-like COMMON structures, I was going to use the
BASIC LOCK to know that a Phantom was running -- the lock missing would
tell me to start the process.
HOWEVER, if the phantom 'dies' - the documentation makes it read as if the
LOCK will stay engaged -- does anyone have
David Wolverton wrote:
(I really wish we had cross-process COMMON for things
like this!)
Amen! That's something that was ripe for development sometime in the 80's
but not a single one of the MV providers added the feature. So many
applications are possible (not least of which would be
Kevin:
You're correct that it's pretty easy to work around this problem...yes
they're in an include. But, IMHO of course, the analysis that it's no big
deal is slightly off the mark.
When we went from 300 baud to 1200 baud then to 4800 baud then to 9600 baud
all of us, well most of us anyway,
David,
I don't currently use a READU on my UD system to LOCK phantoms, but I use it
on UniVerse and it works unless you do an End Task on the UVSH process from
the windows console .. Then you have to do some syntax of the UNLOCK command to
fix things .. On my AIX machine, I use the listuser and
When the phantom dies the semaphore lock get released. We use this
mechanism for all of our background processes and it works like a champ.
Doug
www.u2logic.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of David Wolverton
Sent: Wednesday, December
In UniData 6.1, I need to list the F-type VOC items with counts of the
records and dictionary items.
So I though I'd select the items from VOC, then
LIST VOC A51.COUNT.FILE A51.COUNT.DICT
where those two fields are I-Descriptors that call a subroutine...
SUBROUTINE S.A51.COUNT.RECORDS( RESULT,
Thanks for the link Wally. I will look through it.
Well, I have recently published another article that may be of interest to
you (Rod regarding RFS and deciphering sm.log messages. There is a link to
it on the U2 home page within developerWorks:
Thanks for doing those tests - I just thought someone would have already
done them and spout from memory!
And really thanks for the validation - and I do think it bangs less on the
machine doing it this way.
But GOLLY I wish we had Cross-Process COMMON (and I wasn't even thinking of
MultiPlayer
Wendy
Try this:
RESULT-1 = @SYSTEM.RETURN.CODE
Keep in mind that you are counting multiple records in the VOC.
I used the -1 with your code and got the following result when I listing
10 records from my VOC.
LIST VOC DATA.RECORDS DICT.RECORDS 14:49:55 Dec 28 2005 1
:-)
Bill
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
David Tod Sigafoos
Sent: Wednesday, December 28, 2005 7:56 AM
To: Bill Haskett
Subject: Re[2]: [U2] UD: Named common
Bill,
real programmers don't compile in PICK mode .. or use
rsync is your friend! Put a copy of this on both systems and then run
something similar to this.
echo --- Processing APS ---
/usr/local/bin/rsync -avzrpog --stats /dbms/APS [EMAIL PROTECTED]::dbms
--password-file
=/root/reston
Makes a copy of the CHANGED files from one system to another at the
On 12/28/05, Koser, Mike [EMAIL PROTECTED] wrote:
RESULT-1 = @SYSTEM.RETURN.CODE
Keep in mind that you are counting multiple records in the VOC.
I *think* I'm counting records in the file whose name matches the @ID
of VOC as we're iterating through the select list. I'm only returning
one
We just converted from DG/UX to Win/2003 UV 10.1.11 over the Christmas
extended weekend. Everything went pretty well considering. My problem is my
Alternate indexes will not stay Enabled. We Enable and Rebuild the indexes
but for some reason they won't stay. We do a LIST.INDEX and it shows
From the semi colons (all hail the mighty semi colon!) I will guess
your using c# to do this. If you want to know about your record not
being there, you'll have to use a try catch block to deal with the
exception that will be thrown when the file read occurs and no record
exists.
On 12/28/05,
You will lose your active SELECT list if you don't save it and restore it.
Thank you,
David A. Green
DAG Consulting
(480) 813-1725
www.dagconsulting.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Wendy Smoak
Sent: Wednesday, December 28, 2005 12:16
Wendy
The dictionary entries look like this;
ED DICT VOC DATA.RECORDS
Top of DATA.RECORDS in DICT VOC, 6 lines, 55 characters.
001: V
002: SUBR('COUNT.RECORDS',@ID,,,)
003:
004: DATA}RECORDS
005: 7R
006: M
Also, did you remember to re-catalog the subroutine after the -1 change? I
Wendy wrote
It only ever lists one line. The record count for PERSON is wrong
(but the dictionary count is correct. (Yes, really.) )
/Wendy wrote
Looking at the code I would expect that. The active select list is
exhausted when you hit the first count. Since none of the items in the
person file
You may also have problems if the file ( or dict file ) is invalid or
does not exist. ( You will get a -1 )
Note, can also get around the select list issue, if you include the
select criteria on the list statement.
LIST VOC A51.COUNT.FILE A51.COUNT.DICT WITH F1 = F
Michael Rajkowski
---
Did you create the indexes on the new system or did you copy them over from
the other system?
- Original Message -
From: Mark Waldron [EMAIL PROTECTED]
To: u2-users@listserver.u2ug.org
Sent: Wednesday, December 28, 2005 3:12 PM
Subject: [U2] Alternate Indexes
We just converted from
Maybe the Index Pointer is confused since the Account's path has
changed; try using SET.INDEX DELIN TO NULL and re-create/build the
indexes?
Mark
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Mark Waldron
Sent: Wednesday, December 28, 2005 2:12 PM
To:
We use READU locks - they're much more friendly than semaphores as they can be
given meaningful names that show up with a LIST.READU.
Also they are very quick and easy to test with the RECORDLOCKED() function -
eg.
IF NOT(RECORDLOCKED(filevar, recordname)) THEN ...
AdrianW
DISCLAIMER:
Performing a query from within a query is always going to cause
problems.
I think this topic has been covered before, but in relation to
performing a SELECT from within an subroutine called from an I-type
dictionary item.
Regards,
David
-Original Message-
From: [EMAIL PROTECTED]
On 12/28/05, Mike Rajkowski [EMAIL PROTECTED] wrote:
Looking at the code I would expect that. The active select list is
exhausted when you hit the first count. Since none of the items in the
person file are in the select list no items are counted. ( 0 Items )
The second call to the
Wendy Smoak
A coworker suggested basically the same thing, but with
READLIST/FORMLIST instead of SAVE.LIST/GET.LIST.
That is a lot of mucking about with select-list 0. For more efficiency
without changing your subroutine, try:
SELECT VOC WITH criteria TO 1
LIST VOC FROM 1 A51.COUNT.FILE
Wendy,
I setup the following and got the results I think you are looking for
without the need for selecting the VOC.
AE DICT VOC CTR
001: I
002: SUBR(CTR.VOC,@ID)
003:
004: FILE COUNT DICT COUNT
005: L#21
006: S
AE BP CTR.VOC
001: SUBROUTINE CTR.VOC(RESULT,ID)
002: RESULT =
003: EXECUTE 'COUNT
Bill;
Can't argue much about the 7 character limit on the common names. I'm not sure
why it will compile in basictype 'u' and not 'p'. It just seems awfully
dangerous to let it compile at all.
We moved a lot of systems from Pick to UniData quite easily. ACCT_RESTORE
would even read Pick tapes
37 matches
Mail list logo
