Hi Craig:
It will return single multi-byte character; similarly, index function will
return the multi-byte character position.
Thanks:
Libin
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Craig Bennett
Sent: Tuesday, July
- Original Message -
From: Leroy Dreyfuss [EMAIL PROTECTED]
Date: Mon, 12 Jul 2004 21:08:49 -0600
To: [EMAIL PROTECTED]
Subject: Re: [U2] capturing problem
UniVerse already does that. The file name is capturexnn, where x is
the PID and nn is aa, ab, etc.
That's why I suggested
This may or may not help. I have seen this type of error at our site. But
most sites may not be setup like we are.
We have a separate file system setup for our _PH_, _HOLD_ and SAVEDLISTS.
The theory being that if we have a runaway report or some other event fills
up this file system, we won't
We did run upon this exact scenario last year and corrected this deficiency
in the DataStage engine (ex-universe) component to support pids that exceed
5 digits.
At 11:08 PM 7/12/2004, you wrote:
UniVerse already does that. The file name is capturexnn, where x is
the PID and nn is aa,
Sort of...
Internally, they are stored in what is known as UV-UTF8, a slight variant
of UTF8, but essentially the
same.
We (ok, IBM does now :-) use the encoding scheme to encode the mark
characters and SQL NULL
into unique code positions that don't conflict with the other UTF8 encoded
Ray,
Definitely report it as a bug, though; multiple capture files with the
same date/time should get a different sequencer (the aa or ab on the
end).
That is not really true, so I am refreshing your memory; we do have the
suffix you described and we do not use time/date as part of the key.
Hi All,
We have a process (Universe program) which writes the records in a particular
file (type 19). One of my C process is picking these records and processing
them and the deleting them.
If I run first process independently (C process stopped) it is very fast.
Also if I have few records in
In a message dated 7/12/2004 9:52:47 PM Pacific Daylight Time,
[EMAIL PROTECTED] writes:
Dear All,
do anyone know how to execute the key TAB at Universe run on NT
OS ? This is fine in Unix . Anything related with NLS ?
Thanks Regards
What do you mean by execute ?
If you want to
In a message dated 7/13/2004 7:23:17 AM Pacific Daylight Time,
[EMAIL PROTECTED] writes:
In terms of data-
For writing 1800 records on a file by Universe it takes 3 min. (when C is
not
running)
For processing 1800 records C process takes very less time (when universe is
not running)
Hi,
I've just tried something similar on UV/NT with no problems - until I pushed
my external routine (scanning the directory and deleting the files as fast
as UV writes them) into a tight loop that swallowed the processor. Then it
all crawled, of course.
Could there be anything in your C routine
1) Throttle the polling program so that it does not go out searching
for information repeatedly, but rather waits 1-2 seconds between
polls.
2) Put protections in place that will prevent the receiving program
from attempting to open a file in progress from the sender. This may
be as simple as
Yes, and yet no.
With NLS enabled, internal storage uses an idiosyncratic variant of UTF8 (sometimes
called UV-UTF8) in which the mark characters Char(248) through Char(255) are always
preserved as single-byte characters, so that the usual occupants of these Unicode code
points are re-mapped
I have a problem where a device file exists in an account, and it can be
written to, but can't write the End Of File marker.
Example:
ASSIGN ATM20 TO MTU 0
in a program, it bombs out when trying
WEOF ELSE
and at TCL it bombs out
T.WEOF
Magnetic tape write end of file failed. [ENOTTY] Not a
Ray Wurlod wrote:
Yes, and yet no.
With NLS enabled, internal storage uses an idiosyncratic variant of UTF8 (sometimes called
UV-UTF8) in which the mark characters Char(248) through Char(255) are always preserved as
single-byte characters, so that the usual occupants of these Unicode code points
Anyone out there using Unidata with Tru64
on Alphas. We have having a lot of trouble with
unaligned access pid messages.
We're trying to migrate from an AIX box and thi
is causing us a real show stopper.
Tried convcode and convdata on all our files which helped
a bit but still no functioning
You mentioned convdata and convcode but not convidx.
Did you run convidx?
Then again, the problem could be in your software migration,
not your data migration from AIX to Tru64.
Unaligned access usually isn't fatal, by itself,
but is sometimes caused by bugs that have
other, and fatal,
That is what a user license is! a license per user/connection. Buy thousands
of licenses.
Louie
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Fawaz Ashraff
Sent: 14 July 2004 03:46
To: [EMAIL PROTECTED]
Subject: [U2] UniObject Licening
Hi All,
We are
Hi Fawaz,
We hace only done tests with Uniobjects and we create a connection to the
databse when we need it and then disconnects. In other terms we only use
approx. 5 licenses simultainously. Since you are developing web applications
you should be able to do the same.
/Bjorn
-Ursprungligt
18 matches
Mail list logo
