OpenQM Google Group

Re: LOCATE question

2009-12-18T19:09:03Z

Hi Eugene,
I will check this and let you knowvwhat I find. I am out of the office
today.
Martin Phillips

RE: jBASE to QM and SpoolerPlus

2009-12-19T08:47:04Z

Top posted on purpose. You already know the subject.

The save and restore options mentioned each have some warts. They can be cured but they need to be cured. The Text Mark is one, of course. Others include ID length, characters legal in the ID of some flavors but not others, and probably a few more. If you have any significant amount of data then time itself becomes a wart. Accuterm does a pretty good job of moving files from one flavor to another but it's not too speedy. The fastest transfer requires some programming. Use Openseq/Writeseq or its equivisent on the source system (prepending the Item ID to the body of the record before write) and Openseq/Readseq or its equivisent on the target. If the target is QM then use OnError as well as Else when you read. The OpenSeq will take the Else clause if the OS file does not exist but the file will be created. The command System() will return an error code of zero in this case effectively telling you that the "error" was corrected. The "OnError" clause will effectively bypass any records with item id's that are not acceptable to QM. This is significantly faster than using the OS hashing if the files are large. On a modest sized system with a few million items the convert time using the OpenSeq method reduced from many hours to about an hour.

Re: jBASE to QM and SpoolerPlus

2009-12-19T02:05:34Z

Hi Rick,
Attached is the "Converting Applications to Open QM" document from
Ladybridge Systems with examples of ACCOUNT.RESTORE options for several
source databases - including jBase - on pg. 21.
There is also a QMSAVE program that comes in specific flavors - eg
QMSAVE.MVB to create a save on an mvBase machine that the QMRESTORE command

Re: jBASE to QM

2009-12-18T22:50:09Z

I did one a few years ago. I don't remember how I transferred the
data, though (sorry). I usually use AccuTerm or a program I have that
T-DUMPs an account. In any case it wasn't a big deal.
The only snags I remember were Indexes and shelling out to windows
(this application did a lot of that) but luckily both were pretty well

jBASE to QM

2009-12-18T22:17:03Z

Has anyone moved a jBASE account to QM (both Windows)?
Will a jBASE account-save work on QM?
Thanks for your time,
Rick

Re: LOCATE question

2009-12-18T21:44:01Z

Works the same as above in D3 7.5.3
If you try IT = "A":@VM , you'll get the same as with IT = "". The
result is true, so ELSE not executed.
I think the behaviour is logical and would not want to see it changed.

Re: LOCATE question

2009-12-18T21:36:25Z

Watch your mode settings when using LOCATE. I would suggest that you
use $MODE UV.LOCATE to get the LOCATE statement to operate in a
similar fashion to UniVerse, UniData, mvBASE (and I assume D3). In
this case, I don't think it will make a difference because you've only
got one value in the dynamic array.

Re: LOCATE question

2009-12-18T20:41:24Z

It is not so much being useful as I was just remembering a technique
that I came across at a site and I did an experiment on qm to see if
works the same way.
I just can not remember if it worked differently on other systems if
the record was null. I am still trying to find out differences
between qm and other systems.

Re: LOCATE question

2009-12-18T20:24:50Z

Considering that there are the len() and dcount() functions that can
give you information about the end of the string, how is the above
technique beneficial?
Ashley Chapman

LOCATE question

2009-12-18T18:59:48Z

1= IT="A"
2= LOCATE "" IN IT<1> SETTING POS ELSE
3= IT<1,POS>=1
4= PRINT IT
5= END
This results in A}1
However, if IT="" the program does not generate anything.
I cannot remember if this was the case on Universe and D3 or not. I
know at one site on Universe, this was a pretty common technique for

Re: Umask on linux file

2009-12-11T18:12:51Z

It's hard to tell from the responses here, but it should be clarified
that umask cannot actually grant permissions. It can only remove
(mask) permissions granted by the process setting permissions. IOW,
umask is not equivalent to chmod, which actually does CHange a file's
MODe.

A decent explanation, along with how to calculate umasks, can be found

Re: Umask on linux file

2009-12-11T16:35:52Z

Hi Cedric,

The UMASK command simply sets the underlying operating system umask value so
it should affect all files created from the QM session including use of
CREATE.FILE though this has another way to do this by use of its MODES
option.

Martin Phillips
Ladybridge Systems Ltd
17b Coldstream Lane, Hardingstone, Northampton, NN4 6DB

Re: Umask on linux file

2009-12-11T15:40:59Z

Does this apply to all the files created ? Including create.file ?

--
Cedric Fontaine
[link]

Re: Umask on linux file

2009-12-11T14:31:16Z

It sounds like Cedric needs to do a umask of 000 on the qm side to
give him the same as chmod 777.

We set ours up to put umask of 007 in each user bash profile. I was
not aware of the umask command on the qm side. Interestingly, I typed
umask at tcl and it returned 007.

Eugene

On Dec 10, 4:24 pm, "Martin Phillips" <martinphill...@ladybridge.com >

Re: Umask on linux file

2009-12-10T21:24:45Z

Hi Cedric,
Is there a way to create file with a CHMOD of 0777 with a Write on QM on a linux dir. Of course, I can create my file and then execute a chmod but I'd like a way to do it with a single write.
QM has a UMASK command to set the modes for files created from within QM.
Martin Phillips
Ladybridge Systems Ltd

Re: list file size

2009-12-10T21:22:59Z

Hi Steve,
will len(@record) be equal to:
len(id) + len(record) ???
No. The id is not part of the record.
Martin

Re: Umask on linux file

2009-12-10T21:16:15Z

Cedric,

We have every user have a bash_profile. In that profile we put in
umask 007. What you are going to need is some variation of that umask
number. The 007 creates read and write privileges for the group and
owner but not the "other".

You might play with that to see what umask you will need to create the

Re: list file size

2009-12-10T21:07:53Z

2009/12/10 CDMI - Steve T <c...@swbell.net>:

Nope. @record is the length of the record, EXCLUDING the key.

--
Ashley Chapman

Re: Umask on linux file

2009-12-10T21:06:26Z

Cedric,

I do just that. Include :-
execute 'SH umask 000' capturing umask

before the file create in your program.

Substitute 000 with whatever is appropriate for you.

2009/12/10 Cedric Fontaine <cfonta...@spidmail.net>:

--
Ashley Chapman

Umask on linux file

2009-12-10T20:50:18Z

Hello,

Is there a way to create file with a CHMOD of 0777 with a Write on QM
on a linux dir. Of course, I can create my file and then execute a
chmod but I'd like a way to do it with a single write.

Thanks.

Re: list file size

2009-12-10T20:48:20Z

Thanks a lot.

On 10 déc, 11:38, "Martin Phillips" <martinphill...@ladybridge.com >
wrote:

Re: list file size

2009-12-10T20:36:30Z

quick question sir
will len(@record) be equal to:
len(id) + len(record) ???

Steve Trimble
Computerized Data Mgmt Inc
PO Box 13403
Maumelle, AR 72113
(501) 803-0974 09:00am - 6:00pm CST

Home of self defense products: [link]

______________________________ __

Re: list file size

2009-12-10T16:38:23Z

Hi Eugene/Cedric,
Although this suggestion works for compatibility with Pick systems, I
personally prefer an I-type...
1: I
2: LEN(@RECORD)
as this doesn't rely on some special interpretation of a field number of
9999 and is, therefore, readable by programmers who have not come from a
Pick background.

Re: How to search this google group

2009-12-10T16:19:06Z

Kevin,

I was pretty excited to see that they had fixed it. Thanks for
reporting the problem.

It was very frustrating that they had not fixed it sooner. It is a
pretty important feature to have for a "search company"

Eugene

Re: list file size

2009-12-10T16:17:24Z

Cedric,

Make sure that there is a size item in either dict or the md(voc).

MD SIZE
10 line(s)
----: P
0001: A
0002: 9999
0003: SIZE
0004:
0005:
0006:
0007:
0008:
0009: R
0010: 13
Bottom at line 10
----:

You can list the size just the same way that you do on D3.

Eugene

list file size

2009-12-10T15:32:54Z

Hello,

Is there any way to list item size ? On D3, we can do LIST myfile
SIZE.

Is there an option to do this on QM ?

Thanks,
Cedric Fontaine

Re: How to search this google group

2009-12-09T21:14:39Z

Wow! Finally. I reported this problem to Google a long time ago, and
only after noticing that it hasn't been working for a very long time.

Google Groups does not seem to be a very high priority part of
Google's business. Just check out their support pages.

--
Kevin Powick

Re: How to search this google group

2009-12-09T15:47:05Z

Ashley,

It appears that google may have fixed the problem.

Eugene

On Dec 9, 9:49 am, Ashley Chapman <ashley.chap...@billabong-

Re: How to search this google group

2009-12-09T14:49:40Z

Hi Eugene,
Can you give an example where the Google search falls down? Are you
searching through the plain ol' google.com, or from the google-group?
Ashley

How to search this google group

2009-12-09T14:25:57Z

Google appears to have problems when it comes to searching the posts.

Is there somewhere else on the net that I could go to search this and
other google groups that has a better search capability.

There are times that I need to go back and see what people in the
group have said about a subject. I just find it very difficult to

Re: Accessing file on shaerd mapped drive (windows)

2009-11-27T19:08:43Z

OK. I've had this problem before ...

Look in the ERR.H item in the SYSCOM file. This gives the error
translations.

Error 3007 is 'No VOC record'.

So, looking back at your original post, you have defined a VOC entry -
but there is no dictionary component. So create a dictionary:

CREATE.FILE DICT JUNK

Re: Accessing file on shaerd mapped drive (windows)

2009-11-27T17:48:21Z

Hi All,

Thanks for your response(s)

(1) Does appear when responding to a DOS dir cmd

(2) Does appear when responding to a !dir cmd in QM

(3) Responds with the 3007 error when attempting to open in basic
program in QM

(4) QM version 2.9.9. Have tried setting NETFILES=1 & NETFILES=2 W

Re: Very large variable

2009-11-27T09:21:08Z

Hi Brian/Cedric,
We do something very much like this. We are investigating some performance
issue in management of large strings but the general perception of our users
appears to be that we do not have a significant problem.
Martin Phillips
Ladybridge Systems Ltd
17b Coldstream Lane, Hardingstone, Northampton, NN4 6DB

Re: Accessing file on shaerd mapped drive (windows)

2009-11-26T20:20:43Z

Another couple of things to try:

Open a DOS box from Windows and try to list the shared folder using a
DIR command.

If that doesn't work, you've got general networking issues.

Assuming that does work, try the same thing from the QM command
prompt:

! DIR path_to_sharedfolder

If that works, then you should be able to access the data from

Re: Very large variable

2009-11-26T19:53:17Z

In general, multi-value systems can run run into speed issues when
dealing with large dynamic arrays - but it depends on the
optimisations built into the language and on how you are using it.

Long, long ago, in a galaxy far away, MV databases used to count down
the dynamic array whenever you accessed the variable. This made many

Re: Very large variable

2009-11-26T18:46:54Z

It is an item we read and then on a very large attribute we are doing
locate, insert and delete.

Re: Very large variable

2009-11-26T15:49:52Z

Hi Cedric,
Please explain a bit more about how you update this variable. We regularly
work with data of this size and bigger with no problems.
Martin Phillips
Ladybridge Systems Ltd
17b Coldstream Lane, Hardingstone, Northampton, NN4 6DB
+44-(0)1604-709200

Very large variable

2009-11-26T15:34:07Z

Hello,

We're working on very large variables (300 000 bytes) and it seems too
heavy for QM to work on this large variable and so time is exponential
when adding data to this big variable.

Is there any optimization on QM part we can do to work faster with
big items ?

Re: Accessing file on shaerd mapped drive (windows)

2009-11-25T20:56:53Z

why set a 'mapped drive' ?
why not just:

READV MAPPED.DRIVE FROM PARAMETER.FILE,MAPPED-DRIVE,1 ELSE
MAPPED.DRIVE = "\\CDMI-WEB\CDMIWEB_F\QMSYS"
END
**** something like that above... then
OPENPATH MAPPED.DRIVE TO FILE ELSE
CRT \Could not open \:MAPPED.DRIVE:\ status \:STATUS():\ os.error \:OS.ERROR()

Re: Accessing file on shaerd mapped drive (windows)

2009-11-25T19:45:39Z

Hi Gary,
Try a simple program that attempts to open the file and print the values of
STATUS() and OS.ERROR() in the ELSE clause.
Martin
----- Original Message -----
To: "OpenQM" <openqm@googlegroups.com>
Sent: Wednesday, November 25, 2009 6:31 PM
Hi Martin,
I had already set NETFILES to 1 (or does it need to be set to 2) with

Re: Accessing file on shaerd mapped drive (windows)

2009-11-25T18:31:17Z

Hi Martin,

I had already set NETFILES to 1 (or does it need to be set to 2) with
the same results.
Anything else I should be looking at?

Thanks,

Gary

On 25 Nov, 11:00, "Martin Phillips" <martinphill...@ladybridge.com >
wrote:

Re: Accessing file on shaerd mapped drive (windows)

2009-11-25T18:00:55Z

Hi Gary,
This should work fine but the NETFILES configuration parameter must be set
to enable remote access (See the QM Reference Manual). This is because there
are potential locking issues that QM cannot resolve.
Martin Phillips
Ladybridge Systems Ltd
17b Coldstream Lane, Hardingstone, Northampton, NN4 6DB

Accessing file on shaerd mapped drive (windows)

2009-11-25T17:56:26Z

Hopefully just missing something obvious. Simply want QM to be able to
'see' (read/write) a file on a windows shared mapped drive. I have
tried all sorts of permutations of what I thought the following should
be to no avail ...

VOC entry ...

0001 F
0002 shared name (ie ... z:\ or \\computer\sharedname , etc...

Re: Transfer current application from 2000 server to 2008 server

2009-11-20T13:54:15Z

Thanks

fyi I've already contacted easyco and have scheduled support just in
case I shoot myself in the foot

Rich

On 20 Nov, 08:47, "Martin Phillips" <martinphill...@ladybridge.com >
wrote:

Re: Transfer current application from 2000 server to 2008 server

2009-11-20T13:47:58Z

Hi Rich,
Almost.....
The easiest way to do this is the other way around.
1. Copy the QMSYS directory and all its subdirectories across.
2. Copy all the application account directories across.
3. Now install QM on the new system.
By doing it in this order, you don't have to worry about what parts of the

Re: Transfer current application from 2000 server to 2008 server

2009-11-20T13:15:40Z

1) clean install of new qmsys, so services etc are installed on the
new server

2)windows copy existing qmsys to new c:\qmsys (overwriting)
so catalog space etc is retained

3)move applications to their new destination (yes same location will
be used)
L:\directory name

?? will the indexes require rebuilding/remapping like universe?

Re: Transfer current application from 2000 server to 2008 server

2009-11-20T13:06:49Z

Hi Rich,
QMSYS needs to be installed in a position where it can be seen by all users.
Copy the application account directories unchanged from the old system,
preferably keeping their pathnames unchanged.
The only component in QMSYS that needs updating is the ACCOUNTS file. You
should be able to copy this from teh old system too. If you have changed any

Transfer current application from 2000 server to 2008 server

2009-11-20T12:39:21Z

I have two application that needs to be moved to a client's new 2008
server.

I searched the archives/etc and did not find any information related
to the recommened method of transfer.

My 1st confusion is where to install qmsys since the 2008 server will
have multiple virtual servers defined

Re: Where to add MED (tool) created menu as startup menu after account login

2009-11-18T23:51:13Z

Thanks a lot guys...........that woked...........See Ya....

On 17 Nov, 16:42, "Martin Phillips" <martinphill...@ladybridge.com >
wrote:

Re: Where to add MED (tool) created menu as startup menu after account login

2009-11-17T22:42:19Z

Hi Darryl,
The menu itself should preferably go in the VOC butit can be put in another
file with a remote pointer (R-type record) in the VOC instead. This latter
approach has the advantage of allowing a central menu file that can be
accessed from multiple accounts if that is useful.
To make the menu display when the user logs in, simply reference it from the