Paradox Community

Items in pnews.paradox-programming

Subject:Re: Synchronizing priv tbls with master tbl
Date:Sat, 28 Jul 2018 09:34:50 -0400
From:"Steven Green" <>
all three machines MUST point to the shared net file using the EXACT SAME 
path/mapping, including drive letter

all three machines MUST point to any shared data using the EXACT SAME 
path/mapping, including drive letter

they can't just resolve to the same place, they must be addressed exactly 
the same way


Steven Green
Myrtle Beach, South Carolina, USA

Collectibles and Memorabilia
Vintage Lego Sets and Parts
- and Paradox support, too
"Peter"  wrote in message news:5b5b7423$

Thanks for the reply Steven
In the simplest test I open a table (49 records,22 fields)on each
machine: 1 server/workstation, 2 workstations

I make a change to one field on a computer and in my mind that should
update that same field on the other two machines. It does, sometimes. I
found that on each machine switching to Windows Explorer and under
"View" I click "Refresh"  then after awhile the tables on all the
machines would update without switching to W-Explorer, as expected.

The server/workstation (call it mach #1) has drive E mapped to itself as
drive "I". The other 2 machines have mach #1's drive "E" mapped as
networked drives:

Mach #1 Mach#2 & #3
E   I (mapped to mach#1 E)
I (mapped to E)

I have confirmed that Mach#1 is the MasterBrowser, if that makes a

On 07/27/2018 05:59 AM, Steven Green wrote:
> Peter, you're partially missing the point of the "refresh" concept.. 
> there's a big difference between not seeing changes others have made, and 
> not seeing changes to a record you're trying to lock or update.. or not 
> seeing changes to a calc that your trying to update
> maybe I missed something, but all I'm seeing here is "machine 2 doesn't 
> see what machine 1 did".. how does the code process what machine 1 does, 
> what is machine 2 trying to do that fails, how is your code testing for 
> it, what errors or messages are you getting, etc?
> -- 
> Steven Green
> Myrtle Beach, South Carolina, USA
> Collectibles and Memorabilia
> Vintage Lego Sets and Parts
> - and Paradox support, too
> "Peter"  wrote in message news:5b5aa799$
> More Experimenting
> I found I would get the "DataRefresh" message only when the form AND the
> table were open. The form's "DataRefresh" seemed to be triggered by the
> table change, not directly the result of a field being changed on the
> form displayed on one of the other computers.
> Mach #1 field changed ----->>>> Mach #2, table updated ----->>generates
> a "DataRefresh" which triggers form:action: "DataRefresh"
> Results
>   *Just the form open on each of the computers will not generate a
> "DataRefresh"
>   *The table opened on each of the computers will update the other
> computers, sometimes, ie, not consistently
>   *The "DataRefresh" in the action method is the result of the changes
> on the table which must also be open.
> I guess I'll have to put a machine together to act as a dedicated
> server. Not sure what else to check.
> On 07/26/2018 03:03 PM, Peter wrote:
>> I created a new form, added the table to the datamodel, added a 
>> tableframe, defined 4 fields. Saved, opened the same form on all 3 
>> machines.
>> Inconsistent results, no pattern that I can see. Changing a field value 
>> on any of the machines did not update what was displayed on the other two 
>> machines. Sometimes if I scrolled down, or [PageDown] and then scrolled 
>> back up or [PageUp] it would refresh, most of the time it would not. 
>> Played around with it for hours.
>> At one point I added in the record:action
>> if = DataRefresh then
>>     view("refreshing data")
>> endIf
>> Sometimes it would work, sometimes not.
>> I just closed the form and displayed the tables in edit mode on each 
>> computer. Now when I change a field value on any of the 3 computers the 
>> other two update within the 3 seconds I have set. sigh.
>> On 07/26/2018 02:41 PM, Mark Bannister wrote:
>>> If the tables are viewed in a tableview and not a form they will not 
>>> refresh if they are open before the changes are made. (well, maybe 
>>> sometimes under certain situations).
>>> I don't think your mapping is the issue.
>>> Just slap one in a form and see.
>>> On 7/26/2018 3:34 PM, Peter wrote:
>>>> I just realized that I have a local drive mapped as a network drive.
>>>> When my Pdox starts up (on any computer) it runs a script which reads a 
>>>> table that lists all the aliases and paths. eg, alias|path "Ticket" | 
>>>> "I:\DEV\TickDat". For this reason I need a common letter for all 
>>>> computers.
>>>> The computer that has the files is the master browser and is also a 
>>>> workstation. The files are actually on drive "E" so I mapped it as 
>>>> drive "I"
>>>> Would this have an effect? How often does Windows update the network?
>>>> On 07/26/2018 12:29 PM, Peter wrote:
>>>>> Thanks Mark,
>>>>> I have them all in edit mode, does that make a difference? Sometimes 
>>>>> it does work. I will try it by displaying in a form as you suggest.
>>>>> On 07/26/2018 11:37 AM, Mark Bannister wrote:
>>>>>> Peter:
>>>>>> If the tables are only open in a table view window they will not show

>>>>>> updates.  Display them in a form and they will get refreshed.
>>>>>> On 7/26/2018 12:50 PM, Peter wrote:
>>>>>>> On 07/25/2018 07:36 PM, Liz McGuire wrote:
>>>>>>>> In the BDE on both computers, do you have LOCAL SHARE set to
>>>>>>>> (Config > System > INIT), and STRICTINTEGRITY set to
TRUE (Config
>>>>>>>> > Drivers > Native > PARADOX)?
>>>>>>> Yes. I triple checked and all the BDE settings are the same on
>>>>>>> computers (I added one more)
>>>>>>> I did a lot of experimenting yesterday, used 3 computers and 
>>>>>>> displayed just the 1 table. All were in edit mode. Used record
>>>>>>> changed just 1 alpha field.
>>>>>>> Sometimes the other computers would update, 90% of the time they

>>>>>>> would not. It would work for awhile after all 3 were freshly 
>>>>>>> hard-booted, then it would stop working. Sometimes it would only

>>>>>>> work if I modified from 1 computer but not from the others, they

>>>>>>> each took their turn so no pattern there.
>>>>>>> I did find that if I went into Windows Explorer and clicked View
>>>>>>> Refresh that it would seem to work more consistently after that.
>>>>>>> was able to determine which computer was the master browser, 
>>>>>>> wondering if this is some kind of network refresh problem.
>>>>>>>> Those are the first two things that come to mind.
>>>>>>>> Liz
>>>>>>>> On 25 Jul 2018 17:31, Peter wrote:
>>>>>>>>> So I decided to not use ri. I pondered this and decided
it must 
>>>>>>>>> have
>>>>>>>>> something to do with complicating record locking as opposed
>>>>>>>>> corruption issues.
>>>>>>>>> Anyhow I've started to work on a synchronization method
>>>>>>>>> decided to
>>>>>>>>> go with detecting a datarefresh which will trigger the

>>>>>>>>> synchronizing
>>>>>>>>> routine. I thought there was a way to detect a DataRefresh
>>>>>>>>> DataRefreshOutside but didn't get anywhere with that, yet.
>>>>>>>>> The issue I am having is the failure of updating (refreshing)
>>>>>>>>> in a
>>>>>>>>> table. I set up a simple test:
>>>>>>>>> * 2 computers: 1 workstation, 1 server/workstation (both
>>>>>>>>> XPpro)
>>>>>>>>> * Both computers have the same table open and in edit mode
>>>>>>>>> * Test table has only 12 fields and ~50 records
>>>>>>>>> * Computer 1, manualy change a value in a field
>>>>>>>>> * Computer 2, no changes, have to press [ctrl][F3] to update
>>>>>>>>> * Refresh rate is set to 3(secs), retry is also set to
3 (both 
>>>>>>>>> computers)
>>>>>>>>> * Both computers have the same version of Pdox (v10)
>>>>>>>>> * Both computers run the same BDE (v5.2.0.2, thank you
>>>>>>>>> I doubt "write cache" and "oplocks" have anything to do
with the
>>>>>>>>> situation but both are disabled.
>>>>>>>>> Can someone advise why the tables might not be 
>>>>>>>>> updating/refreshing?
>>>>>>>>> Thank you.
>>>>>>>>> On 07/16/2018 12:46 AM, Leslie wrote:
>>>>>>>>>> Hi Peter,
>>>>>>>>>> It would depend upon the requirements. Her are some
>>>>>>>>>> some of
>>>>>>>>>> which may not be relevant:
>>>>>>>>>> Is this a record level book in/bookout scenario ?
>>>>>>>>>> Can each user only add new data or can they modify
and delete as 
>>>>>>>>>> well ?
>>>>>>>>>> How long can the copy be away from the master - days,
>>>>>>>>>> months ?
>>>>>>>>>> How are you going to identify what has changed ?
>>>>>>>>>> How are you going to handle conflict resolution, ie
two users 
>>>>>>>>>> modify a
>>>>>>>>>> field to a different value within the same timeframe.
Who is 
>>>>>>>>>> correct ?
>>>>>>>>>> How does each User update their copy from Master to
get changes 
>>>>>>>>>> from
>>>>>>>>>> another User without losing any pending changes of
their own ? ie 
>>>>>>>>>> is
>>>>>>>>>> it a push up and then trickle down process ?
>>>>>>>>>> What happens when Master is currently being updated
from User1 
>>>>>>>>>> and
>>>>>>>>>> User2 also wants to update ?
>>>>>>>>>> As you can see there is a lot to think about. We wrote
our own
>>>>>>>>>> Synchronization engine to handle these questions plus
>>>>>>>>>> So if it were me I would start writing a detailed spec
>>>>>>>>>> doing
>>>>>>>>>> any coding as that process will help your mind focus
on the 
>>>>>>>>>> problem.
>>>>>>>>>> FWIW
>>>>>>>>>> Leslie.
>>>>>>>>>> On 13/07/2018 6:58 AM, Peter wrote:
>>>>>>>>>>> I am building a database where each user works
on a copy of the 
>>>>>>>>>>> master
>>>>>>>>>>> table. Their copy is located in their own priv
>>>>>>>>>>> Before I start writing a long complicated way of
keeping the 
>>>>>>>>>>> master tbl
>>>>>>>>>>> and each user's priv tbl synchronized, I was wondering
if anyone 
>>>>>>>>>>> had
>>>>>>>>>>> some ideas of how I could approach it.
>>>>>>>>>>> * Should/could I make referential integrity links?
>>>>>>>>>>> * Could I cue a custom method on receiving a network
>>>>>>>>>>> action?
>>>>>>>>>>> * Other?
>>>>>>>>>>> Thank you for any tips.
>>>>>>>>>>> Peter
>>>>>>>>>> ---
>>>>>>>>>> This email has been checked for viruses by AVG.

Copyright © 2004