Board index » delphi » Major,Major Bde FLAW...row locking..Developers Beware!

Major,Major Bde FLAW...row locking..Developers Beware!

Don,

Please do not make such false and misleading statements without having a
good understanding
of the issues you are referring too. It is obvious from your message that
you are a bit
confused with what optimistic locking is and how it is supposed to work.

Just for the record the Optimistic locking scheme of the BDE works just
like
it should. How the BDE behaves, also depends on the back-end used.
MS SQL SERVER and SYBASE SQL SERVER for example only support page locks.
Any serious client/server developer should consider the locking mechanisms
of
the back-end and develop the front-end accordingly.

I have been doing real-world client/server 2-tier and n-tier for the past
7+ years and
I can confidently recommend the Borland products to any serious C/S server
developer.

Hope this helps.

Nick Liristis

Don Higgins <dhigg...@wrld.com> wrote in article
<33C6541A.1FA3F...@wrld.com>...

Quote
> There  is a fatal flaw in the Optimistic locking scheme of the BDE. It
> _assumes_ that the record you are editing will not be changed by the
> time you want to post it. Boo. Hiss.

> This kind of locking is great for speed & reports, but not so great when
> you are the one who has spent a fair amount of time working on a record
> only to find out that you cannot post your changes.

> All the BDE would need to do would be to provide a means to test the
> lock status of the record! We could all handle it from there.

> Borland's comments on optimistic locking are that most SQL servers these
> days do not provide a row locking mechanism.
> HUH?

> This situation has presented a major problem for the application we are
> currently involved in.  Borland advertises its development tools as
> being for the serious Client/Server developer, but the serious Cl/Svr
> developer doesn't use dbase or Paradox. And NEEDS to know up front If a
> row is LOCKED!!

> Come on, Borland. You've seen enough complaints on this issue, and all
> I've seen so far is ....
> ..."uh, see the explanation on optimistic locking".

> This makes me pessimistic about continuing to use Borland products.

> ----------------------------------------------
>                  Don Higgins
>         Applications Development
>            Auto Europe, Inc.

>      39 Commercial Street, Portland, Maine 04112 USA
>      Phone: +(1) 207 842 2052
>      Fax:       +(1) 207 842 2239
>      http://www.wrld.com/ae/

 

Re:Major,Major Bde FLAW...row locking..Developers Beware!


Hi

Sorry that is not a bug that is the way 'optimistic' schemes work. You're
betting that most times the record won't have changed by the time you are
ready to post it. If it has, you inform the user and give them a chance to
re-edit the data or whatever.

Many C/S databases do not have row locking, thus they try to keep things
unlocked as much as possible.

For instance Sybase 10 and MS SQL 6 do not have row locking.

Regards, Frederick C. Wilt

Re:Major,Major Bde FLAW...row locking..Developers Beware!


No, I am not confused about what optimistic locking is..
Nor did I mention the word "bug".

I did, however mention the word flaw. The flaw, as mentioned in the
original post, is that there is no mechanism in the BDE to test the lock
status of a row;  which I  _do_ consider to be a very important issue.
Apparently so do many other developers as I have seen comments about
this all over the place.

The BDE is not a bad, product...don't get me wrong; it just fine for
most situations...there's just a piece missing, which has not been made
obvious, is a big piece,  does not appear to be being addressed by
Borland.

I feel confident that if there were a mechnism to do this, you would
find a use for it right away.

If  we don't bring these issues into the light, they won't get
addressed.

Other Threads