#6 No error checking?

[Paul]

How can I tell if a call to ORM was successful?

Example 1)
If I successfully fetch a record and then attempt to
update it, and someone behind the scenes has already
deleted it, I would expect to get an error when I then
try to commit my changes. I do not receive an error.

Example 2)
If I successfully fetch a record and then attempt to
delete it, and someone behind the scenes has already
deleted it, I would expect to get an error when I then
try to delete it. I do not receive an error.

Example 3)
If I successfully fetch a record and update it, when I
issue the COMMITALL command, I would expect an error
if someone else has already updated the record. I do not
receive an error.

[Sven Aelterman]

Logged In: YES
user_id=1096815

This is basically the support request I opened.

It's not error checking, it's checking for concurrency issues
such as lost updates, etc.

[Nobody/Anonymous]

Logged In: NO

The program operates under the approach of optimistic
locking. While it is possible to implement more complex
locking, traditionally this kind of page locking impedes the
greater good. If person A gets a record, person B gets a
record, person A saves a record and then person B saves
record well you'd end up with what person B put in anyway.
99% of applications can operate under this safe assumption.
However, I feel your pain but know that implementing more
complex lock handling will only take away from the
performance of this solution.