Re: Record Lock under Commitment Control

Hello John,

Am 24.02.2022 um 20:29 schrieb John Yeung <gallium.arsenide@xxxxxxxxx>:

This issue shows to me: Things are often not that easy as they look at a first glance. ;-)

True. However, my take on RPG and display file mechanics is that they're designed to be straightforward to use for straightforward cases.

I think this is a good overall approach.

When one user has a database record open on their screen, and that user has the capability to update that record, then it seems to me that the simplest, safest, and most obvious thing to do is to not let any other user open that record for update. And that is exactly the behavior you described.

Yes, that's how I expected things to happen, and exactly how I want it wo work. My mishap was (probably, still need to verify) due to not updating the record but skipping the update (because the DSPF contents weren't changed), the lock was still existing through the former CHAIN. And I assumed that the lock would still exist *if* the update was done.

It's fairly unsophisticated, but effective. And if you stick to this model, it *is* about as easy as it looks.

Trees. Many of them. Until you can't see the forest. ;-)

:wq! PoC