I should add that (in our test account) I switched to the READU statement along with a LOCKED clause ->
READU STU.REC FROM STU.FILE, STU.ID LOCKED CRT STU.ID:" RECORD IS LOCKED" END THEN STU.REC<48> = INST.CRED STU.REC<49> = INET.CRED WRITE STU.REC ON STU.FILE, STU.ID END I locked one of the records in the list I was processing. When the program was working on that record, the CRT statement was executed, and the program moved right along processing the rest of the list; and record I had locked was not updated. Looking through the 'UniBasic Commands Reference' PDF, I found 'RECORDLOCKED (file.var, rec.id.expr)'. Perhaps this could be used within a LOOP to control things, allowing terminal input where the user can choose what to do about the record lock? I've decided to write the data to a new file, so this is no longer a big issue for me. If I had created the program using our vendor's tool kit, the generated code would have added all the appropriate gunk the system uses for record locks. The responses have been interesting as I've been looking for a chance to better understand how Unidata shops handle locks. Thank you, Jon Wells ________________________________ From: Tony Gravagno <3xk547...@sneakemail.com> To: u2-users@listserver.u2ug.org Sent: Friday, February 8, 2013 2:30 AM Subject: Re: [U2] Record locks (Unidata 7.2) I think people understand exactly what you're saying, but disagree. Most applications don't use the more rigorous coding for having a Locked clause that provides additional information to the user, retries, etc. That's not "wrong", it's just not necessary ... depending on the site. That said, when we used to have a lot of people in an office and a dumb terminal beeping was a recognized message to everyone, it was OK (generally) to just do the READU. These days with more alternative UIs, thick-client, thin-client, web services, mobile, etc, there's no one to hear that beep. So for modern development I would agree that you really don't want an unconditional lock hanging around, you want a Locked clause with logic to rollback the transaction and exit gracefully, perhaps with a message sent to the site admin. The penalty for Not having a Locked clause with a new UI is that the port will hang, the UI will receive a timeout (which many clients aren't prepared to process), and the DBMS port could be sitting there indefinitely with no one aware of the condition. This can cascade to multiple ports until suddenly everything grinds to a halt and GUI users around the world are unable to access the app. I believe some heads will be nodding here because I hear about this happening occasionally. T > From: Wjhonson > Then you're misunderstanding what I'm saying. > From: Kevin King > +1. Well stated. > > Woodward, Bob wrote: > > In an active user session where users are obtaining a lock for a > > length of time, you're right. There are lots of times, though, that > > lock intervention is not a benefit. It all depends on the need of the > > application and many situations where I would highly recommend > > using the default locking system of READU. > > From: Wjhonson > > I wouldn't however *recommend* this approach to locking as the user > > can never tell if their terminal is waiting on a lock, or just "hung" > > in some other odd manner, or perhaps processing a batch of something. > > > > So it's rather a nasty trick to play on the user, to just use READU > > without a LOCKED clause _______________________________________________ U2-Users mailing list U2-Users@listserver.u2ug.org http://listserver.u2ug.org/mailman/listinfo/u2-users _______________________________________________ U2-Users mailing list U2-Users@listserver.u2ug.org http://listserver.u2ug.org/mailman/listinfo/u2-users
