Re: Forms 4.5 bug?

Tyr putting your logic in a block (master block) level POST-INSERT, POST-UPDATE, or POST_COMMIT. One of these should handle all your attempts to commit data to the database without doing the check.

In article <4mrqff$83d_at_dfw-ixnews4.ix.netcom.com>, maricobb_at_ix.netcom.co says...
>
>I have a form level KEY-COMMIT trigger that ensures at least one detail
>record exists on a master-detail form. Works fine; the trigger failure
>is raised with a message, except if the next thing a user does is to
>hit the exit button, or choose action/exit, Oracle presents its little
>dialog asking them if you want to save your changes. Oracle allows it,
>bypassing the KEY-COMMIT logic apparently. I ran the thing in debug,
>and :system.form_status was 'CHANGED', but there was no obvious
>indication that an error had occurred. When do those system error
>variables get set?
>
>Am I missing something? Do you have to re-do all your edits in the
>KEY-EXIT logic too? I tried passing the no-validate option into the
>exit_form built in if :system.form_status <> 'QUERY' (as recommended by
>online help) - but that trashes valid pending updates too. I only want
>to trash invalid updates.
>
>Is it still true that a user can back-tab out of a field and bypass the
>WHEN-VALIDATE logic?
>
>Rules of thumb in what triggers to use where would be GREATLY
>appreciated.
>
>Mari Cobb
>maricobb_at_ix.netcom.com
Received on Fri May 10 1996 - 00:00:00 CEST