Re: Trigger question -- really batting a thousand

[CRPence <CRPbottle@xxxxxxxxx>]

James Lampert wrote:
> The easiest way to veto a change from a "before" trigger is what?
> Cause the trigger program to fail with a fatal error?
> I know I have an example of a trigger program that vetoes everything
> unconditionally (I wrote one to exercise QuestView's trigger veto
> detection feature some years ago), but I can't remember what it did,
> or where I put it.

  Optionally send a diagnostic message, for example to the requester of 
the I/O, and then... Send an Escape message to the program previous to 
the trigger program; i.e. to the DB I/O program.  When that DB I/O 
program receives the escape message, that will effect termination of the 
trigger and manifest that as an I/O error to the program that requested 
the I/O.  Due to implementation details, that escape message will become 
irretrievable; it will not be visible, neither in a trace nor joblog, 
and why the diag message may be valuable if the I/O requester needs some 
information about why the I/O was prevented\vetoed.

Regards, Chuck