Re: I4GL 4.10 input array annoyance

Mike, See your point, not much you can do to overcome this problem. The only suggestion I have, which still means you have to recompile your apps, is to use some form of pre-processor, like cpp/m4 to define macros that perform your validation. In this way you will only have one place to change the code, but still need to compile all the apps. Not as simple a solution as you might require, but it will work. }Mark.Denham@bbc.co.uk (Mark Denham) wrote: }>Try defining the fields in the form as formonly, without any constraints on }>them. Do all the necessary validation (not null, ranges etc) }>programatically. }I was afraid you were going to say that. Just as I get my syscolval/att }clean }and manageable, I find I can't use it thanks to another Informix feature. }I've used the formonly tactic for selection windows, but I really wanted to }avoid it for database fields. }>Do any complete row checking (for blank rows) in the after row and after }>input statements. You will also need to put the INPUT ARRAY inside a WHILE }>loop if you want to prevent them leaving a row if the after row checks throw }>up a problem. You will also need a NO ENTRY field, defined as the last one }>in the row, that you can use to force input array to skip to next row. }>Skeleton code below: }> }>LET exit_input = FALSE }>LET curr_row = 1 }>WHILE NOT exit_input }> }> INPUT ARRAY }> BEFORE ROW }> IF( curr_row > arr_curr() ) THEN }> NEXT FIELD no_entry_field }> END IF }> LET curr_row = arr_curr() }> }> AFTER INPUT ^^^^^ | }After ROW you mean? :-) ^^^^^^^^^^^^^^^^^^^^^^ My mistake. }Yes, I am aware of this construct, and in fact, it is already set up that way }in my generated code. I guess my real concern is duplicating all that }validation code throughout my applications. That's why I was using the }syscolval and syscolatt--so I would have one place to make changes, }then apply them with a quick form compile of all my applications. }Am I SOL here? }Mike Mark Denham BBC London,UK Mark.Denham@bbc.co.uk