> The ADO components were some of the worst and buggiest components Borland
> has produced so I surely am not going to model IBO after them.

*g* ... well, I can imagine that, yes. But at least they don't hide errors
:-)

> I could make a ColumnAttribute you set once at the connection level or
> dataset level and then it causes all string based assignments to validate

in

> this way.

Somewhere along the lines of that would be great. As I said: at least that
won't
hide the error!

I just happened to stumble across this while testing something... I wonder
if
this is the expected behaviour for IBO users?!

Martijn Tonies
Database Workbench - tool for InterBase, Firebird, MySQL, Oracle & MS SQL
Server
Upscene Productions
http://www.upscene.com
Database development questions? Check the forum!
http://www.databasedevelopmentforum.com

> > > I designed it to work the way it does. To my knowledge
> > other component
> > sets
> > > also work this way.
> >
> > Well, ADO, for one, doesn't.
> >
> > > The best way to get an error message is to write a generic
> > validation
> > > routine. It could be a procedure that assigns an open
> > string value to a
> > > TIB_Column and inside it will do the check.
> >
> > The thing is, as I wrote earlier, as a user, I do not know -
> > until I get
> > back my data
> > from the database - that is has been cut off, while if I
> > would have gotten a
> > database
> > error, it would have been clear from the start.
> >
> > Could this behaviour be optional? (default ON not to annoy
> > other IBO users,
> > or
> > perhaps default OFF to make sure errors happen when they should?)