Weird button behavior

[J. Landman Gay]

On 2/5/11 5:31 PM, Peter Haworth wrote:
> I just figured out what was going on and I guess it's worth
> mentioning since it points out the dangers of testing modal dialogs.
>
> In the preOpenCard handler for the modal dialog there was this
> statement:
>
> set the width of this card to 425
>
> This, of course, is not a valid statement since cards don't have a
> width property.

They do, actually, but it is read-only. You can't set it.

>
> So I guess now my question for the community is - What techniques do
> you use to test modal dialogs?

I debug first without using a modal, like you did. That catches most 
stuff. Then if it behaves differently while modal, sometimes placing the 
word "breakpoint" on a line works, where setting a red-dot temporary 
breakpoint doesn't. The engine interprets "breakpoint" and will halt the 
script when possible; the IDE interprets the dots and can't always 
intercept. If that fails, you can resort to putting info into the 
message box after every suspicious statement. In particular, putting the 
result works well for lots of things. But if the script is aborting, you 
might have to wrap it in a "try" structure instead:

try
  set the width of this card to 425
catch tErr
  put "Setting width:" && tErr
end try

Usually testing the stack as a non-modal first catches almost everything 
though.

-- 
Jacqueline Landman Gay         |     jacque at hyperactivesw.com
HyperActive Software           |     http://www.hyperactivesw.com