switch case question

[Kay C Lan]

On 11/5/06, Dar Scott <dsc at swcp.com> wrote:
>
> The 'switch x' seems to convert x to a string and the 'case y' seems
> to convert y to a string, and then comparisons are made.

Excellent Detective work Dar. It also reinforces a habit I'm trying to
develop when writing switch statements, and that is to always use the
default as a catch for cases I haven't considered:

switch L
  case (5)
     handler1
  break
  case (5.0)
     handler2
   break
  case ("5")
     handler3
  break
  case ("5"&"."&"0")
     handler4
  break
  default
    answer "Switch statement error" & return &
   "Handler: myHandler" & return &
   "Paramater L: " & L
   exitGracefully
end switch

Obviously, from what you're saying, handler3/4 would never run, but
more importantly, if 5. (decimal point byt no zero) got into L, I
should get immediate feedback that I haven't correctly taken it into
account in my switch structure.

I've wasted a lot of time stepping through scripts that don't do what
they're suppose to do and often it is because I've added a feature to
and old stack and don't realise that this new feature now puts new
values into properties that get checked by some buried switch
structure which completely ignores it because it was never part of the
original cases considered. Using default as a catch nails it
instantly.

Hope that helps someone:-)