declare variable - bug or poor design?

[Joe Chasan]

On Thu, Feb 14, 2013 at 11:19:31AM -0500, Kenneth Brody wrote:
> On 2/13/2013 5:26 PM, Joe Chasan wrote:
> > consider the following line of code:
> > declare local test_field(len(1),edit(1))
> [...]
> > BUT - if you do not save the processing table and retokenize, rclerk
> > still has the old length/type (dclerk works).
> [...]
> 
> As Ken Cole correctly pointed out, the field length/type is determined at 
> compile time, not runtime.  (This is true for both dclerk and rclerk.  The 
> difference is that dclerk compiles it when you load it, whereas rclerk uses 
> the pre-compiled file from rcabe.)

2 points and i'll drop it.

1) if you add the feature in to use this construct, and don't document
anywhere that i'd have to recompile any prc's when map that uses one
of the referenced fields changes, why should i have to guess that fp
did not add into rclerk some minimal runtime map/field check to
automagically make this happen?  kinda takes away a lot from the
usefulness.

2) regarding "just use dclerk" - again, i must have fallen asleep at some
point and missed where fp-tech got all the new customers, because most
of mine are longtime customers - a few used quickstart for security, but
most used it because it was the way to go to save money on licensing at
the time.  i guess i must have not realized the point when all these
legacy fp users became the minority.
 
--
-Joe Chasan-                           Magnatech Business Systems, Inc.
joe - at - magnatechonline -dot- com   Hicksville, NY - USA
http://www.MagnatechOnline.com         Tel.(516) 931-4444/Fax.(516) 931-1264