At 11:26 PM 3/11/98 -0500, Jack Beidler wrote:
>                                    There are still a few defects in
>the system, but it is in reasonable form and I would appreciate your
>suggestions.  I plan to make some improvements before I release the
>software for general use.

>I appreciate your comments.  Suggestions are always welcome.

   Great tool, now for some citicisms.  (Some of these are I am sure
already on your TBD list, but might as well get you the feedback.)

   Package renames confuse your program.  I don't know if it makes sense to
list the renamings at all, but right now, a package renaming inside a unit
gets listed as the containing unit.

   Generic formal parameters, especially types should be listed as such,
(generic formal type), if at all, and as if contained in the generic unit.
I could make a good case for not listing the generic formals at all--you
don't list other formal parameters do you?  But I think that generic formal
types and packages should probably be listed and generic formal subprograms
and objects shouldn't be.

   Can you change your sort order to put operators last?

   Declarations in package private parts should not be listed.  I saw at
least one case, but it may have been due to the package renaming in the
private part.

   It would be very handy to have a list of types, and to be able to create
on the fly a list of all subprograms and generics with parameters or
results of that type.  (Don't ask much do I?  That's really a job for ASIS.)

   I didn't see any task types or task objects listed.  Of course, there
may not have been any in the GNAT library.




                                        Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...