"bad deriv in da" error

From: Ian Edwards (ian_at_pandora.panorama.com)
Date: Thu Jul 09 1998 - 11:10:51 BST

  • Next message: Larry McVoy: "Re: Thank you!" 
    ----- Begin Included Message -----

>From ups-users-approval_at_ukc.ac.uk Wed Jul  8 08:59:41 1998
Return-Path: <ups-users-approval_at_ukc.ac.uk>
To: ups-users_at_ukc.ac.uk
Subject: "bad deriv in da" error
From: bob_at_fla.fujitsu.com (Bob Carragher)
Reply-To: bob_at_fla.fujitsu.com (Bob Carragher)
Date: Tue, 07 Jul 1998 22:42:21 -0700
Sender: bob_at_fla.fujitsu.com

Hello!

I am brand new to this mailing list, so if I commit any
faux pas or should be consulting a well-known FAQ, please
forgive me and point me in the right direction!

I have been using UPS for years, and it has been quite
robust until recently.  Currently, I am trying to debug a
relatively large (for me) program (roughly 60K lines) and
when I try to expand on variables of certain types
(typedef-ed by me), UPS core dumps with the following
message:

     Fatal internal error: bad deriv in da (aborting) ...
     Dumping core ... Abort (core dumped)

It is very consistent in that it will core dump whenever
I try to explore any variable or field of certain typedefs.
(By "explore," I mean highlighting that variable or field
and then clicking the "Expand" button.)  Funnily enough,
if I try to explore these variables/fields by explicitly
naming a subfield of that typedef, UPS has no problem.

Other oddities include:  it core dumps only for certain
typedefs, not all of them; the set of typedefs on which
UPS core dumps _changes_ as I make large modifications to
the program (e.g. adding more typedefs).

My questions are these:

     1.  Has anyone on this list ever seen this problem
         before?  Is it a bug or feature of UPS?  B-)
     2.  Can anyone recommend a cure (other than "don't
         do it"), or is this a hopeless problem?

I am using UPS on C code.  The platform is an UltraSparc 2,
single-processor.  I use gcc to compile my program (I also
used gcc to make UPS).

Any help would be greatly appreciated!

Thanks!

				Bob Carragher


----- End Included Message -----

    This archive was generated by hypermail 2.1.4 : Wed Feb 13 2002 - 21:55:19 GMT