[Bug c++/6994] [3.4 regression] ICE in find_function_data

["reichelt at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=6994


reichelt at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rth at redhat dot com
             Status|RESOLVED                    |REOPENED
           Priority|P3                          |P1
         Resolution|FIXED                       |
            Summary|ICE in find_function_data   |[3.4 regression] ICE in
                   |                            |find_function_data
   Target Milestone|---                         |3.4
            Version|3.1.1                       |3.4


------- Additional Comments From reichelt at gcc dot gnu dot org  2003-09-26 08:24 -------
Richard's patch

  http://gcc.gnu.org/ml/gcc-cvs/2003-09/msg00666.html

makes the testcase testsuite/g++.dg/ext/vla1.C fail again,
but the underlying problem is a deeper one, according to him
(excerpt from a private mail from Richard):

| My patch merely exposed a previous problem via segv.
| 
| It *should* have been failing before.  I don't know why it wasn't.
| Note that by reverting my change, you get no error on line 12.
| 
| The trees being produced by the c++ front end in this case are
| completely bogus.  It is mere accident that the bogus trees were
| being remapped to error_mark_node, and thus silently ignored by
| the middle-end.
| 
| The front end should be generating an error message and forcing
| error_mark_node there.