c++/1520: [regression] namespace.exp, gcc HEAD, dwarf-2

[mec dot gnu at mindspring dot com]

>Number:         1520
>Category:       c++
>Synopsis:       [regression] namespace.exp, gcc HEAD, dwarf-2
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jan 19 05:18:00 UTC 2004
>Closed-Date:
>Last-Modified:
>Originator:     mec.gnu@mindspring.com
>Release:        gdb HEAD 2004-01-18 04:30:29 UTC
>Organization:
>Environment:
target=native, host=i686-pc-linux-gnu, osversion=red-hat-8.0
gdb=HEAD 2004-01-18 04:30:29 UTC
cc=gcc HEAD 2004-01-15 13:08:06 UTC
as=binutils 2.14
ld=binutils 2.14
libc=glibc 2.2.93-5-rh
gformat=dwarf-2
glevel=2
>Description:
gdb HEAD is performing worse than gdb 6.0 with namespaces with gcc HEAD -gdwarf-2.

Build gdb.cp/namespace with gcc HEAD -gdwarf-2.  I used both gcc HEAD 2004-01-08 and gcc HEAD 2004-01-15, so it wasn't a recent change in gdb.

Then go into gdb.  break on C::D::marker2 and run there.  Then do these commands:

  (gdb) print c
  (gdb) print cc
  (gdb) print 'C::cc'
  (gdb) print C::cc
  (gdb) print cd
  (gdb) print C::D::cd
  (gdb) print 'E::cde'
  (gdb) print E::cde
  (gdb) print shadow
  (gdb) print E::ce
  (gdb) ptype C
  (gdb) ptype E

gdb 6.0 can resolve some of these names, but gdb HEAD fails on all of them.  This is with the same executable built with gcc HEAD.
>How-To-Repeat:
Do the above steps by hand, or just run namespace.exp with gcc HEAD, -gdwarf-2 (which is the default), and compare the results with gdb 6.0 and gdb HEAD.

Don't use the gdb 6.0 version of namespace.exp.  That way lies incomparable results.  Instead, go into your installation directory for gdb and copy or symlink to the gdb 6.0 executable, to run the same test script with a different gdb executable.  Or smash your version of gdb.cp/ in your 6.0 tree to be the same as gdb HEAD.
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: