c++/9143: 64BIT: Error in dump class hiearchy file generated for basic inheritance
Christian Ehrhardt
ehrhardt@mathematik.uni-ulm.de
Fri Jan 3 10:38:00 GMT 2003
http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9143
The following mail form Janis about this PR got accidentally picked up
by GNATS as a new PR. Let's hope that this mail makes it into the audit
trail of the right PR.
Is the generated code incorrect also, or just the dump
output?
GCC generates the same dump information as that reported in
the PR for a powerpc64-linux cross compiler built from the
current mainline and from the 3.3 branch.
With a native powerpc64-linux compiler I get:
Vtable for B
B::_ZTV1B: 7 entries
0 8
8 0
16 &_ZTI1B
24 0
32 0fffffff8
40 &_ZTI1B
48 A::f()
With a native 32-bit compiler on a powerpc64-linux system
I get:
Vtable for B
B::_ZTV1B: 7 entries
0 4
4 0
8 &_ZTI1B
12 0
16 4294967292
20 &_ZTI1B
24 A::f()
For comparison purposes I got the same dumps for two other
targets. With the mainline from 20021210 on ia64-linux:
9143.C:1: sorry, unimplemented: `fdesc_expr' not supported
by dump_expr
9143.C:1: sorry, unimplemented: `fdesc_expr' not supported
by dump_expr
9143.C:5: sorry, unimplemented: `fdesc_expr' not supported
by dump_expr
9143.C:5: sorry, unimplemented: `fdesc_expr' not supported
by dump_expr
I didn't look into this, but it suggests that perhaps some
of the dump functionality used by the cross compiler isn't
supported.
With the mainline from 20021220 on i686-linux:
Vtable for B
B::_ZTV1B: 7 entries
0 4
4 0
8 &_ZTI1B
12 0
16 0fffffffc
20 &_ZTI1B
24 A::f()
This information might be useful for someone who understands
vtable layout and the dump functions.
More information about the Gcc-bugs
mailing list