public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug c++/65189] New: Malformed (C++) class-hierarchy dump on abstract class (in comparission to GCC 4.6.x)
@ 2015-02-24 11:33 meisenmann.lba@fh-salzburg.ac.at
  2015-02-24 12:49 ` [Bug c++/65189] " jakub at gcc dot gnu.org
  0 siblings, 1 reply; 2+ messages in thread
From: meisenmann.lba@fh-salzburg.ac.at @ 2015-02-24 11:33 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65189

            Bug ID: 65189
           Summary: Malformed (C++) class-hierarchy dump on abstract class
                    (in comparission to GCC 4.6.x)
           Product: gcc
           Version: 4.9.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: meisenmann.lba@fh-salzburg.ac.at

Following minimal sample (Ie. test.cpp) does generate a noticeable different
class-hierarchy dump with GCC 4.9.2 - compared to GCC 4.6.4:

class A
{
public:
 virtual ~A();
 virtual int getVal() const = 0;
};

Calling
 gcc -c -fdump-classhierarchy test.cpp && type test.cpp.002t.class

a) generates with GCC 4.6.4 (as expected):

Vtable for A
A::_ZTV1A: 5u entries
0   (int (*)(...))0
4   (int (*)(...))(& _ZTI1A)
8   (int (*)(...))A::~A
12  (int (*)(...))A::~A
16  (int (*)(...))__cxa_pure_virtual

Class A .. dropped; Ie. no difference

b) with GCC 4.9.2, the output is

A::_ZTV1A: 5u entries
0   (int (*)(...))0
4   (int (*)(...))(& _ZTI1A)
8   (int (*)(...))0u
12  (int (*)(...))0u
16  (int (*)(...))__cxa_pure_virtual

The two d'tor-entries (offset 8 and 12) are 0u ?!

This issue also affects the generated dump with option
"-fdump-translation-unit"
... which is used by the ABI Compliance Checker (ABICC).
For example, if I compare the ABI-compatibility of abstract classes between
different GCC-versions, a incorrect/false-positive mismatch of the vtable
(concerning the virtual d'tor entries) is determined.

I've already tested some additional options to dtermine a work-around, but
without luck!

Best regards from Salzburg,
Markus


^ permalink raw reply	[flat|nested] 2+ messages in thread

* [Bug c++/65189] Malformed (C++) class-hierarchy dump on abstract class (in comparission to GCC 4.6.x)
  2015-02-24 11:33 [Bug c++/65189] New: Malformed (C++) class-hierarchy dump on abstract class (in comparission to GCC 4.6.x) meisenmann.lba@fh-salzburg.ac.at
@ 2015-02-24 12:49 ` jakub at gcc dot gnu.org
  0 siblings, 0 replies; 2+ messages in thread
From: jakub at gcc dot gnu.org @ 2015-02-24 12:49 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65189

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |jason at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This changed with r208845.


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2015-02-24 12:18 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-02-24 11:33 [Bug c++/65189] New: Malformed (C++) class-hierarchy dump on abstract class (in comparission to GCC 4.6.x) meisenmann.lba@fh-salzburg.ac.at
2015-02-24 12:49 ` [Bug c++/65189] " jakub at gcc dot gnu.org

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).