[Bug c++/53173] New: PROD02

[Bug c++/53173] New: PROD02

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

             Bug #: 53173
           Summary: PROD02
    Classification: Unclassified
           Product: gcc
           Version: 4.6.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: shalja.rudra@gmail.com


Hi All,

I am trying to upgrade (GCC) 4.4.0 to (GCC) 4.6.2.  I see bunch of incompatible
error from code which works with (GCC) 4.4.0 but NOT with (GCC) 4.6.2. 

The machine information is following: Linux version 2.6.32-220.el6.x86_64  (gcc
version 4.4.5 20110214 (Red Hat 4.4.5-6) (GCC) ) 

Encountered errors:

1. error: ‘constexpr’ needed for in-class initialization of static data member
2. error: no matching function for call to ‘std::pair<const long unsigned int,
boost::shared_ptr<const
3. /usr/include/sigc++-2.0/sigc++/signal.h:38:11: error: 'ptrdiff_t' does not
name a typeFix: #include <cstddef> 
4. error: no matching function for call to ‘make_pair(std::string&,
std::string&)’
5. error: declaration of ‘~typename
6. error: call of overloaded ‘isnan(double&)’ is ambiguous

e.t.c

I  do refer https://wiki.edubuntu.org/GCC4.6 to fix some of the issue. I
rebuilt boost_1_47_0,  SQLAPI-3.7.35, etc. with (GCC) 4.6.2 as well to remove
incompatibilty between these.


I am suspicious if some of the issue is already fixed in (GCC) 4.6.3 (already
released). Please let me know if we can use (GCC) 4.6.3 instead of (GCC) 4.6.2.
Also please point  me from where we can refer to get some more information to
fix our code copilation and linking issue with  (GCC) 4.6.2.

Thanks

[Bug c++/53173] PROD02

[marc.glisse at normalesup dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #1 from Marc Glisse <marc.glisse at normalesup dot org> 2012-04-30 20:02:59 UTC ---
Uh, where are you reporting a bug in gcc?

(In reply to comment #0)
> I am trying to upgrade (GCC) 4.4.0 to (GCC) 4.6.2.  I see bunch of incompatible
> error from code which works with (GCC) 4.4.0 but NOT with (GCC) 4.6.2. 

Yes, g++ becomes better at detecting illegal code.

> 1. error: ‘constexpr’ needed for in-class initialization of static data member

Are you using -std=c++0x? Why?

> 2. error: no matching function for call to ‘std::pair<const long unsigned int,
> boost::shared_ptr<const

Obviously missing most of the message.

> 3. /usr/include/sigc++-2.0/sigc++/signal.h:38:11: error: 'ptrdiff_t' does not
> name a typeFix: #include <cstddef> 

actually stddef.h if you want ptrdiff_t and not just std::ptrdiff_t (unless
there is a using namespace std, as 6. makes me fear)

> 4. error: no matching function for call to ‘make_pair(std::string&,
> std::string&)’

#include <utility>

> 5. error: declaration of ‘~typename

Missing most of the message again

> 6. error: call of overloaded ‘isnan(double&)’ is ambiguous

PR48891 maybe?

> I  do refer https://wiki.edubuntu.org/GCC4.6 to fix some of the issue. I
> rebuilt boost_1_47_0,  SQLAPI-3.7.35, etc. with (GCC) 4.6.2 as well to remove
> incompatibilty between these.

Gcc release notes often also contain relevant information, too.

> I am suspicious if some of the issue is already fixed in (GCC) 4.6.3 (already
> released).

What do you mean, fixed? The bugs are in your code.

> Please let me know if we can use (GCC) 4.6.3 instead of (GCC) 4.6.2.

Sure, more bugs fixed.

[Bug c++/53173] PROD02

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #2 from gcc <shalja.rudra at gmail dot com> 2012-04-30 20:31:49 UTC ---
Thanks Marc !

Sorry on missing most words in few error lines.

2. error: no matching function for call to ‘std::pair<const long unsigned int,
boost::shared_ptr<const a::b::c::dConfig> >::pair(long unsigned int, const
a::b::c::Config*)’

5. error: declaration of ‘~typename a::b::c::dControl<Data, PublicSecurityID,
PrivateSecurityID, ControlQueue, StatusQueue, Queue>::dControl’ as member of
‘a::b::c::dControl<Data, PublicSecurityID, PrivateSecurityID, FeedControlQueue,
StatusQueue, Queue>’

Please suggest.

Thanks

[Bug c++/53173] PROD02

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-05-01 09:36:50 UTC ---
(In reply to comment #2)
> 2. error: no matching function for call to ‘std::pair<const long unsigned int,
> boost::shared_ptr<const a::b::c::dConfig> >::pair(long unsigned int, const
> a::b::c::Config*)’

There is no implicit conversion from Config* to shared_ptr<Config>, it was a
bug in GCC that the code compiled previously.

You need to create the pair with a shared_ptr<Config> as the second argument,
not a Config*

> 5. error: declaration of ‘~typename a::b::c::dControl<Data, PublicSecurityID,
> PrivateSecurityID, ControlQueue, StatusQueue, Queue>::dControl’ as member of
> ‘a::b::c::dControl<Data, PublicSecurityID, PrivateSecurityID, FeedControlQueue,
> StatusQueue, Queue>’

You're declaring a destructor with the wrong type.

[Bug c++/53173] PROD02

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-05-01 09:41:34 UTC ---
Also, please give more useful bug descriptions, "PROD02" is completely
meaningless

[Bug c++/53173] PROD02

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-05-01 09:40:53 UTC ---
(In reply to comment #0)
> 4. error: no matching function for call to ‘make_pair(std::string&,
> std::string&)’

If you're calling make_pair with an explicit template argument list e.g. 
  pair<string&, string&> p = make_pair<string&, string&>(s1, s2);
then that won't work in C++11

Just construct a pair directly, it's pointless to use make_pair if you don't
want to deduce the argument types:

  auto p = pair<string&, string&>(s1, s2);
or
  pair<string&, string&> p(s1, s2);

[Bug c++/53173] GCC 4.6.2 errors

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #6 from gcc <shalja.rudra at gmail dot com> 2012-05-02 14:53:25 UTC ---
Thanks for the point, Jonathan ! I will make changes accordingly in our code.

(In reply to comment #4)
> (In reply to comment #0)
> > 4. error: no matching function for call to ‘make_pair(std::string&,
> > std::string&)’
> If you're calling make_pair with an explicit template argument list e.g. 
>   pair<string&, string&> p = make_pair<string&, string&>(s1, s2);
> then that won't work in C++11
> Just construct a pair directly, it's pointless to use make_pair if you don't
> want to deduce the argument types:
>   auto p = pair<string&, string&>(s1, s2);
> or
>   pair<string&, string&> p(s1, s2);

[Bug c++/53173] GCC 4.6.3 errors

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

gcc <shalja.rudra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|4.6.2                       |4.6.3
            Summary|GCC 4.6.2 errors            |GCC 4.6.3 errors
              Alias|                            |GCC4.6.3_and_std::a

--- Comment #7 from gcc <shalja.rudra at gmail dot com> 2012-05-04 19:36:30 UTC ---
Hi All,

Now I am using GCC 4.6.3 and fixing our code to work with. The following
linking error got reported.

I looked on http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49445 and according to
this GCC 4.6.3 doesn't have it. Please suggest.

boost_1_47_0/boost/lockfree/detail/freelist.hpp:124: undefined reference to
`std::atomic<boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>
>::operator
boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>()
const'

boost_1_47_0/boost/lockfree/detail/freelist.hpp:124: undefined reference to
`std::atomic<boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>
>::operator
boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>()
const'

/boost_1_47_0/boost/lockfree/detail/freelist.hpp:124: undefined reference to
`std::atomic<boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>
>::operator
boost::lockfree::detail::tagged_ptr<boost::lockfree::detail::freelist_node>()
const'

collect2: ld returned 1 exit status

Thanks

[Bug c++/53173] GCC 4.6.3 errors

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

gcc <shalja.rudra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Alias|GCC4.6.3_and_std::a         |GCC4.6.3_errors

--- Comment #8 from gcc <shalja.rudra at gmail dot com> 2012-05-04 19:37:53 UTC ---
Fixing the alias of bugzilla update.

[Bug c++/53173] GCC 4.6.3 errors

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-05-04 19:47:09 UTC ---
Bugzilla is not a help forum.  Please find somewhere more appropriate to ask
how to solve your problems, such as the boost mailing list or the gcc-help
list.

[Bug c++/53173] GCC 4.6.3 errors

[shalja.rudra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53173

--- Comment #10 from gcc <shalja.rudra at gmail dot com> 2012-05-04 20:11:50 UTC ---
I looked into the similar post in boost mailing list
http://lists.boost.org/Archives/boost/2011/07/184266.php and it directed to
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49445 

The email got bounced before using the gcc-help@gcc.gnu.org, may be the email
was not plain text format . It is fine now.

Thanks



(In reply to comment #9)
> Bugzilla is not a help forum.  Please find somewhere more appropriate to ask
> how to solve your problems, such as the boost mailing list or the gcc-help
> list.