Re: Export _Prime_rehash_policy symbols

[François Dumont <frs dot dumont at gmail dot com>]

On 01/30/2013 11:02 AM, Paolo Carlini wrote:
> Hi,
>
> On 01/29/2013 10:22 PM, François Dumont wrote:
> > Hi
> >
> >     This long standing issue has been reported to my attention by 
> > Paolo some weeks ago. This is the patch to export 
> > _Prime_rehash_policy symbols.
> >
> > 2013-01-29  François Dumont  <fdumont@gcc.gnu.org>
> >
> >     * include/bits/hashtable_policy.h
> >     (_Prime_rehash_policy::_M_next_bkt)
> >     (_Prime_rehash_policy::_M_need_rehash): Move definition...
> >     * src/c++11/hashtable_c++0x.cc: ... here.
> >     * config/abi/gnu.ver (GLIBCXX_3.4.18): Export _Prime_rehash_policy
> >     symbols.
> >
> >     Tested under Linux x86_64.
> >
> >     Ok for trunk ?
> Ok, but please add a one-line comment before the
>
> #include <unordered_set>
>
> clarifying that it's a bit it hackish including it and in principle 
> we'd like to include some sort of hashtable*.h.
> >     I also notice a weird thing in hashtable_c++0x.cc. It is 
> > including shared/hashtable-aux.cc within namespace std scope. However 
> > hashtable-aux.cc includes <bits/config.h> itself. If 
> > hashtable_c++0x.cc was not including <bits/config.h> it would fail to 
> > compile. Do you want me to remove <bits/config.h> include from 
> > shared/hashtable-aux.cc at the same time ?
> Yes, I agree.
>
> Thanks!
> Paolo.
I have limited the inclusions to:

#include <initializer_list>
#include <tuple>
#include <bits/hashtable_policy.h>

rather than <unordered_set>

I have tested it with success. Now I am going to check with the 
--enable-symvers option and will resubmit the patch.

François