Re: [Boost-users] thread/call_once compile error on trunk on my code

Subject: Re: [Boost-users] thread/call_once compile error on trunk on my code
From: Vicente J. Botet Escriba (vicente.botet_at_[hidden])
Date: 2013-04-20 11:47:48

Le 20/04/13 17:03, Olaf Peter a écrit :
> Hello,
>
> The error is about
>
> /usr/local/include/boost/thread/detail/invoke.hpp: In function 
> 'typename boost::enable_if_c<(! 
> boost::is_member_function_pointer<Fp>::value), Ret>::type 
> boost::detail::invoke(Fp&&, Args&& ...) [with Ret = void, Fp = void 
> (*)({anonymous}::id_string&), Args = {{anonymous}::id_string}, 
> typename boost::enable_if_c<(! 
> boost::is_member_function_pointer<Fp>::value), Ret>::type = void]':
> /usr/local/include/boost/thread/pthread/once_atomic.hpp:127:9: 
> instantiated from 'void boost::call_once(boost::once_flag&, 
> Function&&, ArgTypes&& ...) [with Function = void 
> (*)({anonymous}::id_string&), ArgTypes = {{anonymous}::id_string&}]'
> /home/olaf/work/token_ids.cpp:93:59:   instantiated from here
> /usr/local/include/boost/thread/detail/invoke.hpp:406:65: error: 
> invalid initialization of non-const reference of type 
> '{anonymous}::id_string&' from an rvalue of type '{anonymous}::id_string'
>
>
> Is this intentionally or do I miss something?
>

Hi,

I see it breaks also code on msvc.

I'm running the regression test on my computer with the attached patch.
Could you tell me if it works for you?

Best,
Vicente