[boost] Re: BOOST_NO_EXPLICIT_FUNCTION_TEMPLATE_ARGUMENTS and gcc

From: David Abrahams (dave_at_[hidden])
Date: 2003-08-15 15:52:05

"John Maddock" <jm_at_[hidden]> writes:

>>Currently, BOOST_NO_EXPLICIT_FUNCTION_TEMPLATE_ARGUMENTS
>>is not defined for gcc. However, the following URL in the gcc bug
>>database
>>
>>http://gcc.gnu.org/bugzilla/show_bug.cgi?id=7676
>>
>>leads me to believe that the macro should be set on for the appropriate
>>versions of gcc. Matter of fact, I run with this problem myself and it
>>can be
>>workedaround with techniques similar to those employed for MSVC. See
>>for instance definitions of get() and workaround_holder in
>>
>>boost/tuple/detail/tuple_basic_no_partial_spec.hpp
>>
>
> Thanks,
>
> The issue with gcc seems to be a little more specific than we normally set
> the macro for, but I don't see any reason why we shouldn't set it.  Am I
> right in thinking that this is specific to gcc 3.1 and 3.2?  Also do you
> have a test case that can be added to the appropriate config test?

We need a more-specific macro.  The problem is that when the function
*doesn't* have type and non-type parameter overloads, the workaround
breaks those very same versions of GCC.  See
http://boost.sourceforge.net/regression-logs/cs-Linux/developer_result_page.html#crc.
(http://tinyurl.com/k5or)

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com