bind/bind_template.hpp + EDG 238 trouble + patch

Hi, The changes to boost/bind/bind_template.hpp from 10/25 break the Boost.Python build under IRIX/MIPSpro 7.3.1. The problem is due to a known bug in the old EDG frontend (it is the obscure "ambiguous by inheritance" bug). Attached is a simple patch that restores the Boost.Python build. I've tested the patch on 18 other platforms, incl. Windows, Mac OS X, Tru64 Unix, and a bunch of Linux distributions. If the patch looks reasonable and I get an OK I'd be happy to check it in. Thanks! Cheers, Ralf Index: bind_template.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/bind/bind_template.hpp,v retrieving revision 1.6 diff -u -r1.6 bind_template.hpp --- bind_template.hpp 25 Oct 2005 21:01:41 -0000 1.6 +++ bind_template.hpp 28 Oct 2005 17:28:28 -0000 @@ -38,7 +38,8 @@ BOOST_BIND_RETURN l_(type<result_type>(), f_, a, 0); } -#if !defined(BOOST_NO_FUNCTION_TEMPLATE_ORDERING) +#if !defined(BOOST_NO_FUNCTION_TEMPLATE_ORDERING) \ + && !BOOST_WORKAROUND(__EDG_VERSION__, <= 238) template<class A1> result_type operator()(A1 const & a1) { @@ -66,7 +67,8 @@ BOOST_BIND_RETURN l_(type<result_type>(), f_, a, 0); } -#if !defined(BOOST_NO_FUNCTION_TEMPLATE_ORDERING) +#if !defined(BOOST_NO_FUNCTION_TEMPLATE_ORDERING) \ + && !BOOST_WORKAROUND(__EDG_VERSION__, <= 238) template<class A1, class A2> result_type operator()(A1 const & a1, A2 & a2) { __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com

Ralf W. Grosse-Kunstleve wrote:
Hi,
The changes to boost/bind/bind_template.hpp from 10/25 break the Boost.Python build under IRIX/MIPSpro 7.3.1. The problem is due to a known bug in the old EDG frontend (it is the obscure "ambiguous by inheritance" bug). Attached is a simple patch that restores the Boost.Python build. I've tested the patch on 18 other platforms, incl. Windows, Mac OS X, Tru64 Unix, and a bunch of Linux distributions.
If the patch looks reasonable and I get an OK I'd be happy to check it in.
Please go ahead. The only potential minor problem with it might be g++ -Wundef.

--- Peter Dimov <pdimov@mmltd.net> wrote:
Please go ahead. The only potential minor problem with it might be g++ -Wundef.
The patch is in the CVS now. Thanks for the fast response! Cheers, Ralf __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com
participants (2)
-
Peter Dimov
-
Ralf W. Grosse-Kunstleve