|
Boost : |
From: John Maddock (John_Maddock_at_[hidden])
Date: 2002-03-01 07:29:16
>Intel C++ uncovered the following with a warning:
------------------------
typedef int& r_type;
typedef const r_type cr_type;
------------------------
type_traits_test.hpp(278): warning #21: type qualifiers are meaningless in
this declaration
typedef const r_type cr_type;
^
------------------------
I wonder if we've been testing the right things?
I changed it to:
typedef int const& cr_type;
Everything seems to continue to pass.
On the other hand, the pragmas used to disable warnings with MSVC suggest
that this was intentional. Was it? A comment would be a big help, here.
<
Yikes that's the wrong change, it was intentional, but as you say should be
documented - it uncovers bugs in some compilers (gcc and Borland at least).
I'll change it back and comment it.
- John Maddock
http://ourworld.compuserve.com/homepages/john_maddock/
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk