From: John Maddock (John_Maddock_at_[hidden])
Date: 2001-08-24 06:34:36
>Leaving aside whether the above is legal (I don't see is_const<>::t
>anywhere, this modification fixes the problem:
On the issue of legality:
1) the static member is used in a way that does not require the definition
to be present.
2) providing a definition would cause the code to fail to compile in many
cases (what kind of constructor do you use?)
3) its a hack for a broken compiler in any case...
On your fix:
Sorry, but you just can't do that - your proposed fix would cause is_const
to fail to compile with reference types, which would stop is_reference from
working, which would stop compressed_pair, call_traits, functional.hpp and
heaven knows what else to stop working. (try building the type traits test
programs if you don't believe me).
Finally, I notice that another_tuple_test_bench.cpp seems to contain
spurious carriage returns - it's been checked in as a text file so I don't
know what the issue is - but it needs fixing or unix users are going to
start complaining - as does my win32 text editor.
- John Maddock
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk