Re: [boost] [config] borland/codegear problems with C++0x macros

Subject: Re: [boost] [config] borland/codegear problems with C++0x macros
From: Beman Dawes (bdawes_at_[hidden])
Date: 2009-03-01 13:39:36

On Sun, Mar 1, 2009 at 12:12 PM, John Maddock <john_at_[hidden]> wrote:
>> * trunk/boost/config/compiler/borland.hpp: It looks like John Maddock
>> and I independently added C++0x macros. The resulting inconsistencies
>> need to be reconciled. The BOOST_CODEGEAR_0X_SUPPORT macro doesn't
>> appear to be documented, by the way.
>>
>> John, do you want to take a whack at this?
>
> Well I've had a go...

Thanks!

> the BOOST_CODEGEAR_0X_SUPPORT macro is entirely
> internal to that header BTW, it gets turned on when the -Ax compiler option
> is used.

Ah! #  pragma defineonoption BOOST_CODEGEAR_0X_SUPPORT -Ax

That's a new one to me. Live and learn.

>> * trunk/boost/config/compiler/codegear.hpp: Doesn't have the negative
>> C++ macros yet. I'd like to borland.hpp finished before trying to do
>> codegear.hpp.
>
> It's basically the same as Borland.hpp, so I took a whack at that one as
> well, we'll see how the tests go.

Great!

> BTW BOOST_HAS_EXPLICIT_CONVERSION_OPS and BOOST_HAS_SCOPED_ENUM appear to
> have no new alternatives?

They should be there:

#  define BOOST_NO_EXPLICIT_CONVERSION_OPERATORS
...
#  define BOOST_NO_SCOPED_ENUMS

And they did get documented, too:-)

I do need to check the gcc 4.4.x compiler series to make sure we are
covering the features they added after the 4.3 series. And I expect a
stream of C++0x features from all of the compiler vendors over the
next year or so. With both GCC and Microsoft shipping 0x features, the
others won't dare to be left behind:-)

--Beman