RE: [boost] Problems compiling MPL

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2002-07-29 21:57:06

Jaap Suter wrote:
> I'm sorry if this actually belongs in the Boost user group 
> instead of here, 

Not yet (if it would ever), - until the library is under review, this list
is exactly the right place to post questions to.

> but I think the MPL is currently only 
> being discussed in here.

Yep.

> 
> Anyway, I'm having a few problems using the MPL. Here's what 
> I have done.

[...]

> 4. Then I wrote the following code:
> 
> typedef boost::mpl::vector< float, double > vars;
> typedef boost::mpl::push_back< vars, unsigned char >::type ext_vars;
> typedef boost::mpl::at_c< 2, ext_vars >::type third_var;
> BOOST_STATIC_ASSERT( (boost::is_same< double, third_var >::value) );
> 
> 5. I compiled everything with Microsoft Visual Studio 7 
> (.NET) , and got the
> following errors:
> 
> c:\Development\Library\boost_1_28_0\boost\mpl\push_back.hpp(36) :
> 
> error C2504: 
> 'boost::mpl::push_back_traits<Tag>::algorithm<Sequence,T>' :
> 
> base class undefined

[...]

> If anybody knows what I'm doing wrong, that would be great. 
> AFAIK the MPL compile-matrix shows that VC-7 should 
> compile my code.

Oh, sorry, I should have updated 'push_back' test (make it empty for now :).
The error is not VC 7.0 specific; it's the result of the interface change,
which, in its turn, is the result of a review comment. The change is that in
order to allow uniform insertion into any Extensible Sequence the family of
'vector' templates was gifted a 'push_front' operation - for the expense of
'push_back'. The "for the expense" part of the change is temporary - it's
perfectly possible to implement a vector that supports both 'push_front' and
'push_back',   I just didn't have resources for that. 

To summarise the above 'push_back' will be back soon :), but in the meantime
please ignore it.

> 
> Sincerely,
> 
> Jaap "hoping he didn't overlook a really obvious mistake" Suter

You didn't, sorry for a faulty test.

Aleksey