On Tue, Feb 10, 2009 at 8:10 PM, Joel de Guzman wrote:

Stjepan Rajko wrote:

...

Hi,

I tried to make boost::array a conforming MPL sequence by including:

#include // makes array a conforming fusion sequence #include // makes fusion sequences conforming MPL sequences

but,

typedef boost::array array; assert(boost::mpl::is_sequence<array>::type::value); // fails

Am I misinterpreting the fusion documentation in expecting that this should work? Is there another way of making boost::array an mpl sequence?

I think the basic problem is that is_sequence is not really part of the requirements for an MPL sequence. You don't have to be a conforming MPL sequences. I am not sure how to best approach this. Perhaps the MPL authors have a better idea?

Joel, thanks for your response. My original problem was with a simple application of mpl::transform, which produced an error stating: no class template named 'apply' in 'struct boost::mpl::clear_implboost::mpl::non_sequence_tag The above code was just an attempt at a minimal example. Also, the same above code passes if boost::array is replaced with fusion::vector<int>, so my impression is that there is something happening for fusion-proper sequences that is not happening for the adapted array. I can provide a minimal example that involves mpl::transform if that is helpful, but I imagine (perhaps erroneously) that whatever fixes mpl::is_sequence would also fix my problem with mpl::transform. Stjepan

On Thu, Feb 12, 2009 at 10:12 AM, Stjepan Rajko wrote:

I was able to get the fusion-adapted boost::array to be a conforming mpl sequence by adding:

namespace boost { namespace mpl {

template struct tag > { typedef boost::fusion::fusion_sequence_tag type; };

template struct sequence_tag > { typedef boost::fusion::fusion_sequence_tag type; };

}}

Furthermore, I am able to execute a transform on the array after adding this: http://pastebin.com/f40b33e1d This works perfectly, that is, minus the fact that adding it in gives me some really bizarre compilation errors (gcc 4.0.1, darwin): /Development/boost/boost/fusion/adapted/array/array_iterator.hpp: In instantiation of 'boost::fusion::array_iterator, 0>': ... /Development/boost/boost/fusion/adapted/array/array_iterator.hpp:29: error: 'boost::array::<anonymous enum>' is/uses anonymous type /Development/boost/boost/fusion/adapted/array/array_iterator.hpp:29: error: trying to instantiate 'template struct boost::fusion::detail::enable_comparison' The offending line in array_iterator is: BOOST_MPL_ASSERT_RELATION(Pos, <=, Array::static_size); Commenting out that line results in a successful compile, and the program runs as expected. I would really appreciate it if someone could help me out here, I am in total dark as to why the BOOST_MPL_ASSERT_RELATION is not working correctly, and what fusion::detail::enable_comparison has to do with anything. TIA, Stjepan

On Thu, Feb 12, 2009 at 7:52 PM, Steven Watanabe wrote:

AMDG

Stjepan Rajko wrote:

...

/Development/boost/boost/fusion/adapted/array/array_iterator.hpp:29: error: 'boost::array::<anonymous enum>' is/uses anonymous type /Development/boost/boost/fusion/adapted/array/array_iterator.hpp:29: error: trying to instantiate 'template struct boost::fusion::detail::enable_comparison'

The offending line in array_iterator is: BOOST_MPL_ASSERT_RELATION(Pos, <=, Array::static_size);

Commenting out that line results in a successful compile, and the program runs as expected.

I would really appreciate it if someone could help me out here, I am in total dark as to why the BOOST_MPL_ASSERT_RELATION is not working correctly, and what fusion::detail::enable_comparison has to do with anything.

Wild guess: try casting Array::static_size to int?

That worked - thanks! Created ticket https://svn.boost.org/trac/boost/ticket/2761 Stjepan