- Boost-users - lists.preview.boost.org

The unholy triumvirate of Boost, MacOS X & gcc
by Paul-Michael Agapow 23 Apr '03

23 Apr '03

Does any one have recent experience with compiling / installing Boost on a Mac using gcc? A brief vox-poll seems to say that most everyone on the Mac uses it with CodeWarrior, manually including it in their project. However I'm more interested in using gcc / ProjectBuilder, and distributing the source code to my programs. So, I need to know if my end-users are going to have any problems installing Boost? My attempts have been been discouraging. The first install of Spirit almost immediately consumed all the memory on my machine and thrashed. Installation of the full Boost failed with link errors reporting "unrecognised option shared". Editing this option out of the regex makefile, lead to the next problem, failures with the unittests. Suspecting I was missing something obvious, I handed the problem over to someone else who spend a few days following the same trail, eventually commenting out the unittests. Sure, I could include the boost source with every program I use it in, but this strikes me as untidy and means every time Boost updates, I have to update the x separate copies on my disk. Opinions or sympathies? p -- Dr Paul-Michael Agapow (p.agapow(a)ucl.ac.uk) Dept. Biology, University College London

3 2

Using for_each on a std::list<boost::function0<void> >
by Russell Hind 23 Apr '03

23 Apr '03

I have a std::list of boost::function0<void> I would like to call all items using a for_each statement. Is there a standard function object that I can pass into the for_each call that will call operator() on each element? Thanks Russell

2 2

dijkstra_shortest_paths with undirectedS example
by jfw1000 23 Apr '03

23 Apr '03

Hello, I am new to BGL, I was trying to test dijkstra_shortest_paths() with undirectedS, the result is not what I wanted, anyone giving help? Thanks, source code like dijkstra-example.cpp in the graph/example. int main(int, char *[]) { typedef adjacency_list < listS, vecS, undirectedS, no_property, property < edge_weight_t, double > > graph_t; typedef graph_traits < graph_t >::vertex_descriptor vertex_descriptor; typedef graph_traits < graph_t >::edge_descriptor edge_descriptor; typedef std::pair<int, int> Edge; const int num_nodes = 3; Edge edge_array[] = { Edge(0, 1),Edge(1, 2), Edge(2, 0) }; double weights[] = {2.0, 1.0, 1.0}; int num_arcs = sizeof(edge_array) / sizeof(Edge);// #if defined(BOOST_MSVC) && BOOST_MSVC <= 1300 graph_t g(num_nodes); property_map<graph_t, edge_weight_t>::type weightmap = get (edge_weight, g); for (std::size_t j = 0; j < num_arcs; ++j) { edge_descriptor e; bool inserted; tie(e, inserted) = add_edge(edge_array[j].first, edge_array [j].second, g); weightmap[e] = weights[j]; } #else graph_t g(edge_array, edge_array + num_arcs, weights, num_nodes); property_map<graph_t, edge_weight_t>::type weightmap = get (edge_weight, g); #endif std::vector<vertex_descriptor> pred(num_vertices(g)); std::vector<double> d(num_vertices(g)); vertex_descriptor s = vertex(0, g); #if defined(BOOST_MSVC) && BOOST_MSVC <= 1300 // VC++ has trouble with the named parameters mechanism property_map<graph_t, vertex_index_t>::type indexmap = get (vertex_index, g); dijkstra_shortest_paths(g, s, &pred[0], &d[0], weightmap, indexmap, std::less<double>(), closed_plus<double>(), std::numeric_limits<double>::max(), 0, default_dijkstra_visitor()); #else dijkstra_shortest_paths(g, s, predecessor_map(&p[0]).distance_map(&d [0])); #endif std::cout << "distances and parents:" << std::endl; graph_traits < graph_t >::vertex_iterator vi, vend; for (tie(vi, vend) = vertices(g); vi != vend; ++vi) { std::cout << "distance(" << *vi << ") = " << d[*vi] << ", "; std::cout << "parent(" << *vi << ") = " << pred[*vi] << std::endl; } std::cout << "min paths tree" << std::endl; adjacency_list<> tree(num_nodes); for(tie(vi,vend) = vertices(g); vi != vend; ++vi) if (*vi != pred[*vi]) add_edge(pred[*vi], *vi, tree); print_graph(tree); std::cout << std::endl; return EXIT_SUCCESS; } The result is : distances and parents: distance(0) = 0, parent(0) = 0 distance(1) = 2, parent(1) = 0 distance(0) = 1, parent(2) = 0 min paths tree 0 -> 1 2 1 -> 2 -> I think the correct one should be: distances and parents: distance(0) = 0, parent(0) = 0 distance(1) = 2, parent(1) = 2 distance(2) = 1, parent(2) = 0 min paths tree 0 -> 1 1 -> 2 2 -> Any idea what is wrong? thanks!

3 3

RE: [Boost-Users] simplest pool_alloc example does not work on VC 6
by scleary＠jerviswebb.com 23 Apr '03

23 Apr '03

> -----Original Message----- > From: Ben Hutchings [mailto:ben.hutchings@businesswebsoftware.com] > > scleary(a)jerviswebb.com wrote: > > Sorry, Boost.Pool does not currently support MSVC, due to that > compiler's > > extreme lack of template support. > <snip> > > For VC++ 6, all that should be needed is a _Charalloc function (instead > of struct rebind) and a change to the type of deallocate's first > parameter (to void *, which actually makes more sense). The _Charalloc function requires each allocator to be able to allocate an arbitrary amount of memory. This goes against the entire design of Boost.Pool, and would be non-trivial (to say the least) to implement. > The VC++ 7 compiler is good enough to support standard allocators and > the library does use them. With the addition of a variable name and > semi-colon, the given example can be compiled with this version. Glad to hear it. Too bad it's not released yet. -Steve

2 1

RE: [Boost-Users] Boost::Filesystem: Directory iterator and dangling links
by Ben Hutchings 23 Apr '03

23 Apr '03

You wrote: <snip> > In the case of dangling links I obtain a not_found_error. > This seems to be a contradiction. First I find something (and obtain an > iterator) and without any interference from > outside is_directory() tells me it cannot find the object. > The problem is that directory_iterator refers to the link whereas > is_directory() and the following exception refer to the target (which is > missing). <snip> To make this consistent, then, maybe is_directory() and friends should be using lstat rather than stat on Unix. Is it desirable that links be resolved? I suppose it is if the iterator is being used to walk a whole tree - but not if it's being used to list a single directory. Perhaps there is a need to include explicit support for links in the API. Windows has supported them since 2000, so they aren't just a Unix thing now.

2 1

RE: [Boost-Users] simplest pool_alloc example does not work on VC 6
by Beman Dawes 22 Apr '03

22 Apr '03

At 01:40 PM 4/22/2003, scleary(a)jerviswebb.com wrote: >> -----Original Message----- >> From: Ben Hutchings [mailto:ben.hutchings@businesswebsoftware.com] >>... >> The VC++ 7 compiler is good enough to support standard allocators and >> the library does use them. With the addition of a variable name and >> semi-colon, the given example can be compiled with this version. > >Glad to hear it. Too bad it's not released yet. Ben is probably talking about 7.0, which has been released for a year or so. VC++ 7.1 will be released officially Thursday, the 24th, IIRC. Using the VC++ 7.1 final beta, test_pool_alloc is passing OK. --Beman

1 0

RE: [Boost-Users] simplest pool_alloc example does not work on VC6
by Ben Hutchings 22 Apr '03

22 Apr '03

scleary(a)jerviswebb.com wrote: > Ben Hutchings <ben.hutchings(a)businesswebsoftware.com> wrote: > > scleary(a)jerviswebb.com wrote: > > > Sorry, Boost.Pool does not currently support MSVC, due to that compiler's > > > extreme lack of template support. <snip> > > The VC++ 7 compiler is good enough to support standard allocators and > > the library does use them. With the addition of a variable name and > > semi-colon, the given example can be compiled with this version. > > Glad to hear it. Too bad it's not released yet. Where have you been for the last year?

1 0

Newbie boost help needed..
by Bryan Coon 22 Apr '03

22 Apr '03

Hi, I am totally new to boost and just wanted to ask a couple of simple questions... First, I compiled the boost libraries okay with msvc. I updated the include path in my vc++ 6.0, and can include boost files via #include <boost/xxx.hpp>. So far so good! I wanted to try some things out, like boost's mersenne twister random number generator. I first did this: #include <boost/random/mersenn_twister.hpp> And got no errors. Then I read there was a random.hpp for convenience so I changed this to: #include <boost/random.hpp> And received as a reward complaints: d:\dev\devproject\boost_1_30_0\boost\random\uniform_01.hpp(44) : error C2275: 'std::numeric_limits<`template-parameter258'>::is_integer' : illegal use of this type as an expression d:\dev\devproject\boost_1_30_0\boost\random\uniform_01.hpp(95) : see reference to class template instantiation 'boost::uniform_01<UniformRandomNumberGenerator,RealType>' being compiled There were 6 of these errors. What am I doing wrong? Also, is there a way to include the boost libs in vc++ where I get the autocompletion if I use boost:: ? It is very difficult to figure out how the libs are supposed to be used (for me) by looking at example code or the documentation. Autocompletion as if boost were part of my project would be very useful. Thanks, B

1 0

RE: [Boost-Users] simplest pool_alloc example does not work on VC6
by Ben Hutchings 22 Apr '03

22 Apr '03

scleary(a)jerviswebb.com wrote: > Sorry, Boost.Pool does not currently support MSVC, due to that compiler's > extreme lack of template support. <snip> For VC++ 6, all that should be needed is a _Charalloc function (instead of struct rebind) and a change to the type of deallocate's first parameter (to void *, which actually makes more sense). The VC++ 7 compiler is good enough to support standard allocators and the library does use them. With the addition of a variable name and semi-colon, the given example can be compiled with this version. Is there some other reason why VC++ cannot be supported?

1 0

simplest pool_alloc example does not work on VC6
by John Harris 22 Apr '03

22 Apr '03

Using Boost 1.30.0. MSVC6 Tow problems: (1) I have to #define __WIN32__, and (2) Doesn't compile Am I doing something wrong, here? #include "afxwin.h" #include <map> #define __WIN32__ // why do I have to define this??? #include "boost/pool/pool_alloc.hpp" int main() { std::map<int,int,std::less<int>, boost::pool_allocator<int> > return 0; } :\Program Files\Microsoft Visual Studio\VC98\INCLUDE\xtree(587) : error C2664: 'deallocate' : cannot convert parameter 1 from 'struct std::_Tree<int,struct std::pair<int const ,int>,struct std::map<int,int,struct std::less<int>,class boost::pool_al locator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win32_mutex,32> >::_Kfn,struct std::less<int>,class boost::pool_allocator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win32_m utex,32> >::_Node *' to 'int *const ' Types pointed to are unrelated; conversion requires reinterpret_cast, C-style cast or function-style cast C:\Program Files\Microsoft Visual Studio\VC98\INCLUDE\xtree (587) : while compiling class-template member function 'void __thiscall std::_Tree<int,struct std::pair<int const ,int>,struct std::map<int,int,struct std::less<int>,class boost::poo l_allocator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win32_mutex,32> >::_Kfn,struct std::less<int>,class boost::pool_allocator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win 32_mutex,32> >::_Freenode(struct std::_Tree<int,struct std::pair<int const ,int>,struct std::map<int,int,struct std::less<int>,class boost::pool_allocator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win32_mutex,32 > >::_Kfn,struct std::less<int>,class boost::pool_allocator<int,struct boost::default_user_allocator_new_delete,class boost::details::pool::win32_mutex,32> >::_Node *)'

1 0