- Boost-users - lists.boost.org

[tuple] easy way to bind to get
by Chris Weed 19 Oct '07

19 Oct '07

Hi, I would like to bind to the boost::get<N> function for use by transform iterator. This is the pseudo code if get worked like a simple function. std::vector<tuple<int,int,float> > v(2); v[0] = tuple<int,int,float>(2,2,2.2f); v[1] = tuple<int,int,float>(3,3,3.3f); std::accumulate( boost::make_transform_iterator( v.begin(), boost::lambda::bind(&boost::get<1>,_1)), boost::make_transform_iterator( v.end(), boost::lambda::bind(&boost::get<1>,_1)), 0); I find bind is difficult to use for functions that deduce template arguments from function arguments. I thought if I changed boost::get<1> -> boost::get<1,tuple<int,int,float> > That didn't work either. Digging through tuple.hpp makes my head spin. Is there an easy way to do this. Thanks, Chris

3 2

Warning binding member functions returning template types on MSVC80
by Josef Grahn 19 Oct '07

19 Oct '07

Hi, Visual Studio 2005 gives me warning C4180 ("qualifier applied to function type has no meaning; ignored") when binding member functions that has a templatized type as return type with boost::bind, but I can't see any reason why it should. The following example reproduces the warning: #include <boost/bind.hpp> template <typename T> struct Foo {}; struct Bar { Foo<void> f() { return Foo<void>(); } int g() { return 0; } }; Foo<void> f() { return Foo<void>(); } int main(int argc, const char **argv) { Bar b; boost::bind(&Bar::f, &b); // Warning C4180 (see below) boost::bind(&Bar::g, &b); // No warning boost::bind(f); // No warning return 0; } So far as I have tested, the program appears to function properly. The full warning is as follows (paths shortened): C:\...\boost/bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored C:\...\boost/bind.hpp(1609) : see reference to class template instantiation 'boost::_bi::add_cref<Pm,I>' being compiled with [ Pm=Foo<void> (__thiscall Bar::* )(void), I=1 ] ..\..\..\src\main.cpp(14) : see reference to class template instantiation 'boost::_bi::dm_result<Pm,A1>' being compiled with [ Pm=Foo<void> (__thiscall Bar::* )(void), A1=Bar * ] GCC 4.1.2 on Linux compiles the code without warnings. Any insights are greatly appreciated! Best regards, Josef Grahn

1 0

BGL: Bundled Properties with pointer type
by Benoît Dagon 19 Oct '07

19 Oct '07

Hi all, It's my first time submitting a question on a newsgroup so pardon me if I'm wrong in my way of doing. I have a graph of type BoostGraph_t defined like typedef boost::adjacency_list <vecS, vecS, undirectedS, NodeProperty*, EdgeProperty*> BoostGraph_t; with vertex and edge properties types defined as pointer on the following stuctures struct EdgeProperty { long m_id; double m_length; }; struct NodeProperty { long m_id; }; I'm now trying to use the dijkstra_shortest_paths() function with the edge length as parameter for the weigh_map function argument. But the call to the following function get(&EdgeProperty::m_length, m_graph); returns me the following errors ****** e:\Boost\boost\graph\properties.hpp(337) : error C2825: 'Bundle::*': cannot form a qualified name e:\Boost\boost\graph\properties.hpp(337) : error C2226: syntax error : unexpected type 'Graph' e:\Boost\boost\graph\properties.hpp(337) : error C2334: unexpected token(s) preceding ':'; skipping apparent function body e:\Boost\boost\graph\properties.hpp(342) : error C2825: 'Bundle::*': cannot form a qualified name e:\Boost\boost\graph\properties.hpp(342) : error C2657: '`global namespace'::*' found at the start of a statement (did you forget to specify a type?) e:\Boost\boost\graph\properties.hpp(342) : error C2653: 'operator``global namespace''' : is not a class or namespace name e:\Boost\boost\graph\properties.hpp(342) : error C2501: 'boost::bundle_property_map<Graph,Descriptor,Bundle,T>::pm' : missing storage-class or type specifiers with [ Graph=boost::adjacency_list<boost::vecS,boost::vecS, boost::undirectedS,NodeProperty *,EdgeProperty *>, Descriptor=boost::property_map<boost::adjacency_list<boost::vecS, boost::vecS,boost::undirectedS,NodeProperty *,EdgeProperty *>, double EdgeProperty::* >::descriptor, Bundle=boost::property_map<boost::adjacency_list<boost::vecS, boost::vecS,boost::undirectedS,NodeProperty *,EdgeProperty *>, double EdgeProperty::* >::actual_bundle, T=double ] ****** I know it's working correctly if I use NodeProperty and EdgeProperty for graph template arguments instead of pointers. But I would like to know if it's possible to do it in that way and how would be the correct syntax for the get() to work properly. Many thanks in advance for your help! ~Benoît

2 2

Re: [Boost-users] [fusion] subset view
by dan marsden 19 Oct '07

19 Oct '07

>I would like to make a subset view of the element in a tuple. It >would, for example, give me a view into the first, third, and fifth >element of the tuple. It isn't clear how to do this with the available >views. You could do the following (untested): * Use zip to zip an mpl range of numeric indexes to your sequence * Use filter to extract the positions you want, based on the indices zipped on above * Use transform to remove the position labels so the sequence elements are just your desired data. The above works only if you need const access, as the algorithms return views to const data. If you need non-const data, you can use the corresponding views. The above is somewhat awkward, if there is some criteria other than position index that can be used to extract your elements, you may be able to do a lot better. Or you could choose different data structures to start off with, for example if you really have 2 subsets of elements, sub0 and sub1, and you know they're separate, store them in 2 fusion containers, and when you need to work on the whole, use fusion::join to glue them together. HTH Cheers Dan ___________________________________________________________ Yahoo! Answers - Got a question? Someone out there knows the answer. Try it now. http://uk.answers.yahoo.com/

2 1

sorting an ublas matrix by a certain column
by Alexander Schmolck 19 Oct '07

19 Oct '07

Hi, If I have a boost::numeric::ublas::vector<double> v, I can use the STL's sort and boost::lambda. sort(v.begin(), v.end(), _1<_2); How do I do the same with matrix<double>, using, say column 3 as sort key (i.e. I want to reorder the rows in the matrix by their 3rd entry)? sort(m.begin1(), m.end1(), _1[3]<_2[3]); won't do the trick, and I really have a hard time figuring out what does. In case it's not clear what I want, here's an example: [0,0,1,4 0,0,2,3 0,0,3,2 0,0,4,1] => [0,0,4,1 0,0,3,2 0,0,2,3 0,0,1,4] cheers, 'as

1 0

How well does Boost support cross-compilation?
by n.torrey.pines＠gmail.com 19 Oct '07

19 Oct '07

I'm dreaming of compiling MinGW, OS X (Intel and PowerPC) and Linux x86 versions of my program in one go, perhaps on Linux. The program uses little other than several Boost libraries. Is this a possible and reliable usage scenario? This doesn't seem like a particularly original idea, so I'm sure others explored it. Are there any how-to's? I couldn't find any.

3 5

Boost Regex use of &*string.begin()
by Sudarshan Gaikaiwari 19 Oct '07

19 Oct '07

Hi The boost regex library makes liberal use of &*string.begin() For example ---------------------------------------------------------------------------------------- std::string result(++bytes, '\0'); bytes = ::LCMapStringW( id, // locale identifier LCMAP_SORTKEY, // mapping transformation type p1, // source string static_cast<int>(p2 - p1), // number of characters in source string reinterpret_cast<wchar_t*>(&*result.begin()), // destination buffer *of bytes* bytes // size of destination buffer ); -------------------------------------------------------------------------------------------------------------------- Is this construct safe? Are its semantics specified by the standard? regards Sudarshan -- Sudarshan Gaikaiwari www.sudarshan.org sudarshan(a)acm.org

2 1

[xpressive] crash again :(
by Ovanes Markarian 19 Oct '07

19 Oct '07

Well, sorry I did not check mails before. Actually, I should have thought about it, that token_iterator applies a transformation on each iteration. Many thanks. It was already late in the night, when I faced this. Sorry for disturbing your sister's wedding. Hope you will enjoy it! Best wishes to your sister and her husband! With Kind Regards, Ovanes On 10/18/07, Eric Niebler <eric(a)boost-consulting.com> wrote: > > Eric Niebler wrote: > > Eric Niebler wrote: > >> Ovanes Markarian wrote: > >>> Hello Eric, > >>> > >>> sorry for bothering. I use xpressive in the following context: > >> <snip> > >> > >> I can't get your code to compile. Please send a minimal and complete > >> source file that demonstrates the problem. > > > > After fiddling with your code a bit, I got it to compile, but I'm not > > seeing any assertions, exceptions or crashes. What version are you > using? > > Me again. I realized I wasn't seeing the problem because I hadn't put > any strings in the vector over which you were iterating. When I do, I > can reproduce the assert. This is not a bug in xpressive, it is a bug in > your code. The problem is here: > > std::tr1::bind(&apply_regex, sem_separated, td::tr1::placeholders::_1) > > In this call to bind, you pass sem_separated to bind() *by value*. You > return a regex_token_iterator, which holds a reference to the > (temporary) regex object. When the temporary regex object goes away, the > regex_token_iterators are left holding dangling references. > > The lifetime of a regex_token_iterator cannot exceed the lifetime of the > regex it holds. > > If you change the code to ... > > std::tr1::bind(&apply_regex, std::tr1::ref(sem_separated), > td::tr1::placeholders::_1) > > ... it works just fine. > > HTH, > > -- > Eric Niebler > Boost Consulting > www.boost-consulting.com > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

1 0

Feature request for serialization
by Robert Bielik 19 Oct '07

19 Oct '07

It would be nice to be able to specify a callback that can notify how far a serialization (at loading) has come. I have a structure that can take a LONG time to load, and it would be good to be able to present a progress bar for the user. /R

4 5

visual studio 2005/2008: example runs in debug-mode, does not run in release mode
by Vincent Vega 19 Oct '07

19 Oct '07

Hi there! - used the BoostPro Binary Installer for Visual C++ to obtain boost 1.34.1 - downloaded the following example for maximum_cardinality_matching: http://www.boost.org/libs/graph/example/matching_example.cpp - created an project using VS2005 as well as VS2008 Got the following problem (using both studios): - everythings fine while using the debug-mode - project runs but produces wrong results when in release-mode Anybody has an idea what causes my problem? Thanks in advance Dirk -- Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten Browser-Versionen downloaden: http://www.gmx.net/de/go/browser

2 1