July 2009 - Boost-users - lists.preview.boost.org

testing boost cross-compiled
by McLemon, Sean 17 Jul '09

17 Jul '09

Having configured the Boost library for our embedded target, built Boost and tested a few examples successfully we'd like to incorporate Boost into our testing framework. I've a couple of queries, though: 1. When we run the run.py script it checks out a fresh version of Boost from SVN, perhaps clobbering local changes or at least wasting a lot of time before testing. The workaround suggested in the website (http://beta.boost.org/development/running_regression_tests.html) - that is passing "--user=" to the script - doesn't change this. Is there any way to prevent this happening reliably? 2. Any executable we generate will need to be run using our command line simulator (i.e. we might create test.dxe, and run using jitccs532.exe test.dxe), how should I convey this to the test suite? 3. Our simulator does support input redirection via the command line (i.e. $ yes | jitccs532 test.dxe) or command line arguments - does this mean that some of tests won't be usable? - Sean

2 2

[Multi-Index] has_index<BMI,TAG> meta-function
by Dominique Devienne 17 Jul '09

17 Jul '09

I'd like to use enable_if/disable_if with a meta-function that determines whether a given multi-index container has a given tagged index. But I'm not having success writing has_index<BMI,TAG>, even looking at BMI's own index<Tag> meta-function :( Does anyone know how to implement it? Below how I'd like to use it. Thanks, --DD struct ByName {}; template <class BMI> struct VCursor { virtual void filter_by_name(const string& name) { do_filter_by_name(name); } private: void do_filter_by_name( const string& name, typename boost::enable_if<has_index<BMI, ByName> >::type* /*dummy*/ = 0 ) { typedef typename BMI::template index<ByName>::type index_type; ... } void do_filter_by_name( const string& name, typename boost::disable_if<has_index<BMI, ByName> >::type* /*dummy*/ = 0 ) { throw "no name index"; } };

3 6

[multi-index] Key extractor for 1:n relationship?
by Boris Schaeling 17 Jul '09

17 Jul '09

All the key extractors in the documentation of Boost.MultiIndex return one value for an element in a container. Is it somehow possible to use a 1:n relationship between element and value? The documentation uses this container definition (among others): typedef multi_index_container< employee, indexed_by< hashed_unique< const_mem_fun<employee, std::string, &employee::name> > > > employee_set; Let's assume an employee can have more than one name (std::string becomes std::vector<std::string>). Now if someone searches for Tom or John the same employee should be returned by the container. Is this possible? Boris

4 4

Using vectors of ptrs
by Robert Jones 17 Jul '09

17 Jul '09

Hi All Here's a quickie for you.... Given struct A { void a(); }; type std::vector<boost::shared_ptr<A> > VecType; VecType v; for (VecType::iterator i=v.begin(); i != v.end(); ++i ) (*i)->a(); How can I write that as a for_each loop? for_each( v.begin(), v.end(), boost::bind( ????? ) ); Thanks, - Rob. ps I can't use lambda because boost.bind is already used, and the placeholder name collisions just aren't worth the pain.

2 3

Re: [Boost-users] UTF16
by Tan, Tom (Shanghai) 17 Jul '09

17 Jul '09

I have some code that does conversion between UTF16 and MBCSs on Windows only: template<typename FROM, typename TO> struct convert { basic_string<TO> operator()(const basic_string<FROM>& from) { return from; } }; template<> struct convert<wchar_t, char> { string operator()(const wstring& from) { return utf16_to_mbcs(from); } private: string utf16_to_mbcs(const wstring& ws) { if(ws.empty()) return string(); const size_t BUFFER_SIZE = (ws.size() << 1) + 1; shared_array<char> p_mcb(new char[BUFFER_SIZE]); bool has_utf16le_bom = (0xFEFF == ws[0]); int count = ::WideCharToMultiByte( AreFileApisANSI() ? CP_THREAD_ACP : CP_OEMCP, WC_NO_BEST_FIT_CHARS, ( has_utf16le_bom ? ws.substr(1) : ws).c_str(), has_utf16le_bom ? ws.size() - 1 : ws.size(), p_mcb.get(), BUFFER_SIZE, 0, 0 ); return (0 == count) ? string() : string(p_mcb.get(), count ); } }; template<> struct convert<char, wchar_t> { wstring operator()(const string& from) { return mbcs_to_utf16(from); } private: wstring mbcs_to_utf16(const string& s) { if(s.empty()) return wstring(); const size_t BUFFER_SIZE = (s.size() << 1) + 1; shared_array<wchar_t> p_ws(new wchar_t[BUFFER_SIZE]); int count = ::MultiByteToWideChar( AreFileApisANSI() ? CP_THREAD_ACP : CP_OEMCP, MB_PRECOMPOSED, s.c_str(), s.size(), p_ws.get(), BUFFER_SIZE ); return (0 == count) ? wstring() : wstring(p_ws.get(), count ); } }; >Date: Thu, 16 Jul 2009 10:39:31 +0200 >From: plarroy <plarroy(a)promax.es> > >My approach is using std::string, etc. all the time and using UTF-8 >internally, only converting to other charsets when it's needed. >I use IBM icu library and made a boost::iostreams filter to convert >encoding, once it's done takes a lot of complexity away, I use it like: > // setup a conversion from charset to utf-8 > filt_streamb.push(ucnv_filter(charset.c_str(), "utf-8")); > istream is(&filt_streamb); >Perhaps there's interest to push this charset conversion into >boost::iostreams filters examples. >Regards. Robert Dailey wrote: > Oh, I also forgot to mention, I am also using boost::filesystem::path. I > guess this means I need to use wchar_t everywhere (std::wstring, > boost::filesystem::wpath, etc) and just let wxWidgets do the > encoding/decoding? If I don't have to do any encoding/decoding myself, then > there really is no need for a special object. But just in case I would like > to have the encoding/decoding abilities. > > On Sun, Jun 14, 2009 at 12:27 PM, Robert Dailey <rcdailey(a)gmail.com> wrote: > > >> Hi everyone, >> I did a bit of googling to see if Boost 1.39 as any portable support for >> UTF-16 encoded strings, but I did not find any. I'm currently using >> wxWidgets in my application, and I need a decent string object to use. I >> know that wxWidgets has UTF-16 string support through wxString, however I do >> not want to expose this object in my interfaces. I want to remain as >> abstracted away from wxWidgets as possible. Having said that, if someone >> could tell me if there is any existing UTF-16 string support in Boost, I'd >> appreciate it. I did not find anything in the vault, sandbox, or trunk in >> Boost. >> >> If boost has no such string object, could someone give me a head start on >> where to look? Thanks. >>

1 0

[Graph] Defining coordinates for vertices
by Maxime van Noppen 17 Jul '09

17 Jul '09

Hi, I can't figure out if it is possible (and if it is, how) to assign coordinates to graph vertices on planar graphs. I've tried setting the property vertex_index_t to std::pair<int, int> but it doesn't seem to work (make_maximal_planar doesn't use them for example). I'm trying to represent a 2D terrain via a triangular mesh. So far it is easy : I add all the points of the mesh as vertices and create the appropriated edges. But now I'm trying to make this graph evolve, which means adding new vertices (at some precise coordinates) and having them automagically linked to all other vertices possible while keeping the graph planar. This is what I do for now but that doesn't work (full source code is attached) : ------------------ typedef adjacency_list< vecS, vecS, undirectedS, property< vertex_index_t, std::pair<int, int> >, property<edge_index_t, int> > graph_t; typedef graph_traits<graph_t>::edge_descriptor edge_t; typedef graph_traits<graph_t>::vertex_descriptor vertex_t; // Create the graph graph_t g; // Add the three vertices of a triangle add_vertex(std::make_pair(0, 0), g); add_vertex(std::make_pair(10, 0), g); add_vertex(std::make_pair(5, 10), g); // Create the three edges add_edge(0, 1, g); add_edge(0, 2, g); add_edge(1, 2, g); // Add some random points add_vertex(std::make_pair(5, 5), g); add_vertex(std::make_pair(0, 5), g); // The goal is now to make g planar w.r.t. the coordinates // 1- Either make_connected or add at least one edge to make the graph // connected // 2- make_biconnected_planar // 3- make_maximal_planar ------------------ The input graph is (using boost::print_graph) : 0 <--> 1 2 1 <--> 0 2 2 <--> 0 1 3 <--> 4 <--> The output graph is : 0 <--> 1 2 3 4 1 <--> 0 2 3 2 <--> 0 1 3 4 3 <--> 2 4 0 1 4 <--> 3 2 0 It isn't palanar w.r.t. the given coordinates because 3 can't be connected with 4 without crossing one of the edges of the triangle (3 is supposed to be inside the triangle whereas 4 is outside). I suppose that either I'm doing something wrong or that it might not be possible to use the library this way. Any help appreciated, Thanks ! -- Maxime

2 4

[numeric_cast] Warning with MSVC
by Alexander Lamaison 17 Jul '09

17 Jul '09

I'm using boost::numeric_cast to convert the size of a wstring (presumably a size_t) to an int. const int size = boost::numeric_cast<int>(my_string.size()); MSVC still gives me a warning about possible loss of data: c:\program files\boost\boost_1_38\boost\numeric\conversion\ cast.hpp(41) : warning C4267: 'argument' : conversion from 'size_t' to 'unsigned int', possible loss of data c:\users\awl03\documents\visual studio 2005\projects\swish_trunk (padd)\swish\utils.hpp(91) : see reference to function template instantiation 'Target boost::numeric_cast<int,__w64 unsigned int>(Source)' being compiled with [ Target=int, Source=__w64 unsigned int ] c:\users\awl03\documents\visual studio 2005\projects\swish_trunk (padd)\swish\utils.hpp(121) : see reference to function template instantiation 'std::basic_string<_Elem,_Traits,_Ax> swish::utils::ConvertString<`anonymous-namespace'::Narrow>(const std::basic_string<wchar_t,std::char_traits<wchar_t>,std:: allocator<_Ty>> &)' being compiled with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char>, _Ty=wchar_t ] Unless I misunderstand, isn't the idea of numeric_cast that warnings aren't needed because the casts are checked for loss of data? Any advice is greatly appreaciated. I am using the MSVC 8.0 compiler with Boost 1.38.0 on 32-bit Windows Vista. Thanks. Alex Lamaison

3 3

BOOST_CLASS_TRACKING compilation error : tracking_level not a class or a namespace
by G, Benjamin (NSN - IN/Bangalore) 17 Jul '09

17 Jul '09

Hi, I want to disable tracking because of the now somewhat famous (STATIC_ASSERT compilation errors). I have a few classes within a namespace in a header file. For serialization of these, I want to disable tracking so I placed the macros: BOOST_CLASS_TRACKING(mynamespace::myClassA, boost:serialization::track_never); BOOST_CLASS_TRACKING(mynamespace::myClassB, boost:serialization::track_never); outside mynamespace. For this I am getting compilation errors saying : 'struct boost::serialiazation::tracking_level<mynamespace::myClassA>::type' is not a class or a namespace. I've included the boost header files, etc. in my header file. What am I missing? All help appreciated. - Benjamin Jacob.

2 2

[Serialization] Questions / Comments
by Rutger ter Borg 17 Jul '09

17 Jul '09

After having used the serialization library for quite a while, it seems that I'm approaching a point where I are seeing lots duplicate code which could be inside a serialization function. In turn, this head lead me to some questions / comments. The use of serialization in my case is both file and network oriented. This means reading and writing a great deal of short-lived archives. Works fine in between c++ programs, but outside that it's quickly becoming an infeasible task. Creating a custom archive could be an option, however, I would need to spend serious time on understanding the code due to the heavy use of macros. The question to be answered first is whether the current serialization architecture is able to fulfill my needs, or whether I should consider using other options. Considering the following two scenarios / questions: 1) Suppose I would like to create a custom XML archive which can both read shuffled XML data (its attributes/values in arbitrary order), and serialize to child attributes as well as named child nodes. Currently, name-value pairs are in the Serializable concept. I think they belong in the Archive Concept. Otherwise, it may be impossible to read XML- formatted data where the nodes are not in the same order as in the C++ code? In addition, suppose serializing a program structure is like a tree or graph, then name-value-pairs do not allow for any other nesting than the C++ code (i.e., it assumes everything is a child node). To allow for making distinction between nesting levels, one could think of, e.g., template< typename Archive > void serialize( Archive& ar, const unsigned int version ) { ar.serialize_node( "subtree", m_contained_class ); ar.serialize_attribute( "use_count", m_int ); } Is something like this possible? To this end, would it be possible to plug- in another parser, e.g., RapidXML? 2) I would like to create an archive that speaks the Action Message Format (AMF). Although I do have trouble finding a clear distinct parsing layer in the archives, I think this is possible to do. However, perhaps you think archive/serialization should not be used in conjunction with these kinds of uses. Many thanks in advance, Kind regards, Rutger ter Borg

2 3

Phoenix 2, 3 and related stuff
by joel 17 Jul '09

17 Jul '09

I'm currently fiddling with Phoenix 2 and i ended up needing look over my lambda statement to reorder/optimize them. AFAIK, Phoenix 2 is not proto-powered but phoenix 3 is not out yet. What should be the best strategy for me right now : waiting for Phoenix 3 or is there a way to hack inside Phoenix 2 so i can have a rough look at whixh type to match in my meta-function ? -- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35

3 6