July 2007 - Boost-users - lists.boost.org

Using lambda bind as destruct functor to shared_ptr
by Scott French 04 Jul '07

04 Jul '07

Hello, Should the following code sample work? It is the simplest reproduction case I could come up with: #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/shared_ptr.hpp> #include <stdlib.h> using namespace std; using namespace boost; using namespace boost::lambda; int main(void) { void * myResource = malloc(5); shared_ptr<void> mySharedResource(myResource, bind(free, _1)); } I get a compile error using g++ 4.0.1 on Mac OS X: /Documents/boost_1_34_0/boost/detail/sp_counted_impl.hpp: In member function 'void* boost::detail::sp_counted_impl_pd<P, D>::get_deleter(const std::type_info&) [with P = void*, D = boost::lambda::lambda_functor<boost::lambda:: lambda_functor_base<boost::lambda:: action<2, boost::lambda::function_action<2, void> >, boost::tuples::tuple<void (&)(void*), const boost::lambda::lambda_functor<boost::lambda::placeholder<1> >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >]': test.cpp:13: instantiated from here /Documents/boost_1_34_0/boost/detail/sp_counted_impl.hpp:150: error: no match for ternary 'operator?:' in '(+ ti)->std::type_info::operator==(((const std::type_info&)((const std::type_info*)(& _ZTIN5boost6lambda14lambda_functorINS0_19lambda_functor_baseINS0_6action ILi2ENS0 _15function_actionILi2EvEEEENS_6tuples5tupleIRFvPvEKNS1_INS0_11placehold erILi1EE EEENS7_9null_typeESG_SG_SG_SG_SG_SG_SG_EEEEEE)))) ? boost::lambda::operator&(const boost::lambda::lambda_functor<T>&) [with Arg = boost::lambda::lambda_functor_base<boost::lambda::action<2, boost::lambda::function_action<2, void> >, boost::tuples::tuple<void (&)(void*), const boost::lambda::lambda_functor<boost::lambda::placeholder<1> >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> >]() : 0' Of course, just using "free" instead of bind(free, _1) works, but my real case is more complicated and this appears to be a general conflict between lambda::bind and shared_ptr. Thanks, Scott French

2 1

[IBD] Weekly news: 2007-06-27 to 2007-07-03
by Matias Capeletto 04 Jul '07

04 Jul '07

================================================== Hi Boost, Once a week, we will send you a brief of our work for the Improving Boost Docs project. We need more hands, join us! We are waiting you on the boost-docs list. More information: http://svn.boost.org/trac/boost/wiki/ImprovingBoostDocs ================================================== 2007-07-03 ------------------------------------------------------------------------------------- * A new community builder tool for the project: Linked In. Be part of our network of trust. * Change date format to be ISO 8601 compliant. * Ohloh statistics for our project. 2007-07-02 ------------------------------------------------------------------------------- * New project: Boost Debugger Visualizers. Welcome Filip! * Formal presentation of the project. 32 downloads on the first 12 hours! * Initial structure of the Boost Glue docs project. 2007-07-01 ------------------------------------------------------------------------------- * First Release of Boost Docs! 2007-06-30 ------------------------------------------------------------------------------- * We will use ohloh as a community builder tool and for statistics about our project. * Two new objectives: Standard C++ Library Boosted Docs and Documentation Best Practices. 2007-06-29 ------------------------------------------------------------------------------- * Stuart Dootson will be our browser testing manager. Thanks! * Add fancy graphics to the wiki. 2007-06-28 ------------------------------------------------------------------------------- * The formal objectives of the project have been outlined. * We are trying to build a long term community of people that cares and constantly improve Boost docs. With that in mind new sections have been created: Subprojects, People, and Acknowledgments; including pictures of us to make us feel as a group. * The community is growing: Peter Foley starts the SVG icons set project. Welcome Peter! 2007-06-27 ------------------------------------------------------------------------------- * Help needed for a new project: Boost specific WikiMacros * "Quickbook as a WikiProcessor for our Trac", let us know we can count on you! Best regards Improving Boost Docs Crowd

2 3

Iteratiing on regex results with pointer-to-member-function
by Alain Leblanc 04 Jul '07

04 Jul '07

Hi, I wrote some code where I want to iterate on the result of a regex matching and call a member function for each match. It goes roughly like this: ====================================================================== class XX { typedef bool (XX::*FuncPtr)(const boost::match_results<std::string::const_iterator>& what); FuncPtr funcPtr; }; bool XX::regex_callback(const boost::match_results<std::string::const_iterator>& what) { return true; } void XX::addDelayedVariables(string expr) { funcPtr = &XX::regex_callback; regex re("([[:alpha::]][[:alnum]]+)(\[[^\\]]+])"); sregex_iterator m1(expr.begin(), expr.end(), re); sregex_iterator m2; match_results<string::const_iterator> results; (*) bool a = (this->*funcPtr)(results); (*) for_each(m1, m2, (this->*funcPtr)); (#) } =========================================================================== When compiling (gcc 4.1 on FC 6) I get an error message at (#) that says: invalid use of non-static member function I added the two lines with a (*) as a sanity check to see that I declared the pointer-to-member-function correctly, and I have no problem compiling those. So my question is: is this a problem with me not understanding how to handle pointer-to-member-functions in C++, or is there something with the boost libraries that I'm missing or is preventing me from performing this operation? I would appreciate any help or opinion. thanks, Alain

2 2

Can't compile boost 1.3.3.1 with stlport 4.6.2
by Hongquan 04 Jul '07

04 Jul '07

Hi Buddy: I down the boost 1.3.3.1 and stlport 4.6.2 and then using the command < bjam.exe -sTOOLS=msvc-stlport -sSTLPORT_4.6.2_PATH=D:\SDK\Stlport\STLport-4.6.2 > to build the boost source ,but some error was threw out in compile process just as: E:\boost_1_33_1\boost/date_time/date_formatting.hpp(64) : fatal error C1001: INT ERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 1794) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information I searched the web and find out some clues at microsoft web site http://support.microsoft.com/kb/226110/en-us and Microsoft has comfirmed that this is a bug of vc6 . But I'm thinking if there is any other way to build the boost with stlport . If anyone knows it,please give me a tip! Thank you very much! Hongquan I-ZQ

3 2

Simple regex question
by david v 04 Jul '07

04 Jul '07

Hello, I'm trying to use regex_search to look for regular expressions. regex = (david|divad) stringtosearch = "david" So I'm looking for regex above in the string "david". As you see a very basic example. I'm expecting to find two hits, but I thought I should recover "david" and "david" and not twice the same match "david" ??? Here is a snippet of the code: .... boost::smatch match; start = stringtosearch.begin(); //that would be the start of "david" end = stringtosearch.end(); // that would be the end of "david" old_pos = 0; cout << "Looking for regex:" << regex << " in string " << stringtosearch << endl; while (boost::regex_search(start,end,match,regex)) { old_pos += match.position()+1; std::string t = match[0]; std::cout << "\tFound:" << match[0] << endl; old_pos += match.length(); start = match[0].second; } catch(const std::exception &regex) { std::cerr << "Exception:" << regex.what() << std::endl; exit(-1); } Here is the output: Looking for regex:(david|divad) in string david Found david Looking for regex:(david|divad) in string david Found david As you can see, pattern "david" is found twice ??? I was expecting two different patterns to be found "david" and "divad". By the way how can I recover the total number of matches found ?? thanks, david _________________________________________________________________ MSN Messenger : discutez en direct avec vos amis ! http://www.msn.fr/msger/default.asp

2 1

[signal] How to disconnect?
by Boris 04 Jul '07

04 Jul '07

void Hello() { } boost::signal<void ()> sig; sig.connect(&Hello); // OK sig.disconnect(&Hello); // Error Can anyone tell me how I disconnect the callback function Hello? I can see in the documentation that I can call disconnect with a group_type and with a slot_type but not with a slot_function_type. I'm surprised to run into such a problem as from the point of view of a library user it can't get much easier than that? Boris PS: This is with Boost 1.33.1.

4 13

[parameter] _type not defined in function signature
by Krishna Roskin 03 Jul '07

03 Jul '07

Hey everyone, I'm trying to use Boost.Parameter to make my 9 parameter BGL function more manageable. My code is very close to the example given in the Parameter docs for depth_first_search: BOOST_PARAMETER_NAME(simplified) BOOST_PARAMETER_NAME(original) BOOST_PARAMETER_NAME(initial_vertex) BOOST_PARAMETER_NAME(final_vertex) BOOST_PARAMETER_FUNCTION( (void), simp, tag, (required (in_out(simplified), *) (original, *) ) (optional (initial_vertex, (typename graph_traits<original_type>::vertex_descriptor), original.initial_vertex) (final_vertex, (typename graph_traits<original_type>::vertex_descriptor), original.final_vertex) ) ) { simplify(simplified, original, target_io_edge_map<simplified_type, original_type>(), accumulate_edge_weight(), initial_vertex, final_vertex); } When compiling this under GCC 4.0.1 I get an "'original_type' has not been declared" error in the function signature. The original_type in the function body is fine. If I change the type of initial_vertex and final_vertex to *, then everything works fine. Any idea why this code (which is almost the same as the example given for depth_first_search) doesn't work? Thanks for any help anyone can give and I hope everyone's having a good summer. -krish I'm using 1.34 and here's what gcc -v says: Using built-in specs. Target: i686-apple-darwin8 Configured with: /private/var/tmp/gcc/gcc-5250.obj~12/src/configure --disable-checking -enable-werror --prefix=/usr --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.0/ --with-gxx-include-dir=/include/c++/4.0.0 --build=powerpc-apple-darwin8 --with-arch=pentium-m --with-tune=prescott --program-prefix= --host=i686-apple-darwin8 --target=i686-apple-darwin8 Thread model: posix gcc version 4.0.1 (Apple Computer, Inc. build 5250)

2 2

[BGL] Troubles with floyd_warshall_all_pairs, johnson_all_pairs
by Alex Ivanov 03 Jul '07

03 Jul '07

Hi, guys! I'm stumbled upon a weird problem using SUBJ. functions! Lets suppose that we don't know how much Vertexes we are going to have (for instance Vertexes might be got from file or stdin). So then, in my opinion, is completely impossible to execute SUBJ functions! Let me explain: =============== int xx; //"xx" is runtime-defined vertices size. double* cmat=new(xx*xx); //"cmat" is adjacency matrix actually, it's also defined during runtime. InitCmat(); //here I fill in "cmat" array typedef adjacency_list<listS, vecS, directedS, no_property, property<edge_weight_t, double> > Graph; //This is from some example Graph G(xx); for (int i=0;i<xx;i++) { for (int j=0;j<xx;j++) { if (cmat[i*xx+j]!=0) { add_edge(i,j,cmat[i*xx+j],G); } } } //and now the most interesting part of the code //AFAIK it's impossible to declare array using directive below in runtime: double [xx][xx] D; //Compiler says the same: "C2057: expected constant expression" //even code: double [(const int)xx][(const int)xx] D; //won't work, so the only option left is to make a pointer: double* D = new [xx*xx] ; //but both "floyd_warshall_all_pairs_shortest_paths" and "johnson_all_pairs_shortest_paths" are declared to take a reference, not a pointer, so you'll get multiple compile errors, if you try to compile like below: floyd_warshall_all_pairs_shortest_paths(G, D); //the instruction above will cause multiple compile errors like this: \boost_1_34_0\boost\graph\floyd_warshall_shortest.hpp(115): error C2109: subscript requires array or pointer type P.S. Compiler is MSVC7.1, but I also read that gcc does the same. P.S.S. Sorry for duplication, if this also appears in [devel] list. =============== 1. What can I do to get this code work? 2. If no workaround is available, don't you think that it's a good idea to change the DistanceMatrix type in prototypes to pointers? Thanks in advance! //Alex.

2 2

[wave] thread safety???
by Ovanes Markarian 03 Jul '07

03 Jul '07

Hello Hartmut! I would like to start multiple preprocessing threads. Is wave thread-safe or there some points I should take in account? With Kind Regards, Ovanes Markarian

3 7

[Boost.Graph] Parameter for the Dijkstra-Algorithm
by Ewgenij Sokolovski 03 Jul '07

03 Jul '07

Hello, I read in the documentation that the parameter weight_map of dijkstra_shortest_paths should be a model of Readable Property Map. http://www.boost.org/libs/graph/doc/dijkstra_shortest_paths.html Readable Property Map interface defines a function get(), which has to be contained in each implementation of the interface. OK so far. But when you look at the example provided with the boost library, you see that a simple STL vector is used as weight map. http://www.boost.org/libs/graph/example/dijkstra-example.cpp And STL vector doesn't provide a get() function at all. So the whole thing shouldn't work at all. So where is the magic explanation?:) Ewgenij -- GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS. Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail

2 2