September 2007 - Boost-users

Problem while using Dijkstra algorithm
by abhishek.v＠tcs.com 12 Sep '07

12 Sep '07

Hi Considering the example that comes with BGL i.e examples/dijkstra-example.cpp. I have already a graph which is undirected now i want to apply dijkstra algorithm on it. My vertex discriptor is "Vertex_name" and edge discriptor is "edge_value". The following is the code provided in the example. :- std::vector<vertex_descriptor> p(num_vertices(g)); std::vector<int> d(num_vertices(g)); vertex_descriptor s = vertex(A, 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, &p[0], &d[0], weightmap, indexmap, std::less<int>(), closed_plus<int>(), (std::numeric_limits<int>::max)(), 0, default_dijkstra_visitor()); #else dijkstra_shortest_paths(g, s, predecessor_map(&p[0]).distance_map(&d[0])); #endif My code for the corresponding code is as follows :- std::vector<vertex_descriptor> p(num_vertices(g)); std::vector<int> d(num_vertices(g)); vertex_descriptor s = vertex(L1, g); //Giving error Undeclared identifier L1. Here L1 is the name of my vertex. Also i tried Vertex_name(u) to this parameter but it is giving compilation error. What parameter need to be passed for this. Secondly i m not defining BOOST_MSVC so i'm elimination the IF BLock and directly writing the code as dijkstra_shortest_paths(g, s, predecessor_map(&p[0]).distance_map(&d[0])); Here kindly explain what the distance_map signifies and where it is being defined. Thanks Abhishek Vyas Tata Consultancy Services Mailto: abhishek.v(a)tcs.com Website: http://www.tcs.com ____________________________________________ Experience certainty. IT Services Business Solutions Outsourcing ____________________________________________ =====-----=====-----===== Notice: The information contained in this e-mail message and/or attachments to it may contain confidential or privileged information. If you are not the intended recipient, any dissemination, use, review, distribution, printing or copying of the information contained in this e-mail message and/or attachments to it are strictly prohibited. If you have received this communication in error, please notify us by reply e-mail or telephone and immediately and permanently delete the message and any attachments. Thank you

2 1

Failing to build dynamic regexp libraries on Mac OS X
by Tommy Nordgren 12 Sep '07

12 Sep '07

Building dynamic regular experssion libraries fails on Mac OS X Tiger , when attempting to link against ICU. I would appreciate any hint on how to fix this. All the ICU libraries are present in /usr/local/lib, so I wonder if I need to add -licudata to the build somehow. (As can be seen from the build log, only -licui18n and -licuuc are added to the command line. (I've tried to edit the boost distribution files that refer to theese flags, to also include -icudata , but the command line passed to the tools don't change, so how do I do it) ./tools/jam/src/bin.macosxppc/bjam -sICU_PATH=/usr/local --user- config=user-config.jam Building Boost.Regex with Unicode/ICU support enabled Using ICU in /usr/local/include ...patience... ...patience... ...found 4691 targets... ...updating 4 targets... darwin.link.dll bin.v2/libs/regex/build/darwin/release/threading- multi/libboost_regex-mt-1_34_1.dylib ld: warning can't open dynamic library: libicudata.36.dylib referenced from: /usr/local/lib/libicui18n.dylib (checking for undefined symbols may be affected) (No such file or directory, errno = 2) ld: Undefined symbols: _icudt36_dat referenced from libicuuc.36 expected to be defined in libicudata.36.dylib /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/libtool: internal link edit command failed g++ -dynamiclib -o "bin.v2/libs/regex/build/darwin/release/ threading-multi/libboost_regex-mt-1_34_1.dylib" "bin.v2/libs/regex/ build/darwin/release/threading-multi/c_regex_traits.o" "bin.v2/libs/ regex/build/darwin/release/threading-multi/cpp_regex_traits.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/cregex.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/fileiter.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/icu.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/instances.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/posix_api.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/regex.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/ regex_debug.o" "bin.v2/libs/regex/build/darwin/release/threading- multi/regex_raw_buffer.o" "bin.v2/libs/regex/build/darwin/release/ threading-multi/regex_traits_defaults.o" "bin.v2/libs/regex/build/ darwin/release/threading-multi/static_mutex.o" "bin.v2/libs/regex/ build/darwin/release/threading-multi/w32_regex_traits.o" "bin.v2/libs/ regex/build/darwin/release/threading-multi/wc_regex_traits.o" "bin.v2/ libs/regex/build/darwin/release/threading-multi/wide_posix_api.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/winstances.o" "bin.v2/libs/regex/build/darwin/release/threading-multi/ usinstances.o" -licui18n -licuuc ...failed darwin.link.dll bin.v2/libs/regex/build/darwin/release/ threading-multi/libboost_regex-mt-1_34_1.dylib... darwin.link.dll bin.v2/libs/regex/build/darwin/debug/libboost_regex- d-1_34_1.dylib ld: warning can't open dynamic library: libicudata.36.dylib referenced from: /usr/local/lib/libicui18n.dylib (checking for undefined symbols may be affected) (No such file or directory, errno = 2) ld: Undefined symbols: _icudt36_dat referenced from libicuuc.36 expected to be defined in libicudata.36.dylib /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/libtool: internal link edit command failed g++ -dynamiclib -o "bin.v2/libs/regex/build/darwin/debug/ libboost_regex-d-1_34_1.dylib" "bin.v2/libs/regex/build/darwin/debug/ c_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/ cpp_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/cregex.o" "bin.v2/libs/regex/build/darwin/debug/fileiter.o" "bin.v2/libs/regex/ build/darwin/debug/icu.o" "bin.v2/libs/regex/build/darwin/debug/ instances.o" "bin.v2/libs/regex/build/darwin/debug/posix_api.o" "bin.v2/libs/regex/build/darwin/debug/regex.o" "bin.v2/libs/regex/ build/darwin/debug/regex_debug.o" "bin.v2/libs/regex/build/darwin/ debug/regex_raw_buffer.o" "bin.v2/libs/regex/build/darwin/debug/ regex_traits_defaults.o" "bin.v2/libs/regex/build/darwin/debug/ static_mutex.o" "bin.v2/libs/regex/build/darwin/debug/ w32_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/ wc_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/ wide_posix_api.o" "bin.v2/libs/regex/build/darwin/debug/winstances.o" "bin.v2/libs/regex/build/darwin/debug/usinstances.o" -licui18n - licuuc -g ...failed darwin.link.dll bin.v2/libs/regex/build/darwin/debug/ libboost_regex-d-1_34_1.dylib... darwin.link.dll bin.v2/libs/regex/build/darwin/release/ libboost_regex-1_34_1.dylib ld: warning can't open dynamic library: libicudata.36.dylib referenced from: /usr/local/lib/libicui18n.dylib (checking for undefined symbols may be affected) (No such file or directory, errno = 2) ld: Undefined symbols: _icudt36_dat referenced from libicuuc.36 expected to be defined in libicudata.36.dylib /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/libtool: internal link edit command failed g++ -dynamiclib -o "bin.v2/libs/regex/build/darwin/release/ libboost_regex-1_34_1.dylib" "bin.v2/libs/regex/build/darwin/release/ c_regex_traits.o" "bin.v2/libs/regex/build/darwin/release/ cpp_regex_traits.o" "bin.v2/libs/regex/build/darwin/release/cregex.o" "bin.v2/libs/regex/build/darwin/release/fileiter.o" "bin.v2/libs/ regex/build/darwin/release/icu.o" "bin.v2/libs/regex/build/darwin/ release/instances.o" "bin.v2/libs/regex/build/darwin/release/ posix_api.o" "bin.v2/libs/regex/build/darwin/release/regex.o" "bin.v2/ libs/regex/build/darwin/release/regex_debug.o" "bin.v2/libs/regex/ build/darwin/release/regex_raw_buffer.o" "bin.v2/libs/regex/build/ darwin/release/regex_traits_defaults.o" "bin.v2/libs/regex/build/ darwin/release/static_mutex.o" "bin.v2/libs/regex/build/darwin/ release/w32_regex_traits.o" "bin.v2/libs/regex/build/darwin/release/ wc_regex_traits.o" "bin.v2/libs/regex/build/darwin/release/ wide_posix_api.o" "bin.v2/libs/regex/build/darwin/release/ winstances.o" "bin.v2/libs/regex/build/darwin/release/usinstances.o" -licui18n -licuuc ...failed darwin.link.dll bin.v2/libs/regex/build/darwin/release/ libboost_regex-1_34_1.dylib... darwin.link.dll bin.v2/libs/regex/build/darwin/debug/threading-multi/ libboost_regex-mt-d-1_34_1.dylib ld: warning can't open dynamic library: libicudata.36.dylib referenced from: /usr/local/lib/libicui18n.dylib (checking for undefined symbols may be affected) (No such file or directory, errno = 2) ld: Undefined symbols: _icudt36_dat referenced from libicuuc.36 expected to be defined in libicudata.36.dylib /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/libtool: internal link edit command failed g++ -dynamiclib -o "bin.v2/libs/regex/build/darwin/debug/ threading-multi/libboost_regex-mt-d-1_34_1.dylib" "bin.v2/libs/regex/ build/darwin/debug/threading-multi/c_regex_traits.o" "bin.v2/libs/ regex/build/darwin/debug/threading-multi/cpp_regex_traits.o" "bin.v2/ libs/regex/build/darwin/debug/threading-multi/cregex.o" "bin.v2/libs/ regex/build/darwin/debug/threading-multi/fileiter.o" "bin.v2/libs/ regex/build/darwin/debug/threading-multi/icu.o" "bin.v2/libs/regex/ build/darwin/debug/threading-multi/instances.o" "bin.v2/libs/regex/ build/darwin/debug/threading-multi/posix_api.o" "bin.v2/libs/regex/ build/darwin/debug/threading-multi/regex.o" "bin.v2/libs/regex/build/ darwin/debug/threading-multi/regex_debug.o" "bin.v2/libs/regex/build/ darwin/debug/threading-multi/regex_raw_buffer.o" "bin.v2/libs/regex/ build/darwin/debug/threading-multi/regex_traits_defaults.o" "bin.v2/ libs/regex/build/darwin/debug/threading-multi/static_mutex.o" "bin.v2/ libs/regex/build/darwin/debug/threading-multi/w32_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/threading-multi/ wc_regex_traits.o" "bin.v2/libs/regex/build/darwin/debug/threading- multi/wide_posix_api.o" "bin.v2/libs/regex/build/darwin/debug/ threading-multi/winstances.o" "bin.v2/libs/regex/build/darwin/debug/ threading-multi/usinstances.o" -licui18n -licuuc -g ...failed darwin.link.dll bin.v2/libs/regex/build/darwin/debug/ threading-multi/libboost_regex-mt-d-1_34_1.dylib... ...failed updating 4 targets... Not all Boost libraries built properly. ------------------------------------- This sig is dedicated to the advancement of Nuclear Power Tommy Nordgren tommy.nordgren(a)comhem.se

2 2

Re: [Boost-users] Using Boost.Fusion transform on vector<int&>
by dan marsden 12 Sep '07

12 Sep '07

Martin Ecker wrote: >I need something like the following (somewhat contrived) example to >work with boost::fusion::transform on a boost::fusion::vector<int&> or >other vector types that have reference types. >Note that the following doesn't compile. <snip original example> The following code allows references to pass through, as I believe you want. It just uses a slight modification to your function object to allow it to cope with references correctly. namespace { struct pass_through { template<typename Sig> struct result; template<typename T> struct result<pass_through(T&)> { typedef T& type; }; template<typename T> T& operator()(T& t) const { return t; } }; } int main() { typedef fusion::vector<int&> vec; int i = 202; vec v1(i); vec v2(fusion::transform(v1, pass_through())); std::cout << fusion::at_c<0>(v2) << std::endl; // Output 202 } Is this ok for your needs? Cheers Dan ___________________________________________________________ Yahoo! Answers - Got a question? Someone out there knows the answer. Try it now. http://uk.answers.yahoo.com/

2 1

deciding directed and undirected parameter of adjancency list at runtime .
by abhishek.v＠tcs.com 11 Sep '07

11 Sep '07

Hi I have a function which take an input from the user whether the graph is directed or undirected. I'm declaring graph as follows :- typedef adjacency_list < vecS, vecS, undirectedS, property < vertex_name_t,std::string >, property < edge_name_t, std::string > > Graph; Graph g; Now here the parameter in Bold need to be varied as the user input changes. One method could be i could have two different function say,CreateDirected() and CreateUndirected(). But this is not working in my case as i want the declaration should be global so that i can use it throughout my application. Please help. Thanks Abhishek Vyas Tata Consultancy Services Mailto: abhishek.v(a)tcs.com Website: http://www.tcs.com ____________________________________________ Experience certainty. IT Services Business Solutions Outsourcing ____________________________________________ =====-----=====-----===== Notice: The information contained in this e-mail message and/or attachments to it may contain confidential or privileged information. If you are not the intended recipient, any dissemination, use, review, distribution, printing or copying of the information contained in this e-mail message and/or attachments to it are strictly prohibited. If you have received this communication in error, please notify us by reply e-mail or telephone and immediately and permanently delete the message and any attachments. Thank you

2 2

Using Boost.Fusion transform on vector<int&>
by Martin Ecker 11 Sep '07

11 Sep '07

Hi, I need something like the following (somewhat contrived) example to work with boost::fusion::transform on a boost::fusion::vector<int&> or other vector types that have reference types. Note that the following doesn't compile. --------------------------- #include <boost/fusion/sequence/container/vector.hpp> #include <boost/fusion/algorithm/transformation/transform.hpp> struct pass_through { template <typename SignatureT> struct result; template <typename T> struct result<pass_through (T)> { typedef T type; }; template <typename T> T operator () (T value) const { return value; } }; int main() { using namespace boost::fusion; int i = 5; vector<int&> v(i); vector<int&> v2(transform(v, pass_through())); return 0; } --------------------------- This code doesn't compile because in deref_impl for the transform_view when this code static type call(Iterator const& i) { return i.f(*i.first); } calls the pass_through function object's operator (), the compiler deduces T to be int instead of int&. Is there any way to make this work with Boost.Fusion, e.g. by somehow additionally passing the actual type stored in the vector to the function object in some way instead of relying on the compiler to deduce it? I was thinking of maybe using the binary version of transform and passing in a range_c containing integral constants from 0 to <size of vector - 1> and then using this in operator () to retrieve the actual type from the vector by using mpl::at. Is there a better/simpler way? Thanks, Martin

1 0

[spirit]how to ignore the sequence of the rules?
by james.jones＠firstinvestors.com 11 Sep '07

11 Sep '07

> From: moosefly <moosefly(a)gmail.com> > For example, I want to parse the following text: > > <input type=hidden value=1> > > I have defined 2 rules match the "type" field and the "value" field, > but the sequence of 2 rules is uncertain, that means "<input value=1 > type=hidden>" is also legal,so I can't define the rules just like: > > rule_type >> !space_p >> rule_value > > The following works at this simple condition: > > ( (rule_type >> !space_p >> rule_value) | (rule_value >> !space_p > >>rule_type ) ) > > But if there are many out-of-order fields need to parser, how can I > ignore the sequence of the rules? What you seem to want to do (correct me if I'm wrong) is to have some set of N attributes, each of which may or may not be present, but can appear only once, in arbitrary order, and you want to check this in the parser. This is not a great idea, because it leads to a combinatorial explosion of states. There's really no way out of this (strictly within the parser). What you CAN do to avoid the explosion is to delegate the uniqueness checking to helper functions outside the parser. So instead of allowing only rule_type then rule_value or the reverse, allow this: rule_any = rule_type | rule_value; rule_input_attributes = *rule_any; // ignoring whitespace issues here This will match: <input type=hidden value=1> or <input value=1 type=hidden> as you want. But it'll also match: <input value=1 type=hidden type=visible> and <input> which you might not want. So you have to add an action to rule_any, which will pass the name-value pair data to a function which will check the uniqueness of the names in the NVP. I'm not an expert on Spirit, so you might want to consult the Spirit mailing list if you get stuck creating actions. HTH, - James Jones Administrative Data Mgmt. Webmaster 375 Raritan Center Pkwy, Suite A Data Architect Edison, NJ 08837

2 1

Re: [Boost-users] [multi_index] erasing elements
by Jeff Holle 11 Sep '07

11 Sep '07

Joaquín Mª López Muñoz wrote: > The two snippets of code you show are entirely equivalent, in the > sense that > Present inherits the functionality of its first index, YIndex. In > particular > YIndex::iterator is the *same type* as Present::iterator, so I'm > puzzled by the > fact that you're observing different behaviors --from the code you're > showing, > using one or the other variation cannot possibly make any difference > whatsoever. > Could you please check your code for some other differences you might > have not noticed when doing the switch from one version to the other? Thanks for this information. The two code snippets were not equivalent because I was only using the iterator returned from the multi_index container's erase method in the one using YIndex::iterators. I have found that this iterator doesn't point to an element in the container. Instead its pointing to some memory that is zeroed. Prior to erasing the element I inserted some test code that looks though the entire container and doesn't find such an element. Given my constructor, this only makes sense. Do you have an explanation?

2 1

[Serialization] linking problems leading to unregistered class exception
by David Raulo 11 Sep '07

11 Sep '07

Hi, It seems I have linker-related problems leading to "unregistered class" exceptions being raised. Now, I know this sounds like a compilation chain-specific problem, but the thing is I get the exact same problem on Linux (Ubuntu 7.04, gcc4.1), Macosx 10.4 (gcc 4.0) and Windows (Microsoft Visual Studio 2005 Express). My initial problem was that I'm using Boost.Serialization on a rather big class hierarchy (around 200 classes), so instantiating all the serialization code in a single .cpp is not pratical (too much memory consumed by the compiler). So I split my BOOST_CLASS_EXPORT_GUID() declarations into several compilation units, and this worked fine. *Except* if I generate a static library, in which case I get unregistered class exceptions. If I take the same object files and link them into a dynamic library, or directly into my executable, then everything works as intended. So it looks like generating a static library (with gcc or msvc) leads to parts of the serialization code to disappear (a quick look at the file-size of the generated executables seems to confirm that). I attached the smallest test-case I could think of that reproduces this problem. The included Makefile will generate 3 executables: test_s11n, test_s11n_dyn, test_s11n_static. All of them build fine, but the last one fails at run-time. Anyone has some insight into this? Thanks, david.

2 2

Unit Test Framework add
by Chris Miceli 11 Sep '07

11 Sep '07

I have had nothing but troubles with boost test 1.34.1 after switched from 1.33. First I had trouble with the main from shared object issue, and now my code that worked before is not executing a test added to the framework. #include <iostream> #define BOOST_TEST_DYN_LINK #define BOOST_TEST_MAIN #define BOOST_TEST_MODULE File Test #include <boost/bind.hpp> #include <boost/test/unit_test.hpp> template <typename Functor> struct my_test { void execute(int value) { //Never prints this message //Doesn't have to be a print, nothing gets here std::cerr << "Hi all" << std::endl << std::flush; } }; BOOST_AUTO_TEST_CASE( test123 ) { boost::shared_ptr<my_test<int> > tester(new my_test<int>); boost::unit_test::framework::master_test_suite().add(BOOST_TEST_CASE(boost::bind(&my_test<int>::execute, tester, 5))); } this worked when it was not using boost::bind but just the address of a normal function. This is just an simple version that illustrated the problem. I am compiling as such g++ test.cpp -o test -L/usr/local/lib -lboost_unit_test_framework-gcc41-mt This is the output: Running 1 test case... *** No errors detected

2 2

XML Serialization
by tcmichals 11 Sep '07

11 Sep '07

Creating an application to send XML formatted data to this class that uses boost serialization, the issue is the name spaces in the boost_serialization tag; for example shown below... <?xml version="1.0" encoding="utf-8"?> <!DOCTYPE boost_serialization > <boost_serialization xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" signature="serialization::archive" version="4">... looking at the rule for parsing, SerializationWrapper = !S >> str_p(L"<boost_serialization") >> S >> SignatureAttribute >> S >> VersionAttribute >> !S >> L'>' Is there a way to change the rules to allow skipping the namespaces? all of the elements are there but there are a few extra? The issue is trying to change the XML parser not to generate may not be possible .NET.

1 1