August 2006 - Boost-users - lists.boost.org

(no subject)
by Juan Ramos 22 Aug '06

22 Aug '06

Hi everybody, I'm using VC++ .NET 2003 and when compile a file with connected_components.hpp I get the following compile error: Fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 2844) Please choose the Technical Support command on the Visual C++ What can I do to fix this error? Thanks for your help Juan

1 0

compile error under GCC 3.4.4, but successfully compiled by GCC 3.2.3
by sean yang 22 Aug '06

22 Aug '06

I successfully compiled my program with GCC 3.2.3 on a Linux (Red Hat Linux 3.2.3-56) box. When I tried to compile the same program with GCC 3.4.4 on a Solaris box, it gave me the following errors. Can anyone help me figure out why it is happening? #g++ -O -I./boost_1_25_1 -Wall beaconfsa.cpp ./boost_1_25_1/boost/config/compiler/gcc.hpp:57:7: warning: #warning "Unknown compiler version - please run the configu re tests and report the results" In file included from ./boost_1_25_1/boost/graph/graph_utility.hpp:42, from ./boost_1_25_1/boost/graph/detail/adjacency_list.hpp:43, from ./boost_1_25_1/boost/graph/adjacency_list.hpp:293, from beaconfsa.cpp:19: ./boost_1_25_1/boost/graph/depth_first_search.hpp: In function `void boost::depth_first_search(const VertexListGraph&, const boost::bgl_named_params<P, T, R>&)': ./boost_1_25_1/boost/graph/depth_first_search.hpp:167: error: there are no arguments to `make_dfs_visitor' that depend on a template parameter, so a declaration of `make_dfs_visitor' must be available ./boost_1_25_1/boost/graph/depth_first_search.hpp:167: error: (if you use `-fpermissive', G++ will accept your code, bu t allowing the use of an undeclared name is deprecated) In file included from ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp:39, from beaconfsa.cpp:22: ./boost_1_25_1/boost/pending/mutable_queue.hpp: In member function `bool boost::mutable_queue<IndexedType, RandomAccess Container, Comp, ID>::test()': ./boost_1_25_1/boost/pending/mutable_queue.hpp:131: error: `is_heap' is not a member of `std' In file included from beaconfsa.cpp:22: ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp: In member function `void boost::dijkstra_visitor<Visitors>::edg e_relaxed(Edge, Graph&)': ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp:70: error: `m_vis' undeclared (first use this function) ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp:70: error: (Each undeclared identifier is reported only once for each function it appears in.) ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp: In member function `void boost::dijkstra_visitor<Visitors>::edg e_not_relaxed(Edge, Graph&)': ./boost_1_25_1/boost/graph/dijkstra_shortest_paths.hpp:74: error: `m_vis' undeclared (first use this function) beaconfsa.cpp: In function `int transition_fsa(unsigned int*, unsigned int*)': beaconfsa.cpp:282: warning: comparison between signed and unsigned integer expressions make: *** [beaconfsa.o] Error 1 _________________________________________________________________ Windows Live Spaces is here! Its easy to create your own personal Web site. http://spaces.live.com/signup.aspx

2 1

Re: [Boost-users] bind depending on the type
by Todd Greer 22 Aug '06

22 Aug '06

Minkoo Seo wrote: > vector< pair<string, int> > subjects; > subjects.push_back(make_pair("Math", 90)); > subjects.push_back(make_pair("English", 86)); > subjects.push_back(make_pair("Physics", 88)); > subjects.push_back(make_pair("Biology", 97)); > > long total; > total = accumulate(subjects.begin(), subjects.end(), 0, > _1 + bind(&pair<string, int>::second, _2)); > This is a simple score card application that computes the total score. > As you can see, I've used bind in calling accumulate and that > seems to be ugly. So, I modified the code as follows: Seems fine to me, but it's your code, so your opinion's the important one. > So, what I want to do is write a function get_score that successfuly > compiles and run the following: > > total = accumulate(subjects.begin(), subjects.end(), 0, get_score(_1) + > get_score(_2)); > > I've tried my best only to fail. Could anybody write get_score functions > for me? You're motivation is unclear. The usual reason to use Boost.Lambda is that you want to define all of the logic of the algorithm at the call site, rather than having auxiliary definitions elsewhere. If you're going to have some part of the algorithm get its own named definition, then it's simpler just to use a regular function or function object. Then everything is quite simple: total = accumulate(subjects.begin(), subjects.end(), 0, accumulate_scores); This assumes that accumulate_scores is something like: int accumulate_scores(int a, pair<char const*, int> const& b) { return a + b.second; } If you really do want to write "total = accumulate(subjects.begin(), subjects.end(), 0, get_score(_1) + get_score(_2));", you would need to write two get_score() overloads, one that takes a boost::lambda::placeholder1_type, and one that takes a placeholder2_type. They could return types with an overloaded addition operator that returned a boost::function of the appropriate type, or they could return types equivalent to bind expressions. Although the placeholder1_type overload would simply return its argument, the other would return a long, complex, and I think unspecified type, though I guess you could generate the type with boost::result_of. Frankly, it sounds like far too much trouble; I'd take the simple route, but it's your call. -- Todd Greer

1 0

[BGL] How to add a vertex_descriptor in a graph property
by Manuel Yguel 22 Aug '06

22 Aug '06

Hello, I want to make tree class with the BGL and so I need to specify the root of the tree. Previously I did that by deriving a graph and adding a graph_descriptor member like the code below: typedef boost::adjacency_list< boost::listS, boost::listS, boost::directedS, int > Graph; class Tree : public Graph { public: typedef typename boost::graph_traits< Graph >::vertex_descriptor root; Tree(); }; The constructor create a Vertex then defines root equal to that first vertex. But with bundled property I wanted to know if it is possible to define the root in the graph property. In fact I tried it but I am a bit embarrassed by the cyclic problem: - graph property needs to know what a vertex_descriptor is, - graph needs to know what a graph property is, - graph defines what a vertex_descriptor is. I tried the Barton-Nackman trick but I failed. Is it a proper way to do that ? Some advices, or pointers will be really apreciated. Thanks, Manuel Yguel.

2 1

Is there any corresponding C++ component for java.lang.reflect?
by Bo Xie 22 Aug '06

22 Aug '06

Hi all, I know C++ language does not support Java language's "reflect" mechanism. Is there any corresponding C++ component/library for java.lang.reflect? e.g., Does Boost support "reflect"? Thanks! Best Regards, Xie, Bo

5 6

segfault and no segfault with 2 different boost versions.
by Thierry Dumont 22 Aug '06

22 Aug '06

I develop a C++ code which uses boost::threads * This code performs perfectly well with the boost library wich comes with Debian sarge (libboost-thread1.32.0 for thread library). But I want a more complete implementation of boost (I want to use property tree library). So I have installed and compiled boost_1_33_1. * Using this library, all the programs which use the thread library segfault (at thread-create). Where is the problem ? With the Debian sarge library, I compile with the options: -DBOOST_HAS_THREADS -D_REENTRANT -O2 -funroll-all-loops -fomit-frame-pointer Do I need to change something? Did I missed something when I installed boost_1_33_1? In both cases, g++ version is 3.3.5. -- Thierry Dumont. Institut Camille Jordan -- Mathematiques-- Univ. Lyon I,43 Bd du 11 Novembre 1918, 69622 - Villeurbanne Cedex - France. Thierry.Dumont(a)maply.univ-lyon1.fr web: http://maply.univ-lyon1.fr/~tdumont

2 1

Problem while installin boost 1.33.1
by Leonardo José Reyes Diago 22 Aug '06

22 Aug '06

Hello, i´m trying to install boost 1.33.1, ii'm using windowsXP with visual c++ 2005 express, and using the follow command line: bjam "-sTOOLS=vc-8_0" "—prefix=d:\" install. that have to build a folder with the libraries, but does not do that. This is a part of the message that it show in the command prompt: CALL "C:\Archivos de programa\Microsoft Visual Studio 8\Common7\Tools..\..\VC\bin\vcvars32.BAT" >nul "C:\Archivos de programa\Microsoft Visual Studio 8\Common7\Tools..\..\VC\bin\cl" /Zm800 -nologo /EHsc -c -DBOOST_DATE_TIME_STATIC_LINK -DUSE_DATE_TIME_PRE_1_33_FACET_IO /Z7 /Od /Ob0 /EHsc /GR /MTd /Zc:forScope /Zc:wchar_t -I"bin\boost\libs\date_time\build" -I"C:\boost_1_33_1" -Fo"bin\boost\libs\date_time\build\libboost_date_time.lib\vc-8_0\debug\runtime-link-static\threading-multi\greg_month.obj" -Tp"C:\boost_1_33_1\libs\date_time\build\../src/gregorian/greg_month.cpp" ...failed vc-C++ bin\boost\libs\date_time\build\libboost_date_time.lib\vc-8_0\debug\runtime-link-static\threading-multi\greg_month.obj...vc-C++ bin\boost\libs\date_time\build\libboost_date_time.lib\vc-8_0\debug\runtime-link-static\threading-multi\greg_weekday.obj The system cannot find the specified route . The system cannot find the specified route . ...skipped <@boost!lib>libboost_wave-vc80-mt-1_33_1.lib for lack of <@boost!libs!wave!build\libboost_wave.lib\vc-8_0\release\threading-multi>libboost_wave-vc80-mt-1_33_1.lib... ...skipped <@boost!lib>boost_date_time-vc80-mt-gd.lib for lack of <@boost!lib>boost_date_time-vc80-mt-gd-1_33_1.lib... ...skipped <@boost!lib>boost_date_time-vc80-mt.lib for lack of <@boost!lib>boost_date_time-vc80-mt-1_33_1.lib... ...skipped <@boost!lib>libboost_date_time-vc80-mt-sgd.lib for lack of <@boost!lib>libboost_date_time-vc80-mt-sgd-1_33_1.lib... ...failed updating 808 targets... ...skipped 336 targets... ...updated 2 targets... Can someone help me to solve this? Thanks

2 1

Boost regex multi-processor scalability and user-defined allocators
by Roy Emek 22 Aug '06

22 Aug '06

Iâm using boost-regex library, and Iâm running into multi-threading scalability issues. When tested on a machine with several CPUs, my performance tests show improvement of about 60-70% when moving from a single thread to 2 threads. Question 1: Is anyone aware of this problem? Is there a known solution? >From some testing weâve done, it seems that the cause of the scalability issue is problematic implementation of std:allocator. This is a known issue on some operating systems (e.g., Solaris 8). Some of the boost::regex classes (e.g., match_results) accept a user-defined allocator. However, as far as I can understand thereâs no way to completely override the use of std::allocator. Question 2: is there a way to completely prevent boost / regex from using std::allocator? If no, can such a capability be added? (One solution to this problem would be global overriding the ânewâ and âdeleteâ operators â but for various reasons this cannot be done in the application Iâm developing). __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

3 2

Re: [Boost-users] [Signals] Relationships to other libraries and terms
by Ames Andreas 22 Aug '06

22 Aug '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Scott Meyers > Sent: Tuesday, August 22, 2006 12:07 AM > Subject: [Boost-users] [Signals] Relationships to other > libraries and terms > > We start with "callable entities" -- things that can be called, e.g., > function pointers, function objects, bound member functions, > etc. As I > understand it: > > - A callback (function or object) is a callable entity > - A slot in Boost.Signals is a callable entity > - boost::bind and boost::lambda create callable entities. > - A boost::function object stores a callable entity. > > We also have collections of callable entities: > > - A .NET delegate is a collection of 0 or more callable > entities with a > given signature. > - A signal in Boost.Signals is a collection of 0 or more callable > entities with signatures compatible with a given signature. > > Objects of these collection types are also callable entities: > > - When a Boost.Signals signal or a .NET delegate is > called, it invokes > the callable entities it contains. > > Is that correct? > > Also, does anybody know where the name "slot" comes from in > this context? > I've found this term to be a stumbling block for trying to > understand this > stuff. I think the callability of signals is just a detail. Libsigc++ seems to prefer signal.emit() (although they also have operator()). A slot corresponds to the Observer (in the corresponding GoF pattern), i.e the receipient of a signal (which comes closest to the subject in the GoF pattern). Regarding the terminology: IIRC, it was introduced by a signal/slot lib in C, unfortunately I can't find the links, sorry (I'd also be interested in them, if anyone can find them). There are quite some implementations in C++: Qt, libsigc++, boost and others. On the libsigc++ mailing list was a discussion about the terminology, you can find it in the thread starting with http://mail.gnome.org/archives/libsigc-list/2005-July/msg00014.html. cheers, aa -- Andreas Ames | Programmer | Comergo GmbH | Voice: +49 69 7505 3213 | ames AT avaya DOT com

1 0

Problem binding a predicate to random_shuffle()
by Nelis Franken 22 Aug '06

22 Aug '06

Good day. After an initial post on the comp.lang.c++ usergroup, I was pointed to using boost::bind for binding member functions as predicates for STL functions (original example applied to sort()). The technique to use Boost's bind() works well, except when used with STL's random_shuffle(). The following code demonstrates a working call (on sort()) and a call that generates an error (on random_shuffle()), both using the same technique. The code: #include <algorithm> #include <functional> #include <vector> #include <boost/bind.hpp> using namespace std; class Foo { public: Foo() { for (int i=0; i < 5; i++) { dataVector.push_back(i*10); idVector.push_back(i); } } bool isLess(int leftID, int rightID) { return (dataVector[leftID] < dataVector[rightID]); } int customRand(int n) { return (rand() % n); } void mySort() { sort(idVector.begin(), idVector.end(), boost::bind(&Foo::isLess, this, _1, _2)); } void myShuffle() { random_shuffle(dataVector.begin(), dataVector.end(), boost::bind(&Foo::customRand, this, _1)); } private: vector<int> dataVector; vector<int> idVector; }; int main(int argc, char** argv) { Foo myObj; myObj.myShuffle(); myObj.mySort(); return 0; } The error: sort.cpp: In member function `void Foo::myShuffle()': sort.cpp:33: error: no matching function for call to `random_shuffle(__gnu_cxx::__normal_iterator<int*, std::vector<int, std::allocator<int> > >, __gnu_cxx::__normal_iterator<int*, std::vector<int, std::allocator<int> > >, boost::_bi::bind_t<int, boost::_mfi::mf1<int, Foo, int>, boost::_bi::list2<boost::_bi::value<Foo*>, boost::arg<1> > >)' /usr/lib/gcc/i686-pc-cygwin/3.4.4/include/c++/bits/stl_algo.h:1766: note: candidates are: void std::random_shuffle(_RandomAccessIterator, _RandomAccessIterator, _RandomNumberGenerator&) [with _RandomAccessIterator = __gnu_cxx::__normal_iterator<int*, std::vector<int, std::allocator<int> > >, _RandomNumberGenerator = boost::_bi::bind_t<int, boost::_mfi::mf1<int, Foo, int>, boost::_bi::list2<boost::_bi::value<Foo*>, boost::arg<1> > >] The code compiles and runs fine using VC7.1 (Visual Studio 2003), but gives an error using g++ (GCC) 3.4.4 (cygming special). Any help will be greatly appreciated. Kind regards, Nelis Franken PS: A follow-up post on the exact same problem to comp.lang.c++ directed me to post it on the Boost list, since it was deemed off-topic. I hope someone else on this list has had a similar STL problem with random_shuffle() and boost::bind?

3 4