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

Re: [Boost-users] [math/special_functions] newbie question: compiler error with factorial
by Sakaie, Ken 11 Jun '08

11 Jun '08

Thank you Steven and John for the fixes, explanation, and fast reply! =================================== P Please consider the environment before printing this e-mail Cleveland Clinic is ranked one of the top hospitals in America by U.S. News & World Report (2007). Visit us online at http://www.clevelandclinic.org for a complete listing of our services, staff and locations. Confidentiality Note: This message is intended for use only by the individual or entity to which it is addressed and may contain information that is privileged, confidential, and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient or the employee or agent responsible for delivering the message to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and destroy the material in its entirety, whether electronic or hard copy. Thank you.

1 0

program_options: emit usage message on unknown option ?
by Stephen Collyer 11 Jun '08

11 Jun '08

Is it possible to have boost::program_options emit a usage message when an unknown option is supplied ? I have code which is parsing its options like this: > void UserAPI::parse_options(const std::vector<std::string>& required_options, > int argc, char* argv[], > std::string usage) > { > po::store(po::parse_command_line(argc, argv, options_desc_, options_style_), vm_); > po::notify(vm_); > > // count how many required options we have > > uint count = 0; > for (std::vector<std::string>::const_iterator s = required_options.begin(); > s != required_options.end(); ++s) > { > count += vm_.count(*s); > } > > // if we have a help option, or we don't have all the required options > // emit the usage message > > if ( vm_.count("help") || count != required_options.size() > ) > { > std::cout << usage << std::endl; > std::cout << options_desc_ << std::endl; > exit(1); > } > } If I pass an unknown option, it seems to be handled automagically: > install/linux/general -orgid ABC -p fred > unknown option -p I don't want this - I want to catch it somehow and emit my usage message, but this seems not to be documented, AFAICS. Anyone know how this can be done ? -- Regards Steve Collyer Netspinner Ltd

2 2

shared_ptr and multi_set
by Robert Kubrick 11 Jun '08

11 Jun '08

I am having a memory error when accessing a shared_ptr in a multi_set collection. I caught this comment in the shared_ptr documentation: int main() { std::vector<FooPtr> foo_vector; std::set<FooPtr,FooPtrOps> foo_set; // NOT multiset! ... http://www.boost.org/doc/libs/1_35_0/libs/smart_ptr/example/ shared_ptr_example.cpp So I guess shared_ptr in multi_set are not supported?

2 1

[math/special_functions] newbie question
by Sakaie, Ken 11 Jun '08

11 Jun '08

I did the following 'hello-world'-level test of the factorial function and get the following error. > cat test.cpp #include "boost/math/special_functions/factorials.hpp" int main() { unsigned int i = 6; printf("%d\n", boost::math::factorial(i)); } > g++ -I/home/sakaiek/ext/src/boost_1_35_0/ test.cpp test.cpp: In function int main(): test.cpp:5: error: no matching function for call to factorial(unsigned int&) I think I set up the include directory correctly: > ls /home/sakaiek/ext/src/boost_1_35_0/boost/math/special_functions acosh.hpp beta.hpp detail ellint_3.hpp ellint_rj.hpp fpclassify.hpp laguerre.hpp math_fwd.hpp sinhc.hpp asinh.hpp binomial.hpp digamma.hpp ellint_rc.hpp erf.hpp gamma.hpp lanczos.hpp powm1.hpp sin_pi.hpp atanh.hpp cbrt.hpp ellint_1.hpp ellint_rd.hpp expm1.hpp hermite.hpp legendre.hpp sign.hpp spherical_harmonic.hpp bessel.hpp cos_pi.hpp ellint_2.hpp ellint_rf.hpp factorials.hpp hypot.hpp log1p.hpp sinc.hpp sqrt1pm1.hpp Any help would be greatly appreciated. Thanks! =================================== P Please consider the environment before printing this e-mail Cleveland Clinic is ranked one of the top hospitals in America by U.S. News & World Report (2007). Visit us online at http://www.clevelandclinic.org for a complete listing of our services, staff and locations. Confidentiality Note: This message is intended for use only by the individual or entity to which it is addressed and may contain information that is privileged, confidential, and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient or the employee or agent responsible for delivering the message to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and destroy the material in its entirety, whether electronic or hard copy. Thank you.

3 3

Re: [Boost-users] Serialization triggers .lib and .exp file generationin VS2005
by spock＠snafu.de 11 Jun '08

11 Jun '08

Robert Ramey wrote: > Thomas Voigt wrote: >> Robert Ramey schrieb: >>> This is a side effect of the steps taken inhibit the compiler from >>> eliminationg code not explicitly referred to. (aka code stripping). >> >> So is there a way around it? Since we use boost in a rather controlled >> environment (only VS2005) we could patch it locally. > > well, we wouldn't put it in there if there was an easy way around it. Sure :-) > You could comment out all the _export keywords in the code. But > its doubtful that serialization of pointers could be made to work. Could you elaborate what the problem is exactly? We don't want to serialize pointers, at least not at the moment. Regards, Thomas

1 0

Using boost file memory mapping (managed_mapped_file) with a vector
by Henry 11 Jun '08

11 Jun '08

I am trying to use a vector with a managed_mapped_file. http://www.boost.org/doc/libs/1_35_0/doc/html/boost/interprocess/vector.html seems to indicate this can be done. When I try to run the code below, it fails with a "bus error" (on OS X) after printing "Entering item in vector". What am I doing wrong? Also, how can I grow this managed_mapped_file as I add things to the vector? I get runtime errors using the managed_mapped_file::grow() function. The code below should just fill up a vector with powers of 2, and print it. I am trying to understand the library before I implement it for something real. Code below: #include <boost/interprocess/file_mapping.hpp> #include <boost/interprocess/mapped_region.hpp> #include <boost/interprocess/containers/vector.hpp> #include <boost/interprocess/managed_mapped_file.hpp> #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/containers/vector.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <iostream> #include <fstream> #include <cstring> #include <cstddef> #include <cstdio> #include <cmath> int main () { using namespace boost::interprocess; try{ std::cout << "Starting" << std::endl; managed_mapped_file mfile (open_or_create, "MyMappedFile", 65536); std::cout << "Mapped file" << std::endl; typedef allocator<int, managed_shared_memory::segment_manager> ShmemAllocator; typedef vector<int, ShmemAllocator> MyVector; const ShmemAllocator alloc_inst (mfile.get_segment_manager()); std::cout << "Making vector" << std::endl; MyVector* vec = mfile.construct<MyVector>("vec")(alloc_inst); std::cout << "Made vector" << std::endl; std::cout << "Filling vector" << std::endl; int i = 0; for(i = 0; i < 63; ++i) { std::cout << "Entering item in vector" << std::endl; vec->push_back(pow(2,i)); std::cout << "Entered item in vector" << std::endl; } std::cout << "Filled vector" << std::endl; std::cout << "Size of vec is:\t" << vec->size() << std::endl; std::cout << "Reverse printing contents of vector" << std::endl; for(i = vec->size()-1; i >= 0; --i) { std::cout << "At " << i << ":\t" << vec->at(i) << std::endl; } std::cout << "Printed contents of vector" << std::endl; } catch(interprocess_exception &ex){ std::remove("MyMappedFile"); std::cout << ex.what() << std::endl; return 1; } std::remove("MyMappedFile"); return 0; }

1 0

Re: [Boost-users] darwin universal static libs for release and debug for 1.35.0
by Daniel Lord 11 Jun '08

11 Jun '08

> > From: Jeff Flinn <TriumphSprint2000(a)hotmail.com> > To: boost-users(a)lists.boost.org Date: Tue, 10 Jun 2008 16:25:43 -0400 Subject: Re: [Boost-users] darwin universal static libs for release and debug for 1.35.0 Jeff Flinn wrote: Jeff Flinn wrote: > >> Trying to build universal static libs for darwin. >> >> bjam --toolset=darwin --architecture=combined stage >> >> appears to only build libboost_*_mt-1_35.dylib's and >> libboost_test_exec_monitor-mt-1_35.a. Whereas the default for 1.34.1 was to >> build all libs. >> >> I need to build release and debug static libs. I'm having problems finding >> the command line flags in the docs. >> >> This is my first foray into mac/unix builds for boost. Any help is >> appreciated. >> >> Thanks, Jeff >> > > Ok, that was easy after reading the boost build docs: > > bjam --toolset=darwin --architecture=combined link=static threading=multi > stage debug release > For Posterity, the full command req'dis: bjam --toolset=darwin --architecture=combined link=static threading=multi address-model=32 macosx-version=10.4 stage debug release Thanks, Jeff Due to issues with the configs for OS X, you have to use the bjam calls and Jeff pointed out the correct ones.Nota Bene: if you want to build 66-bit libs, not all of them build (for shared libs) and you have a to add: macosx-version=10.5 In my shell script I made for building various configurations I use the following command where: link_type = static | shared local_arch = x86 | power local_address_model = 32 | 64 bjam release debug toolset=darwin \ link=${link_type} \ threading=multi \ architecture=${local_arch} \ address-model=${local_address_model} \ macosx-version=10.5 Works for me. -- "It's not how many years you live--it's how much life you put in your years" -- Sierra Mountain Guides

1 0

[Graph] Problem Compiling BGL-Python
by ZSS 11 Jun '08

11 Jun '08

Hi, I'm facing difficulty in compiling BGL-Python (which I'm not familiar with). I followed all the instructions in the README file, however, when I reach the stage where I need to run bjam I recieve this error: error: Could not find parent for project at '.' error: Did not find Jamfile.jam or Jamroot.jam in any parent directory. However, when I move the bgl-python-0.9 directory and put it under boost_1_35_0 I get a different error message: Jamfile:13: in modules.load rule project-root unknown in module Jamfile I don't know what to do, any hints are highly appreciated. Thank you, Zee

1 0

graph visualization
by Dennis Yurichev 10 Jun '08

10 Jun '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, I'm looking for some open-source solution which is able to vizualize boost::graph structures. I mean, I develope software which use boost::graph extensively and I wish to offer end-user to see it visually and edit it. It's not a problem to export to some GDL file for example, and use external tool, but user interaction should be present. Are there anything ready-made or I doomed to develope my own? P.S. Target is Win32 GDI API. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (MingW32) iD8DBQFITurt1YPmFmJG++MRAmYTAKDe0rEYszytT7f70y6up3f5js45ngCfTQHP uJ5SJhSb0tKlNG62DwwMJoI= =193P -----END PGP SIGNATURE-----

3 2

darwin universal static libs for release and debug for 1.35.0
by Jeff Flinn 10 Jun '08

10 Jun '08

Trying to build universal static libs for darwin. bjam --toolset=darwin --architecture=combined stage appears to only build libboost_*_mt-1_35.dylib's and libboost_test_exec_monitor-mt-1_35.a. Whereas the default for 1.34.1 was to build all libs. I need to build release and debug static libs. I'm having problems finding the command line flags in the docs. This is my first foray into mac/unix builds for boost. Any help is appreciated. Thanks, Jeff

1 2

boost filesystems and case issues
by brad＠16systems.com 10 Jun '08

10 Jun '08

I've looked through the documentation and Googled, but I cannot find any info on the topic. If I want to find a file named 'System.dll', but on some systems the file is named 'system.DLL' or 'SYSTEM.DLL', etc. If I could could do something like this: if (boost::filesystem::is_regular(itr->path())) { std::cout << itr->path().lower() << std::endl; } Notice the .lower() In this way, each path would be lowercase and I could search for only 'system.dll' and .lower() would find any occurrence of that filename, no matter the case. I wonder how others do this today? Do you use transform or tolower from another library? Thank you

4 3

A question about boost::bind function
by 风舞影天 10 Jun '08

10 Jun '08

I'm a newbie in boost library and sorry for my poor English. I have a project now, it was created in Visual Studio.Net 2003 and the version of the boost library is 1.33.1, all thing is good. But, When I replaced the boost library with the new version 1.34.1 or 1.35.0, the project can't compiled correct. The error message is 'can't convert const sometype& to sometype&'. So, i force myself to read the source code of boost library and compare them betwwen version 1.33.1 and version 1.44.1. at last, I found the problem and written a test program. The following code is the test program: #include "stdafx.h" #include <boost/bind.hpp> #include <vector> #include <algorithm> struct _sUserData { int m_nUserId; int m_nUserAge; _sUserData() : m_nUserId( rand() ) , m_nUserAge( rand()%100+1 ) { } }; void AddUserAge( int& age ) { ++age; } int _tmain(int argc, _TCHAR* argv[]) { std::vector<_sUserData> vUserData; for( int i = 0; i < 10; ++i ) vUserData.push_back( _sUserData() ); std::for_each( vUserData.begin(), vUserData.end(), boost::bind( AddUserAge, boost::bind( &_sUserData::m_nUserAge, _1 ) ) ); return 0; } If i change the function AddUserAge( int& age ) to AddUserAge( const int& age ), all things is good. but I need change this parameter. In the version 1.33.1, the data member formal boost::bind function was defined as following: template<class R, class T, class A1> _bi::bind_t< R, _mfi::dm<R, T>, typename _bi::list_av_1<A1>::type > BOOST_BIND(R T::*f, A1 a1) { typedef _mfi::dm<R, T> F; typedef typename _bi::list_av_1<A1>::type list_type; return _bi::bind_t<R, F, list_type>( F(f), list_type(a1) ); } In the version 1.35.0, the boost::bind function was defined as following: template< class A1, class M, class T > _bi::bind_t< typename _bi::dm_result< M T::*, A1 >::type, _mfi::dm<M, T>, typename _bi::list_av_1<A1>::type > BOOST_BIND( M T::*f, A1 a1 ) { typedef typename _bi::dm_result< M T::*, A1 >::type result_type; typedef _mfi::dm<M, T> F; typedef typename _bi::list_av_1<A1>::type list_type; return _bi::bind_t< result_type, F, list_type >( F( f ), list_type( a1 ) ); } Although the code that is similar to the version 1.33.1 is existed, but it is written in a macro condition, and in my environment is invalid, so ignore it. The difference is the return type. The boost::_bi::dm_result is defined as following: namespace _bi { template< class Pm, int I > struct add_cref; template< class M, class T > struct add_cref< M T::*, 0 > { typedef M type; }; template< class M, class T > struct add_cref< M T::*, 1 > { typedef M const & type; }; template<class R> struct isref { enum value_type { value = 0 }; }; template<class R> struct isref< R& > { enum value_type { value = 1 }; }; template<class R> struct isref< R* > { enum value_type { value = 1 }; }; template<class Pm, class A1> struct dm_result { typedef typename add_cref< Pm, 1 >::type type; }; template<class Pm, class R, class F, class L> struct dm_result< Pm, bind_t<R, F, L> > { typedef typename bind_t<R, F, L>::result_type result_type; typedef typename add_cref< Pm, isref< result_type >::value >::type type; }; } the compiler will convert the result type to const reference, then pass it to user's function, the error is caught. Now, all things is come out. But I don't understand why add const reference to the result type and the behavior is difference with the version 1.33.1. Wait for your answer and thanks for your time, and sorry for my poor in English:)

4 4

[boost-users][tuple, bind] binding tuple element
by Igor R 10 Jun '08

10 Jun '08

Hello, Given a vector of tuples, like this: vector<tuple<int, std::string> > tuples_; I want create a binder to find a tuple by one of its "attributes". If it would be a struct: struct Tuple { int id_; string name_; }; ...I'd use it like this: find_if(tuples_.begin(), tuples_.end(), bind(&Tuple::name_, _1) == "kuku"); How can bind tuples::get<1>() function? Thanks a lot, Igor'.

2 1

More Granularity in boost::serialization "stream error"?
by eric 10 Jun '08

10 Jun '08

Hello, How can I get more detail when boost::serialization raises the exception "stream error"? I find that this one error can result from a wide variety of dissimilar root causes, but regardless of the nature of the problem (or the size of the buffer) the only information provided by boost::serialization is that one opaque string. It is easy to lose a lot of time tracking down the problem. I have worked with other more traditional XML parsers that raise exceptions in the form "Unexpected character in line x column y." I am aware that the architecture of boost::serialization is completely different to that of a traditional XML parser but the bottom line is the same, if I could get more granularity on this exception it would save a lot of time troubleshooting. Debugging seems not to be an option as my debugger (gdb) gets lost in boost header files and templates. I am left with trial and error, chopping up the stream into smaller pieces until I can isolate the invalid content. Any ideas? Thanks, Eric

1 0

Create boost_1_33_1 boost::archive::text_iarchive with an empty stream
by Gancho Manev 10 Jun '08

10 Jun '08

Hi, The following code does not throw archive_exception always: std::stringstream ss; boost::archive::text_iarchive ia(ss); The archive_exception is thrown always on VC++2005/Windows while on GCC3.2.2/Redhat8 and GCC3.4.6/CentOS a std::length_error: basic_string::resize is thrown sometimes. That means the constructor tries to resize some string with value > basic_string.max_size(). The GCC defines the max_size() to be only about a quarter of the max length the allocator can get. That explains why that exception happens to be raised on GCC. A search of the source code lead me to the overloaded methods 'load' of text_iarchive_impl<Archive>. They look suspicious because do not check the input stream state before using the string size just loaded. The local variable is not initialized nor loaded, so it contains some arbitrary value! In the method load(std::string&) that might cause std::length_error while in load(char *s) - even program crash. The code in the next Boost versions does not seem to be changed. Has anyone encountered such behavior? I've not found that bug reported. Thanks in advance, Gancho

2 2

Using boost::thread with windows events
by Sean Farrow 10 Jun '08

10 Jun '08

Hi: I am neding to use boost::thread with windows events, dos anyone know whether this is possible/whether there are any examples. Basically, I need to exit the thread when a variable becomes fales. Any help apreciated. Sean. __________ Information from ESET NOD32 Antivirus, version of virus signature database 3171 (20080610) __________ The message was checked by ESET NOD32 Antivirus. http://www.eset.com

1 0

yet another logging library
by Mojmir Svoboda 10 Jun '08

10 Jun '08

Hello, i'm trying to create a logging library framework. It is my second attempt, the first can be seen at http://www.codeplex.com/flog . There is some documentation in the doc directory that is perhaps worth reading. The first version used boost::iostreams and not very smart locking mechanism in mt environment (in terms of performance mainly), but served more as a verification 'it can be done and it is not that ugly as i thought' than a real library candidate. Now i'd like to focus on performance and code size, so I made a sketch of the next version. It is really small and is rather a model than anything usable, but before I start implementation, I'd like to know opinion of library writers and users - what do you think about the concept? is it worth continuing, or i'd better go doing something better? :) The sketch of the next version can be seen at http://www.codeplex.com/fflog It's meant to be illustration of the interface and it's usage itself. Short summary of the design: User specifies compile-time vector of requested filters that apply to the logged text. Each filter requiring user's input is tagged and when logging function is invoked, the supplied types are checked against expected type, set to filter and later written to a buffer. Buffers are managed by the sink_manager and likely is one of few places that has to be mt-safe, as the rest is on the stack. I thought about some lock-free queues of buffers that are assigned to caller and filled returned back to write. The actual write of the buffer can be performed in dedicated lower-priority thread. Run-time filtering like level or context is possible, but not implemented yet. The main idea can be seen in the 1st version in runtime_filtering.hpp. Again, runtime values are shared among threads, so they have to be protected in mt environment. Many thanks for your attention, Mojmir

6 13

parallel-iteration
by e r 10 Jun '08

10 Jun '08

Hi, suppose i have container_type<range_type> seqs; i'm looking for iterator, whose instance I call i, such that if(*i == seq [j][k] ){ if(j<seqs.size()){ ++i; *i == seq[j+1][k] }else{ ++i; *i == seq[0][k+1] } if anyone has this or anything similar, could he/she kindly share it? the above is not quite the same as zip_iterator, as the parallel ranges all have the same type. also i'm aware that multi_arrays deal with this, but here i'm interested in arbitrary type. thanks for reading this thread, e.

1 0

is there strided iterator already available somewhere?
by e r 10 Jun '08

10 Jun '08

hello, is there strided iterator already available somewhere? iter = vec.begin(); std::advance(iter,offset) for(unsigned int i=0; i<(vec.size()-offset)/stride; ++i){ /* do something with iter */ std::advance(iter,stride) } alternatively, a permutation iterator does not seem very practical here... thanks, e.

2 2

metafunction invocation performance
by Hicham Mouline 09 Jun '08

09 Jun '08

Hello, I submitted a post a few days ago http://lists.boost.org/boost-users/2008/06/36852.php and I went to investigate on my own. Having not reached a solution to speed up the compilation, I looked again at the post and realized the code didn't compile due to a wrong paste. Apologies (if anyone tried), Anyways, here is the version that is too slow and runs out of mem during compilation on a P4 512Mb minimal redhat box with g++4.1 while, on core2duo 2Gb vista64 with intel 10.1, it compiles in 40s and runtime is correct. The point is to generate all template "curve" specializations for various tmpl arguments (constant, linear....) (cross product of 3 mpl sequences interpolator_types, extrapolator_types and integrator_types , each containing the possible tmpl type arguments 1, 2 and 3), and insert the fully specialized templates curve<T1, T2, T3> in a mpl vector "curve_types" to be used by make_variant_over. Regards, -------------------- #include <iostream> #include <boost/mpl/pair.hpp> #include <boost/mpl/placeholders.hpp> #include <boost/mpl/fold.hpp> #include <boost/mpl/vector/vector0.hpp> #include <boost/mpl/vector.hpp> #include <boost/mpl/joint_view.hpp> #include <boost/mpl/transform.hpp> #include <boost/mpl/for_each.hpp> #include <boost/mpl/size.hpp> #include <boost/variant.hpp> class constant {}; class linear {}; class polynomial {}; class cubic_spline {}; class simpson {}; template <typename inter, typename extra, typename integ> class curve {}; template <typename inter, typename extra, typename integ> struct curve_metafunction { typedef curve<inter,extra,integ> type; }; template<typename T> struct pair_with : boost::mpl::pair<T,boost::mpl::_> {}; template<typename pair> struct triple_with : curve_metafunction<typename pair::first, typename pair::second, boost::mpl::_> {}; template<typename Sequence1,typename Sequence2> struct sequence_product : boost::mpl::fold< Sequence1, boost::mpl::vector0<>, boost::mpl::joint_view< boost::mpl::_1, boost::mpl::transform< Sequence2, pair_with<boost::mpl::_2> > > > {}; template<typename SequencePair, typename Sequence> struct seq_seqpair_product : boost::mpl::fold< SequencePair, boost::mpl::vector0<>, boost::mpl::joint_view< boost::mpl::_1, boost::mpl::transform< Sequence, triple_with<boost::mpl::_2> > > > {}; template<typename Sequence1,typename Sequence2,typename Sequence3> struct seq_product : seq_seqpair_product< sequence_product<Sequence1,Sequence2>::type, Sequence3 > {}; typedef boost::mpl::vector<constant, linear, polynomial, cubic_spline> interpolator_types; typedef interpolator_types extrapolator_types; typedef boost::mpl::vector<simpson> integrator_types; typedef seq_product<interpolator_types, extrapolator_types,integrator_types>::type curve_types; struct type_printer { template <typename T> void operator()( T ) { std::cout<< typeid(T).name() <<std::endl; } }; int main() { std::cout<< boost::mpl::size<curve_types>::type::value <<std::endl; boost::mpl::for_each<curve_types>( type_printer() ); }

4 7

Using boost 1.35.0 on Mac OS X with Xcode 3
by Phil Hystad 09 Jun '08

09 Jun '08

I am rebuilding a library that was last compiled a year ago using an earlier version of boost and an earlier version of Xcode (2.4) and an earlier version of Mac OS X (10.4). I downloaded boost_1_35_0 distro for my Mac and built it using the standard configuration and build procedures. That seemed to go OK. I am compiling C++ source code that formerly compiled, without errors or warnings, but now reports the following errors in many files: "error: 'snprintf' is not a member of 'std'" And, the offending line is the following from the c++locale.h file (the error is reported on the line referencing snprintf): #ifdef _GLIBCXX_USE_C99 const int __ret = std::snprintf(__out, __size, __fmt, __prec, __v); #else const int __ret = std::sprintf(__out, __fmt, __prec, __v); #endif Does anyone know the source of this problem. There is absolutely no changes in my source that is being compiled. I am dual posting this on both the boost-users and the Mac Xcode lists because I am not sure which is the more appropriate forum.

1 0

Re: [Boost-users] [iterators][range] size compile failure in 1.35
by Thorsten Ottosen 09 Jun '08

09 Jun '08

Jeff Flinn skrev: > David Abrahams wrote: >> in 1.35 I'm pretty sure that iterator_range<T>::size() only works when T >> is a random access iterator, and filtered iterators are by necessity >> bidirectional. I suggest you use >> >> std::distance(r.begin(), r.end()) >> >> if you're willing to pay for O(N) execution. >> > > Hmm, it looks like that's what iterator_range<T>::size() did in 1.34.1 > while 1.35 returns m_Begin - m_End. Any idea why this was done? boost::distance(r) could also work. I did it as part of the general clean-up of the range lib. size() is expected to be O(1) by users, distance() is not. -Thorsten

1 0

multi-threaded safe circular_buffer
by Selçuk Giray Özdamar 09 Jun '08

09 Jun '08

Hi everyone, I specialized the boost::circular_buffer template class for concurrent access. But I have no idea about how the locking strategy must be for copy ctor, and also assignment operator. I inserted question marks at the end of the related lines. Thanks in advance... template <class T> class CircularBuffer { typedef boost::circular_buffer<T> internal_buffer; typedef boost::shared_mutex ReadWriteMutex; typedef boost::shared_lock<boost::shared_mutex> ReadLock; typedef boost::unique_lock<boost::shared_mutex> WriteLock; public: // Constructors explicit CircularBuffer(capacity_type capacity = 10) : m_capacity(capacity) , m_buffer(capacity) { } // copy ctor CircularBuffer(const CircularBuffer& other) { //WriteLock w_lock(rw_mutex); ??????????? m_capacity = other.m_capacity; m_buffer = other.m_buffer; } size_type size() const { ReadLock r_lock(rw_mutex); return m_buffer.size(); } void push_back(param_value_type item = value_type()) { WriteLock w_lock(rw_mutex); m_buffer.push_back(item); } // assignment operator CircularBuffer& operator= (const CircularBuffer& other) { //WriteLock w_lock(rw_mutex); ?????????? if (this != &other) { m_capacity = other.m_capacity; m_buffer = other.m_buffer; } return *this; } private:// member variables capacity_type m_capacity; internal_buffer m_buffer; // rw_mutex for internal buffer mutable ReadWriteMutex rw_mutex; };

3 3

serialize shared_array
by Hansi 09 Jun '08

09 Jun '08

Hello, Is it possible to serialize a shared_array?How? Best regards Hansjörg

1 0

Converting seconds to minutes and seconds using boost::datetime
by Sean Farrow 09 Jun '08

09 Jun '08

Hi: Does boost::datetime allow me to convert seconds (such as 164) to minutes and seconds? If yes wha function does this. Sean. __________ Information from ESET NOD32 Antivirus, version of virus signature database 3167 (20080609) __________ The message was checked by ESET NOD32 Antivirus. http://www.eset.com

2 3

serialization, 1.34/1.35 difference: serializing a vector
by Filip Peters 08 Jun '08

08 Jun '08

I just tried serializing a simple vector, std::vector<unsigned int> vectortest; If I serialize out in 1.34, I can't read it in in 1.35. xml result from 1.35: <vectortest> <count>3</count> <item>1</item> <item>2</item> <item>3</item> </vectortest> from 1.34: <vectortest> <count>3</count> <item_version>0</item_version> <item>1</item> <item>2</item> <item>3</item> </vectortest> Notice the item_version, this causes an exception when reading it in with 1.35.... I need a way to get the behavior from 1.34 back.... Otherwise my files aren't readable anymore. small testprogram: #include <boost/config.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> //#include <boost/archive/binary_iarchive.hpp> //#include <boost/archive/binary_oarchive.hpp> #include <boost/serialization/vector.hpp> #include <boost/serialization/export.hpp> #include <fstream> #include <iomanip> int _tmain(int argc, _TCHAR* argv[]) { std::vector<unsigned int> testvector; testvector.push_back(1.); testvector.push_back(2.); testvector.push_back(3.); //this in 1.34 std::ofstream ofs( "testing2.txt", std::ios::binary ) ; boost::archive::xml_oarchive oa(ofs); oa << boost::serialization::make_nvp("vectortest", testvector); ofs.close(); //do this in 1.35 try{ std::ifstream ofs( "testing2.txt", std::ios::binary ) ; boost::archive::xml_iarchive ia(ofs); ia >> boost::serialization::make_nvp("vectortest", testvector2); ofs.close(); } catch(boost::archive::archive_exception testexception){ //reading a file from 1.34, will end up here std::string error = testexception.what(); } return 0; } ____________________________________________________________________________________ You rock. That's why Blockbuster's offering you one month of Blockbuster Total Access, No Cost. http://tc.deals.yahoo.com/tc/blockbuster/text5.com

5 6

[test/signals] trying test thread safe signals with the boost unit test framework
by Richard Dingwall 07 Jun '08

07 Jun '08

Hi guys, Using the boost unit test framework, I'm trying to test a simple scenario: when a person's name is set, their 'updated' signal is triggered. Here's my person class: class person { public: person() : name_() {}; void name(const std::string & name) { name_ = name; updated(*this); }; boost::signal<void(const person & person)> updated; private: std::string name_; }; To test this, I've created a mock handler for the updated event. In it's destructor, it checks has been called (and also with the correct subject): struct mock_handler { mock_handler(const person & subject) : has_been_called_(false), subject_(subject) {}; // Handler. void operator()(const person & subject) { BOOST_TEST_MESSAGE("in mock_handler::operator()"); has_been_called_ = true; BOOST_CHECK_EQUAL(&subject == &subject_, true); }; // This handler must be called before it goes out of scope. ~mock_handler() { BOOST_TEST_MESSAGE("in mock_handler::~mock_handler()"); BOOST_CHECK_EQUAL(has_been_called_, true); }; private: bool has_been_called_; const person & subject_; }; And then writing my test case to use it: BOOST_AUTO_TEST_CASE(setting_name_triggers_update_signal) { BOOST_TEST_MESSAGE("entered test"); person subject; mock_handler handler(subject); subject.updated.connect(handler); subject.name("Richard"); BOOST_TEST_MESSAGE("exiting test"); } Unfortunately, when this code runs, it produces the following output: Running 1 test case... entered test in mock_handler::~mock_handler() test.cpp(46): error in "setting_name_triggers_update_signal": check has_been_called_ == true failed [0 != 1] in mock_handler::~mock_handler() test.cpp(46): error in "setting_name_triggers_update_signal": check has_been_called_ == true failed [0 != 1] in mock_handler::~mock_handler() test.cpp(46): error in "setting_name_triggers_update_signal": check has_been_called_ == true failed [0 != 1] in mock_handler::operator() exiting test in mock_handler::~mock_handler() test.cpp(46): error in "setting_name_triggers_update_signal": check has_been_called_ == true failed [0 != 1] in mock_handler::~mock_handler() *** 4 failures detected in test suite "tests" I'm not sure what's going on here, or why mock_handler's destructor is being called 4 extra times. Can anyone shed some light as to why this might be happening? Cheers, Rich

2 2

Parse data and asign it to variables
by Hansi 07 Jun '08

07 Jun '08

Hello, I have a question. I need some parser with which I can parse plain binary data and then assign the data to e.g. an PropertyTree It should be possible to parse the data with a userdefined pattern. It would be also great if I can use the same thing to make some filter for traffic, with which I can assign the data to some structs...in this case it is for sure better if this parser is made during compiletime... Is there some library in boost available with which it is possible to solve this problem? I don't know may be it is possible to solve the problem with spirit... Thanks for any hint! Best regards Hansjörg

4 9

OS X: Script to compile various combinations of multi-arch Boost 1_35 for OS X 10.5
by Daniel Lord 07 Jun '08

07 Jun '08

All,I have written a shell script to build static and shared libs from various combinations or 32- and 64-bit i386, ppc, ppc64, and x86_64 architectures using bjam calls. It requires use of the 10.5 SDK while the default SDK for Boost on OS X is 10.4--which won't build shared 64-bit libs flawlessly. Many thanks to the Boost developers who gave me a enough hints on this list to figure it all out. If there is a wiki where I can post it, I can do that to make access easier. If someone wants the script in the mean time, let me know. Daniel -- "It's not how many years you live--it's how much life you put in your years" -- Sierra Mountain Guide

1 0

UTF-16 facets
by Johan Råde 07 Jun '08

07 Jun '08

Does Boost have any facets for reading UTF-16 files? I thought there were, but I could not find any. --Johan Råde

1 1

-D_GLIBCXX_DEBUG breaks boost.program_options
by pete＠mu.org 07 Jun '08

07 Jun '08

I've been trying to play with -D_GLIBCXX_DEBUG, but I'm having some problms with boost.program_options. Here's the code and the error. It's on an Ubuntu 8.04 box with g++4.2. #include <boost/program_options.hpp> using namespace boost; using namespace boost::program_options; int main(int ac, char* av[]) { options_description desc("Allowed options"); desc.add_options() ("help,h", "produce a help screen") ; variables_map vm; store(parse_command_line(ac, av, desc), vm); } $ make g++ -g -Wall -ggdb -ansi -I/home/pete/include -DDEBUG -D_GLIBCXX_DEBUG -L/home/pete/lib -lboost_program_options-mt test.cpp -o test /home/pete/tmp/ccJZlQNM.o: In function `basic_command_line_parser': /home/pete/include/boost/program_options/detail/parsers.hpp:41: undefined reference to `boost::program_options::detail::cmdline::cmdline(std::__debug::vector <std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > const&)' collect2: ld returned 1 exit status make: *** [test] Error 1 If I remove -D_GLIBCXX_DEBUG, the problem goes away. Any ideas of how to get this fixed, or am I doing something wrong? Thanks.

1 0

libboost_filesystem-vc90-mt-gd-1_35.lib - Visual C++ 2008
by Fox Fossati 06 Jun '08

06 Jun '08

Excuse me, I'm trying to use the library boost/filesystem but the file libboost_filesystem-vc90-mt-gd-1_35.lib isn't in my directory boost and there is a linker error from the Visual C++ 2008. What can i do for solve the problem? Mr Rainer wrote me: "build the library"..... Can someone explain me this solution? Regards. ___________________________________ Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e la tua opinione! http://www.ymailblogit.com/blog/

2 1

Convert local string to utf-8
by Li, PingShan (Kansas City) 06 Jun '08

06 Jun '08

Scott, Based on your suggestion. I checked http://en.wikipedia.org/wiki/Iconv and http://www.icu-project.org/userguide/codepageConverters.html. I need to support windows and Linux and ICU converter turned out to be a better fit for me. I used ICU before related to Xerces and Xalan library. Just some update on this one. Thank you for your help. On Tue, Jun 3, 2008 at 6:04 PM, Li, PingShan (Kansas City) <pingshan.li_at_[hidden]> wrote: > I need to convert a big5 encoded string to utf-8 format. Is there anything > in boost I can use for that? Or any suggestions? > Boost.Regex uses ICU[1] for Unicode support, so that's the most related thing I can think of. [1] http://www.icu-project.org/

1 0

Client Authentication not happening in asio/ssl
by Arun Prabu 06 Jun '08

06 Jun '08

Hi, I am using the following boost asio example program to create a ssl connection - http://www.boost.org/doc/libs/1_35_0/doc/html/boost_asio/example/ssl/client… - http://www.boost.org/doc/libs/1_35_0/doc/html/boost_asio/example/ssl/server… I am able to create a ssl connection, send/receive packets through the ssl connection. This code is suppose to do one way handshaking only. i.e only server authentication is done. The server sends its certificate to Client and the Client authenticates it. This is the code on server side which adds its certificate into the context. Server: context_.set_options( asio::ssl::context::default_workarounds | asio::ssl::context::no_sslv2 | asio::ssl::context::single_dh_use); context_.set_password_callback(boost::bind(&server::get_password, this)); context_.use_certificate_chain_file("server.pem"); context_.use_private_key_file("server.pem", asio::ssl::context::pem); context_.use_tmp_dh_file("dh512.pem"); This is the code on the client side which initiates the Server Authentication Client : ctx.set_verify_mode(boost::asio::ssl::context::verify_peer); ctx.load_verify_file("ca.pem"); But, I wanted to do Client side authentication too. So, i added the following on server side Server: ctx.set_verify_mode(boost::asio::ssl::context::verify_fail_if_no_peer_cert | boost::asio::ssl::context::verify_peer); ctx.load_verify_file("ca.pem"); I added the following on client side to upload its certificates to the context. Client: context_.set_options( asio::ssl::context::default_workarounds | asio::ssl::context::no_sslv2 | asio::ssl::context::single_dh_use); context_.set_password_callback(boost::bind(&server::get_password, this)); context_.use_certificate_chain_file("client.pem"); context_.use_private_key_file("client.pem", asio::ssl::context::pem); context_.use_tmp_dh_file("dh512.pem"); After adding these code, the Client Authentication is not happening. But, Openssl man page says that setting verify_fail_if_no_peer_cert should initiate Client Authentication. I am missing out something. Can someone over here help me out how to do both server/client authentication over an ssl connection using boost asio. Regards, Arun

1 0

Re: [Boost-users] [iterators][range] size compile failure in 1.35
by David Abrahams 06 Jun '08

06 Jun '08

on Fri Jun 06 2008, Jeff Flinn <TriumphSprint2000-AT-hotmail.com> wrote: > I'm moving from 1.34.1 to 1.35.0 > > The following compiles in 1.34.1 and fails in 1.35.0 on MSVC 8. Any > thoughts on how to correct this? Jeff, in 1.35 I'm pretty sure that iterator_range<T>::size() only works when T is a random access iterator, and filtered iterators are by necessity bidirectional. I suggest you use std::distance(r.begin(), r.end()) if you're willing to pay for O(N) execution. -- Dave Abrahams BoostPro Computing http://www.boostpro.com

1 0

a boost interprocess question
by Alex Markelov 06 Jun '08

06 Jun '08

Hello, several processes need to communicate using managed_windows_shared_memory object:: - "servers": managed_windows_shared_memory shm (open_or_create, "MySharedMemory", shm_size); - "clients": managed_windows_shared_memory shm (open_only, "MySharedMemory"); "Client" and "server" proceses run sequence may be the following: 1) start server 1 2) start client 1 3) exit server 1 4) start client 2 -> oops! interprocess exception thrown: shared memory could not be open. ... Why at the moment when last "server" exited the shared memory was destroyed? Shared memory should exist while at least one user exists. At the moment when the last "server" exited, client 1 was existing and thus the shared memory should be accessible for opening still. Regards, Alex

2 1

[serialization] portable serialization archive
by Chris Weed 06 Jun '08

06 Jun '08

Hi, I have seen several mentions on the mailing list of a portable binary serialization archive, what is the place to get the latest and greatest version? Thanks, Chris

6 13

a problem with interprocess::managed_windows_shared_memory
by Alex 06 Jun '08

06 Jun '08

Hello, in my program a shared memory object is created by: managed_windows_shared_memory shm (open_or_create, "MySharedMemory", shm_size); The problem is that calling to 'shm' object methods, the program cannot detect - was shared memory actually created or it was opened only? Looks like somewhere inside 'managed_windows_shared_memory' class this information is availble but not exposed via its public interface. May it be fixed? Regards, Alex

1 0

[Boost.StringAlgo] How to insert substr into str
by Valentine Rozental 06 Jun '08

06 Jun '08

Dear All, what is the elegant/right way to insert some substr into the given str? For example: // ------------------------------------------------------- ... std::string str("xyz_123_abc_xyyz_a1b"); std::string mark("abc"); std::string substr("$$$"); iterator_range<range_iterator<std::string>::type> it1 = boost::find_first(str, mark); ... // ------------------------------------------------------- I want to insert substr ("$$$") after mark ("abc") if the mark exists in the str. That is, the result in the example above should be: "xyz_123_abc$$$_xyyz_a1b" Thank you in advance!

4 3

How to simplify repetetive constructor calls?
by Zeljko Vrba 06 Jun '08

06 Jun '08

I have a following situation: class X { type1 v1; type2 v2; ... typen vn; other_var ... public: X() : v1(this), v2(this), .., vn(this), other_var(1, 2, 3) { } }; I'm looking for suggestions on how to automatically initialize variables v1..vn with "this" being passed as the only argument to their constructor. In the ideal case, the programmer should not need to specify any constructor calls for v1..vn, so that the end result looks something like: class X { type1 v1; ... typen vn; other_var; public: X() : other_var(1, 2, 3) { // some "magic" arranges that v1(this) .. vn(this) is called } }; Thanks!

3 3

resize() in boost/multi_array.hpp
by Eric Darve 05 Jun '08

05 Jun '08

Hi, I found what I think is a bug in boost/multi_array.hpp. Perhaps some of you can explain what is wrong. Look at the code below. When executed you get the following error: multi_array_ref.hpp:487: failed assertion `std::equal(other.shape(),other.shape()+this->num_dimensions(), this->shape())' This is caused by the line x.resize(extents[n+1][3]); It seems to me that the code is correct and should work properly. Curiously if one adds x[i][j].resize(extents[0][0]); before x.resize(extents[n+1][3]), the error message disappears. I also added another piece of code at the end showing that if one replaces multi_array< multi_array<int,2>, 2> x; by multi_array< vector<int>, 2> y; then the resize seems to work. The code was run on MacOSX with g++. Eric #include "boost/multi_array.hpp" #include <vector> using boost::multi_array; using boost::extents; using std::vector; int main(void) { const int n = 10; multi_array< multi_array<int,2>, 2> x; x.resize(extents[n][3]); for (int i=0;i<n;++i) { for (int j=0;j<3;++j) { x[i][j].resize(extents[i+1][2]); // Size varies with i } } for (int i=0;i<n;++i) { for (int j=0;j<3;++j) { // x[i][j].resize(extents[0][0]); /* failed assertion below disappears if line above is un-commented */ } } x.resize(extents[n+1][3]); /* Failed assertion when executing line above: multi_array_ref.hpp:487: failed assertion `std::equal(other.shape(),other.shape()+this->num_dimensions(), this->shape())' */ // Attempting a similar operation using vector<int> this time multi_array< vector<int>, 2> y; y.resize(extents[n][3]); for (int i=0;i<n;++i) { for (int j=0;j<3;++j) { y[i][j].resize(i+1); } } y.resize(extents[n+1][3]); /* This instruction executes fine */ return 0; }

2 1

build times and Boost.MultiIndex
by gast128 05 Jun '08

05 Jun '08

Dear all, just refactored a component in our (large) application with use of Boost.MultiIndex. However using a test application it seems that by only including the Boost.MultiIndex header in the header files, the build time is 7 times (700%) slower than without that header file. Pulling Boost.MultiIndex to the cpp file is no option, because I need to export the iterator as well. Of course I can make a wrapper around the iterator as well (for the sole purpose to shield the inclusion of the Boost.MultiIndex header), but isn't this a little bit strange? Putting Boost.MultiIndex in the pch is also a no-op because, 1) every pch must be adjusted, 2) VC++ 7.1 chokes on that. It is not something Boost can do much about it, but it as least an irritating point to put it mildly. It greatly cost productive hours to wait for builds to finish. Something for the c++ committee? c# seems to have a superior build system... wkr, me

4 6

Reader/Writer Mutex
by Selçuk Giray Özdamar 05 Jun '08

05 Jun '08

Hi everyone, Below you can see my simple thread safe class which allows multiple readers and exclusive write for single writer. Is this boost implementation correct? Also if anyone has other aspects, I'm open. Regards... typedef boost::shared_mutex ReadWriteMutex; typedef boost::shared_lock<boost::shared_mutex> ReadLock; typedef boost::unique_lock<boost::shared_mutex> WriteLock; class MtClass { public: MtClass() : m_value() { } ~MtClass(){} void setValue(int value) { WriteLock w_lock(rw_mutex); m_value = value; } int getValue(void) const { ReadLock r_lock(rw_mutex); return m_value; } private: int m_value; mutable ReadWriteMutex rw_mutex; };

2 1

Where is Math Toolkit in the Boost Vault?
by John Maddock 05 Jun '08

05 Jun '08

Boris wrote: > Can anyone provide me a link to download the Math Toolkit? If I > remember correctly it could be found in the Boost Vault before (but I > can't find it anymore). I know that it's shipped now with Boost > 1.36.0 but I can't upgrade to the latest version. The version in the vault was outdated so I removed it once it shipped as a released Boost version. What I suggest you do is grab a copy of boost/math and libs/math from the 1.36 release and use those. I'm fairly sure (!) that should work OK with 1.35.1. HTH, John.

1 0

boost documentation in printable format
by e r 05 Jun '08

05 Jun '08

Hi, I'd like to print an entire boost library documentation. Is there a convenient way to do it? Thanks, E.

2 1

Using Boost.Test in a dynamically loaded library
by Florian Winter 05 Jun '08

05 Jun '08

We are trying to write units tests using the Boost.test framework which involve dynamic loading of libraries (i.e. lt_dlopen on Linux or LoadLibrary on Windows) and doing tests using the Boost test tools in code inside the dynamically loaded library. Does Boost.Test support this? We have run into problems when trying to use the quick-and-dirty "include" option for defining the Boost.Test symbols. Dynamically loaded libraries may not contain undefined symbols, and including the implementation headers such as <boost/test/included/unit_test.hpp> in both the dynamically loaded library and the main application seems to lead to duplicate symbols. In particular, if we use BOOST_CHECK macros in both the main application, everything works fine. However, if we use BOOST_CHECK macros in the dynamically loaded library, we get an exception with the following comment: can't use testing tools before framework is initialized We suspect that this is caused by Boost.Test global symbols being defined multiple times - once in the application and once in the dynamic library. We would like to know what is the "best" practice to use Boost.Test in unit tests that require dynamic loading of libraries and tests being made in both application code and dynamic library code. It must work on Linux (using libraries created with libtool and libltdl for dynamic loading) and Windows (using DLLs and LoadLibrary for dynamic loading). Is this even possible and has it been tested before? Would using a static or shared library version of Boost.Test solve our problem? Are there any other potential pitfalls that we should be aware of? Thanks in advance for your help. Best Regards, Florian Winter, Motama GmbH -- Florian Winter Software-Entwickler

2 2

[serialization] reverse bytes in portable_binary_oarchive
by Chris Weed 04 Jun '08

04 Jun '08

I am a little confused by the section of code in the ifdef of portable_binary_oarchive 48 #ifdef BOOST_BIG_ENDIAN 49 cptr += (sizeof(boost::intmax_t) - size); 50 if(m_flags & endian_big) 51 reverse_bytes(size, cptr); 52 #else 53 if(m_flags & endian_big) 54 reverse_bytes(size, cptr); 55 #endif I would expect line 50 to be if(m_flags & endian_little) Chris

1 0

Re: [Boost-users] Boost-users Digest, Vol 1653, Issue 4
by Fox Fossati 04 Jun '08

04 Jun '08

Message: 3 Date: Wed, 04 Jun 2008 17:14:19 +0200 From: Rainer Thaden <RThaden(a)web.de> Subject: Re: [Boost-users] Visual C++ 2008- boost/filesystem - linker error To: boost-users(a)lists.boost.org Message-ID: <4846B14B.50802(a)web.de> Content-Type: text/plain; charset=UTF-8; format=flowed > Excuse me, > I'm trying to use the library boost/filesystem. > I wrote: > > #include "boost/filesystem/operations.hpp" > #include "boost/filesystem/path.hpp" > #include <iostream> > > or > > #include "boost/filesystem.hpp" > #include <iostream> > > but in both cases this error occurs: > > LINK : fatal error LNK1104: cannot open file > 'libboost_filesystem-vc90-mt-gd-1_35.lib' > > How can I solve this problem? Locate libboost_filesystem-vc90-mt-gd-1_35.lib. If not there: build the library. If there: make sure that the containing directory is present in the library path of the linker. If not: Open the properties of your project, go to linker and add the path. Regards, Rainer The file isn't in the boost directory.... Wich library i have to build? I have to donwload another file from boost.org? Can you explane me..... I'm a beginner.... :-( Thanks! ___________________________________ Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e la tua opinione! http://www.ymailblogit.com/blog/

1 0

Proto transform using visitor
by Joel Falcou 04 Jun '08

04 Jun '08

Hello, I'm currently trying to grasp the way proto trasnform are supposed to work and I was wondering fio there is any use cases or examples of how and why a Visitor can be used. Thanks

3 8

breadth_first_visit
by Trevor Bain 04 Jun '08

04 Jun '08

I have a Graph here: ********************************* //! value (i.e. voltage) of vertex struct value_t { typedef boost::vertex_property_tag kind; }; typedef boost::property<value_t, float> Value; /*! Defining the graph type 'digraph' * * The listS and vecS types are selectors that determine the data * structure used inside the adjacency_list (listS has faster time * complexity whereas vecS has lower space complexity). * * The directedS type specifies that the graph should be singly * directed and (versus undirected, or bidirectionalS). */ typedef boost::adjacency_list< boost::vecS, // EdgeList boost::vecS, // VertexList boost::directedS, Value, boost::property<boost::edge_weight_t, float> > Digraph; Digraph net; int N; //! Edge description typedef std::pair<int, int> Edge; //! Vertex description typedef Digraph::vertex_descriptor Vertex; //! Property map to access value (voltage) boost::property_map<Digraph, value_t>::type value; //! Structure for creating BFSvisitor struct rand_val : public boost::base_visitor<rand_val> { typedef boost::on_discover_vertex event_filter; template <class Vertex, class Digraph> void operator()(Vertex v, Digraph& G) { boost::put(value, v, normal_sampler()); } }; ********************************* and I try to implement it here (in the Poly.cpp file) ********************************* void Poly::propagate() { // The source vertex Vertex s = *(boost::vertices(net).first); // BFS needs to "color" the vertices. // using std::vector as exterior property storage. std::vector<boost::default_color_type> colors(N); // Breadth First Visit with appropiate visitor boost::breadth_first_visit( net, s, make_bfs_visitor(rand_val()), colors.begin() ); // i think either 'net' or 's' is broken (not correct models) } ********************************* when I compile I get the error: Poly.cpp: In member function 'void Poly::propagate()': Poly.cpp:61: error: no matching function for call to 'breadth_first_visit(boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS, boost::property<Poly::value_t, float, boost::no_property>, boost::property<boost::edge_weight_t, float, boost::no_property>, boost::no_property, boost::listS>&, size_t&, boost::bfs_visitor<Poly::rand_val>, __gnu_cxx::__normal_iterator<boost::default_color_type*, std::vector<boost::default_color_type, std::allocator<boost::default_color_type> > >)' make: *** [Poly.o] Error 1 which means my arguments don't match the function arguments . . . but I've been problem solving this for a while and I can't figure out which ones don't match? I'm thinking it must be my graph but I meet all of the stated requirements for the graph so I'm confused? Any help would be greatly appreciated. Here are the source files: http://svn.reducis.org/polyneuron/src/Poly.cpp http://svn.reducis.org/polyneuron/src/Poly.hpp - Trevor

1 0

[ptr_container] unordered containers added to trunk
by Thorsten Ottosen 04 Jun '08

04 Jun '08

Dear all, I have wrapped Daniel James' unordered containers s.t. we now have ptr_unordered_set ptr_unordered_multiset ptr_unordered_map ptr_unordered_multimap It would be great if someone would try to use the version found in trunk. There's no new docs yet, but if you know the old ptr_map/ptr_set, it should be straightforward to use them. Any comments are most welcome. best regards -Thorsten

1 0

[statechart] usage in realtime application
by Marc Ruiter 04 Jun '08

04 Jun '08

Hi, I'm trying to convert a custom statemachine implementation in a real-time application to Boost.Statechart. The FAQ mentions that custom allocators have to be provided for Event, StateMachine etc. in order to get deterministic timing behaviour. I'm currently working on an allocator which uses an internal preallocated memory pool. Now my question is whether it is possible to determine how much memory i need to preallocate for the States and Events. Regards, Marc

2 3

Visual C++ 2008- boost/filesystem - linker error
by Fox Fossati 04 Jun '08

04 Jun '08

Excuse me, I'm trying to use the library boost/filesystem. I wrote: #include "boost/filesystem/operations.hpp" #include "boost/filesystem/path.hpp" #include <iostream> or #include "boost/filesystem.hpp" #include <iostream> but in both cases this error occurs: LINK : fatal error LNK1104: cannot open file 'libboost_filesystem-vc90-mt-gd-1_35.lib' How can I solve this problem? Thanks and best regards. ________________________________ Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e la tua opinione! ___________________________________ Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e la tua opinione! http://www.ymailblogit.com/blog/

2 1

Re: [Boost-users] [boost] BOOST documentation in PDF/PS format
by Jason Sachs 04 Jun '08

04 Jun '08

dhruva wrote: > Hello, > I am looking for the BOOST library in either PDF/PS formats so that I > can print the ones I need and read them offline (which I generally > prefer). Reading though the boost documenting link, it appears quite a > daunting task to setup boost doc build to generate PDF files (it has a > bunch of dependencies). > Is there any location from where I can download the PDF files for the > current release version of BOOST or any plans for providing the same > in the near future? Me too. I don't even mind if they are HTML files, as long as it's one HTML page per library and it prints OK. The sourceforge link ( https://sourceforge.net/project/showfiles.php?group_id=7586&package_id=1597… ) is somewhat limited in scope, only a handful of libraries, and it is 2 versions out of date (1.33 vs 1.35). Any other suggestions? I read somewhere about using htmldoc but even that is too daunting at present.

1 0

BGL problem wit prims
by Line Blander Reinhardt 04 Jun '08

04 Jun '08

The following: vector<standard::UnDirectedVertex> parent(num_vertices(graph_)); prim_minimum_spanning_tree(graph_,&parent[0]); gives me the minimum spanning tree using the boost::edge_weight property however I would like to use another property as I use the boost::edge_weight property for other things. I tried as described here but have some problems see error messages. I believe it is because of the parameter boost::default_dijkstra_visitor() however I do not know how to fix it: vector<standard::UnDirectedVertex> parent(num_vertices(graph_)); vector<std::size_t> distance(num_vertices(graph_)); boost::property_map<standard::Graph, mst_cost_t>::type weightmap = boost::get(mst_cost, graph_); boost::property_map<standard::Graph, boost::vertex_index_t>::type indexmap = get(boost::vertex_index, graph_); prim_minimum_spanning_tree(graph_,0, &parent[0], &distance[0],weightmap,indexmap,boost::default_dijkstra_visitor()); The error message is: 1>c:\program files\microsoft visual studio 8\vc\include\limits(109) : error C2440: '<function-style-cast>' : cannot convert from 'int' to 'W' 1> No constructor could take the source type, or constructor overload resolution was ambiguous 1> c:\program files\microsoft visual studio 8\vc\include\limits(107) : while compiling class template member function 'W std::numeric_limits<_Ty>::max(void) throw()' 1> with 1> [ 1> _Ty=W 1> ] 1> c:\boost\include\boost-1_33_1\boost\graph\prim_minimum_spanning_tree.hpp (61) : see reference to class template instantiation 'std::numeric_limits<_Ty>' being compiled 1> with 1> [ 1> _Ty=W 1> ] It would be nice if someone could help. Thanks. Line

1 0

Preprocessor error in mpl/apply.hpp with g++ 4.4.0 20080604
by Samuel Tardieu 04 Jun '08

04 Jun '08

With Boost 1.35.0 packages from Debian unstable, I can no longer use <boost/tokenizer.hpp> with today GCC trunk version. The shortest way to reproduce the problem is: % echo '#include <boost/mpl/apply.hpp>' > w.cc % g++ -E w.cc > /dev/null In file included from /usr/include/boost/mpl/apply.hpp:23, from w.cc:1: /usr/include/boost/mpl/apply_wrap.hpp:81:31: error: missing binary operator before token "(" /usr/include/boost/mpl/apply_wrap.hpp:173:31: error: missing binary operator before token "(" In file included from /usr/include/boost/mpl/bind.hpp:27, from /usr/include/boost/mpl/lambda.hpp:18, from /usr/include/boost/mpl/apply.hpp:25, from w.cc:1: /usr/include/boost/mpl/apply_wrap.hpp:81:31: error: missing binary operator before token "(" /usr/include/boost/mpl/apply_wrap.hpp:173:31: error: missing binary operator before token "(" In file included from /usr/include/boost/mpl/lambda.hpp:18, from /usr/include/boost/mpl/apply.hpp:25, from w.cc:1: /usr/include/boost/mpl/bind.hpp:364:31: error: missing binary operator before token "(" /usr/include/boost/mpl/bind.hpp:531:31: error: missing binary operator before token "(" In file included from /usr/include/boost/mpl/lambda.hpp:22, from /usr/include/boost/mpl/apply.hpp:25, from w.cc:1: /usr/include/boost/mpl/aux_/full_lambda.hpp:230:31: error: missing binary operator before token "(" In file included from w.cc:1: /usr/include/boost/mpl/apply.hpp:138:31: error: missing binary operator before token "(" It worked fine a few weeks ago, I do not know whether this is a boost bug uncovered by recent GCC changes or if this is a GCC bug. Sam

1 0

Compiling boost libraries with stlport and msvc8.0
by Sean Zlatnik 04 Jun '08

04 Jun '08

> > Hi! > > I think between my web searches and experimentation/debugging, I've spent > in excess 10-12 hours trying to solve this problem, so any help that can be > given would be great. > > Here is the relevant section of my user-config.jam file: > > # MSVC configuration > # Configure msvc (default version, searched in standard location > # and PATH). > using msvc : 8.0 : "C:/Program Files/Microsoft Visual Studio > 8/VC/bin/cl.exe" : stdlib=stlport ; > # Borland configuration > # using borland ; > > # STLPort configuration > # Configure, specifying location of STLPort headers. > # Libraries must be either not needed, or available to > # the compiler by default > using stlport : 5.1.5 : C:/STLport_5_1_5/stlport C:/STLport_5_1_5/lib > C:/STLport_5_1_5/bin ; > > I've tried all kinds of things her, including without anything in the msvc > section > > I've tried all kinds of command lines; the basic idea I've been using is > bjam --toolset=msvc --stdlib=stlport --build-type=complete. > > I don't know what the exact line should be or what to change my > user-config.jam file to, I've looked at the documentation and scoured the > web for information, and for the life of me, I can't get the files to show > up with the p in them indicating stlport was used. Any help you can give > would be greatly appreciated. > > Sincerely, > Sean >

1 0

boost & c-style calback issue
by Michael Bradley Jr 04 Jun '08

04 Jun '08

Hi, i'm facing this problem may be someone on the list do have an elegant solution class A; class B; class C { int operator () (A*,B*) }; typedef int (*foobar)(A*,B*); void do_something(foobar cb); C obj; do_something(obj); *ERROR* did try boost::function<int(A*,B*)> ftor = obj; do_something(ftor); *ERROR* compiler generates in both cases the following error messages cannot convert parameter 2 from 'boost::function<Signature>' to 'int (__cdecl *)(A *,B *)' with 1> [ 1> Signature=int (A *, B *) 1> ] 1> No user-defined-conversion operator available that can perform this conversion, or the operator cannot be called I can't make the corresponding class B member function static due to side-effects. Any clue how to get rid of this? Thanks in advance Mike

5 8

program_options config_file questions
by Chris Ross 04 Jun '08

04 Jun '08

I'm trying to load a bunch of information out of a file, and I see that with boost::program_options 1.35.0, there's a way to pass a parameter to parse_config_file() to tell it to read in unknown "options". I am doing this, but with a config file as such: [tale] name = bob name = Jerry value = 7 I get back three options, each of which is a vector<> with only one value. Is there a way to get program_options to return tale.name as a single option with two values in the vector? I would assume so, otherwise it seems odd to have it implemented as a vector. I'm sure I must've missed something somewhere to enable that behavior, but would appreciate any pointers. Thanks. - Chris

2 1

Convert local string to utf-8
by Li, PingShan (Kansas City) 03 Jun '08

03 Jun '08

I need to convert a big5 encoded string to utf-8 format. Is there anything in boost I can use for that? Or any suggestions? Thank you PingShan Li

2 1

Where is Math Toolkit in the Boost Vault?
by Boris 03 Jun '08

03 Jun '08

Can anyone provide me a link to download the Math Toolkit? If I remember correctly it could be found in the Boost Vault before (but I can't find it anymore). I know that it's shipped now with Boost 1.36.0 but I can't upgrade to the latest version. Boris

1 0

[xpressive] compilation woes with gcc 3.3.3
by Christoph Duelli 03 Jun '08

03 Jun '08

compiling #include <boost/xpressive/regex_actions.hpp> fails with gcc version 3.3.3 (SuSE Linux) (running SuSE LES9) Any chance that might be fixed? Anything I can do about it?

4 8

[boost-users][fusion] nested map access
by Istvan Buki 03 Jun '08

03 Jun '08

Hello, I'm trying out some ideas with boost.fusion maps. Here is a small example: namespace fields { struct general { struct first_name ; struct last_name ; struct age ; } ; struct address { struct street ; struct postal_code ; struct country ; } ; } typedef fusion::map < fusion::pair< fields::general::first_name, std::string >, fusion::pair< fields::general::last_name, std::string >, fusion::pair< fields::general::age, int > > general_person ; typedef fusion::map < fusion::pair< fields::address::street, std::string >, fusion::pair< fields::address::postal_code, std::string >, fusion::pair< fields::address::country, std::string > > address_person ; typedef fusion::map < fusion::pair< fields::general, general_person >, fusion::pair< fields::address, address_person > > person ; int main() { using namespace fields ; general_person gp = fusion::make_map< general::first_name, general::last_name, general::age >( "John", "Doe", 99 ) ; address_person ap = fusion::make_map< address::street, address::postal_code, address::country >( "Foo street 1", "12345", "Barland" ) ; person a_person = fusion::make_map< general, address >( gp, ap ) ; std::cout << "Age: " << fusion::at_key< general::age >( fusion::at_key< general >( a_person ) ) << std::endl ; return 0 ; } I'm not happy with the last line that retrieve the value of a specific pair inside a nested map. I was wondering whether this could benefit from some MPL trick to make it more readable. I was thinking about something like the following: get_value< general, age >( a_person ). My questions are: 1) Is MPL the right tool to transform a call to get_value< general, age >( a_person ) into fusion::at_key< general::age >( fusion::at_key< general >( a_person ) ) 2) If it is, could someone provide me some help on how to achieve such a result? Thank you for your help

2 4

interprocess object in memory
by Jan Suchý 03 Jun '08

03 Jun '08

Hallo boost interprocess useres, first of all, many thanks to Ion for his support: std::auto_ptr<managed_shared_memory> segment_ptr is working well and choosing keytype as interprocess::basic_string is working great too. Thank you. But now, I have stoppend when using map in shared memory in combination: typedef std::pair<MyString, XQQuery*> ValueType; where XQQuery* is pointer to object from xqilla X-Query engine: xqilla.sourceforge.net When I'am creating shared map in one process, I can see the content of the map in second process too. process one: mymap->size: 10 first: xqr_full/MW_APS_CPS_APSCreditScoringResult/REQ_BE_CPS_APSCreditScoringResult.xqr, second:0x85665a8 first: xqr_full/MW_APS_CPS_APSCreditScoringResult/RES_BE_CPS_APSCreditScoringResult.xqr, second:0x8c5ba10 first: xqr_full/MW_AccountCreate/REQ_SB_AccountCreate.xqr, second:0x80a5c08 first: xqr_full/MW_AccountCreate/RES_SB_AccountCreate.xqr, second:0x80b6000 first: xqr_full/MW_LoanCreate/REQ_SB_LoanAccountCreate.xqr, second:0x88a9c78 first: xqr_full/MW_LoanCreate/RES_SB_LoanAccountCreate.xqr, second:0x8dfe298 first: xqr_full/MW_ProductFeeGet/REQ_DON_ProductFee.xqr, second:0x8e0eb90 first: xqr_full/MW_ProductFeeGet/RES_DON_ProductFee.xqr, second:0x9620ba8 first: xqr_full/SYM_AccountGetBalance/REQ_MW_AccountBalance.xqr, second:0x97adca8 first: xqr_full/SYM_AccountGetBalance/RES_MW_AccountBalance.xqr, second:0x998ffa8 process two shows the same list.. First value is key basic_string, second is address of XQQuery object. The problem is, that the XQQuery object is ready to use only in process one, not in process two. I think, that it could be becouse of another memory adddressing between these two processes. So the second process can not see the XQQuery* objects on saved adresses. I would like to solve this problem, but I do not exactly know how to do it. I have found some hint about offse_ptr on address: http://igaztanaga.drivehq.com/interprocess/interprocess/offset_ptr.html but I do not now, if using offset_ptr will solve this problem. If yes, It will be perfect. Thank you for any information or example :-) Have a nice day, Jan Suchy

2 1

[ptr_container] No insert(it, [key, ]value) in associative pointer container?
by Rodolfo Lima 03 Jun '08

03 Jun '08

Hi, I'd like to know why there's no method 'insert' that mimicks stl's associative container's insert method that accepts an iterator that points to where the value could be placed, and the value? Regards, rod

2 4

boost.asio with boost.circular_buffer possible?
by james.vanns＠framestore.com 03 Jun '08

03 Jun '08

Hi, I was wondering if anyone has any experience or knows off the top of their head if it is possible (or even sensible) to use the boost.circular_buffer class as the underlying buffer for boost.asio socket reads and writes? I.e. boost::circular_buffer<char> cb (1024); bytes_transferred = sock.read (boost::asio::buffer (cb)); Similarly with write(). Regards, Jim -- James Vanns Systems Programmer Framestore CFC Ltd.

1 0

boost serialization & C++ STL
by sam p 03 Jun '08

03 Jun '08

Is it right to use serialization library in the following environment, ********************** Main1.cpp : Compiled with Gcc stores map in /tmp/z ********************** (1) main1.cpp -------------------- #include <map> #include<string> #include <boost/archive/text_oarchive.hpp> #include <boost/serialization/map.hpp> #include <boost/serialization/string.hpp> int main() { std::ofstream f("/tmp/z"); boost::archive::text_oarchive oar(f); std::map<std::string, std::string> a; ..... //Some data added to this map oar & a; return 0; } ************************** main2.cpp : Compiled with Forte compiler (or say older version of GCC) loads map stored in /tmp/z ******************************* (2) main2.cpp ----------------------------------------------- #include <map> #include<string> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/map.hpp> int main() { std::ifstream f("/tmp/z"); boost::archive::text_iarchive iar(f); std::map<std::string, std::string> a; iar & a; return 0; } *********************************** Can such use create problems especially with STL and their vendor specific implementations? Thanks, Sameer

3 2

program_options to load data from a configuration file?
by Chris Ross 03 Jun '08

03 Jun '08

I want to put a bunch of configuration into a file, and was wondering if boost::program_options could be used to load it. Looking at the documentation for program_options, it's easy to load configuration off of the command line, and then other values for same and additional configuration parameters from a config file. But, what I want to do is have a data file. I want to have a file that looks like, say: [thing1] name=dogooder method=http [thing2] name=bob method=http [thing4] name=mary method=gopher And, I'm wondering if I can use program_options::parse_config_file to load this, or if there's no way for parse_config_file to load things that aren't specifically entered in to the options_decription as expected options... Can it load unexpected options? Is there an iterator that would let me just look at everything that was loaded? Thanks... - Chris

1 0

[boost-users][statechart] processing an event by few orthogonal regions
by Igor R 03 Jun '08

03 Jun '08

Hi, I'd like several orthogonal regions to process some event. However, it seems (according to the manual and to what I see) that after one of the regions processes the event, it's discarded - isn't it? The only trick I could think about is to requeue the event in all the regions' reactions but one. But probably there is some more elegant way to do this? Thanks, Igor'.

2 1

[build] setting alignment (very urgent)
by Ovanes Markarian 03 Jun '08

03 Jun '08

Hello *, is someone aware how to set cutom structure member alignment for boost libs (DLLs) being compiled. We used ...\boost\config\abi\msvc_prefix.hpp #pragma pack(push,4) ...\boost\config\abi\msvc_suffix.hpp #pragma pack(pop) Either these files are not included by boost libs during the build process, or we have to enable their inclusion. Where to do it is unclear. Maybe we can set the correct flag in some .jam file at all. Any help would be greatly appreciated. Many thanks, Ovanes P.S. We are using Microsoft Visual Studio 2003 with boost 1.34.0 distribution.

1 0

connected_components
by Imran A. Pirwani 03 Jun '08

03 Jun '08

Hi, I am new to boost. I would like to use the following: typedef adjacency_list< listS, listS > Graph; Using a graph defined as above will allow me to call "remove_vertex". I would also like to use "connected_components()" to get connected components after I have removed a bunch of vertices (by calling "remove_vertex"). The example at: http://www.boost.org/doc/libs/1_35_0/libs/graph/example/connected_component… doesn't work for me because of the difference in adjacency lists used. Any help will be greatly appreciated! Thanks! -Imran.

2 2

[Units] and angular acceleration
by P Hawley 03 Jun '08

03 Jun '08

Hi, There are defined structs for angular_velocity, moment_of_inertia and angular_momentum as well as torque. Are you going to add angular_acceleration? Thanks Pat

4 3

normal_distribution<> behaviour
by Kevin Martin 03 Jun '08

03 Jun '08

I found this in the archive: http://thread.gmane.org/gmane.comp.lib.boost.user/26685 I am experiencing the same problem. However it works with variate_generator - see attached code. Is this simply a documentation bug which should say `Must always use variate_generator' or are the distributions supposed to work with any source of randomness? In which case they don't in 1.35.0 and I will go report it. Thanks, Kevin Martin #include <boost/random.hpp> #include <iostream> template<typename Randmoness_>void func1(Randmoness_ &r) { boost::normal_distribution<> x(180, 60); boost::uniform_real<> X(5.0, 10.0); std::cout << x(r) << " " << X(r) << std::endl; } template<typename Randmoness_>void func2(Randmoness_ &r) { boost::normal_distribution<> x(180, 60); boost::variate_generator<Randmoness_ &, boost::normal_distribution<> > y(r, x); boost::uniform_real<> X(5.0,10.0); boost::variate_generator<Randmoness_ &, boost::uniform_real<> > Y(r, X); std::cout << y() << " " << Y() << std::endl; } int main() { boost::mt19937 rng; func1(rng); func2(rng); return 0; }

1 0

[boost-users][statechart] state_machine::initiate()
by Igor R 03 Jun '08

03 Jun '08

Hi, Regarding the subj, the reference says: Effects: 1. Calls terminate() ...... 5. Processes all posted events (see process_event()). Returns to the caller if there are no more posted events But on terminate() call, all the posted events are cleared, so initiate() has nothing to process. Thus, I cannot write something like this: output MyMachine::invoke(SomeInput intput) { post_event(EvIput(input)); initiate(); // I want to proceed only after all the input is processed; the states might post internal events return output; } What would be the correct solution for this? Thanks a lot, Igor'.

1 1

BGL problem wit prims
by Line Blander Reinhardt 03 Jun '08

03 Jun '08

Hi I have some problems with using prim_minimum_spanning_tree function My major problem is I want to use a different weight than standard. My code: vector<standard::UnDirectedVertex> parent(num_vertices(graph_)); vector<std::size_t> distance(num_vertices(graph_)); boost::property_map<standard::Graph, mst_cost_t>::type weightmap = boost::get(mst_cost, graph_); boost::property_map<standard::Graph, boost::vertex_index_t>::type indexmap = get(boost::vertex_index, graph_); prim_minimum_spanning_tree(graph_,0, &parent[0], &distance[0],weightmap,indexmap,boost::default_dijkstra_visitor()); The error message is: 1>c:\program files\microsoft visual studio 8\vc\include\limits(109) : error C2440: '<function-style-cast>' : cannot convert from 'int' to 'W' 1> No constructor could take the source type, or constructor overload resolution was ambiguous 1> c:\program files\microsoft visual studio 8\vc\include\limits(107) : while compiling class template member function 'W std::numeric_limits<_Ty>::max(void) throw()' 1> with 1> [ 1> _Ty=W 1> ] 1> c:\boost\include\boost-1_33_1\boost\graph\prim_minimum_spanning_tree.hpp (61) : see reference to class template instantiation 'std::numeric_limits<_Ty>' being compiled 1> with 1> [ 1> _Ty=W 1> ] Can any one help getting the call right??? Thanks Line

1 0

[ptr_map] container equality comparison doesn't compile
by Rodolfo Lima 03 Jun '08

03 Jun '08

Hi, is there any other reason (apart from begin a bug) that the following code doesn't compile with g++-4.3? #include <boost/ptr_container/ptr_map.hpp> int main() { boost::ptr_map<int,int> m1, m2; m1.insert(1,std::auto_ptr<int>(new int(2))); m2.insert(1,std::auto_ptr<int>(new int(2))); return m1 == m2; } gcc spits the following errors (be seated...): /usr/lib/gcc/i686-pc-linux-gnu/4.3.0/include/g++-v4/bits/stl_algobase.h: In static member function 'static bool std::__equal<_BoolType>::equal(_II1, _II1, _II2) [with _II1 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, _II2 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, bool _BoolType = false]': /usr/lib/gcc/i686-pc-linux-gnu/4.3.0/include/g++-v4/bits/stl_algobase.h:818: instantiated from 'bool std::__equal_aux(_II1, _II1, _II2) [with _II1 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, _II2 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>]' /usr/lib/gcc/i686-pc-linux-gnu/4.3.0/include/g++-v4/bits/stl_algobase.h:950: instantiated from 'bool std::equal(_II1, _II1, _II2) [with _II1 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, _II2 = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>]' /usr/include/boost/ptr_container/detail/reversible_ptr_container.hpp:441: instantiated from 'bool boost::ptr_container_detail::reversible_ptr_container<Config, CloneAllocator>::operator==(const boost::ptr_container_detail::reversible_ptr_container<Config, CloneAllocator>&) const [with Config = boost::ptr_container_detail::map_config<int, std::map<int, void*, std::less<int>, std::allocator<std::pair<const int, void*> > > >, CloneAllocator = boost::heap_clone_allocator]' teste.cpp:10: instantiated from here /usr/lib/gcc/i686-pc-linux-gnu/4.3.0/include/g++-v4/bits/stl_algobase.h:789: error: no match for 'operator==' in '((boost::iterator_facade<boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, boost::ptr_container_detail::ref_pair<int, const int* const>, boost::bidirectional_traversal_tag, boost::ptr_container_detail::ref_pair<int, const int* const>, int>*)(& __first1))->boost::iterator_facade<I, V, TC, R, D>::operator* [with Derived = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, Value = boost::ptr_container_detail::ref_pair<int, const int* const>, CategoryOrTraversal = boost::bidirectional_traversal_tag, Reference = boost::ptr_container_detail::ref_pair<int, const int* const>, Difference = int]() == ((boost::iterator_facade<boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, boost::ptr_container_detail::ref_pair<int, const int* const>, boost::bidirectional_traversal_tag, boost::ptr_container_detail::ref_pair<int, const int* const>, int>*)(& __first2))->boost::iterator_facade<I, V, TC, R, D>::operator* [with Derived = boost::ptr_map_iterator<std::_Rb_tree_const_iterator<std::pair<const int, void*> >, int, const int* const>, Value = boost::ptr_container_detail::ref_pair<int, const int* const>, CategoryOrTraversal = boost::bidirectional_traversal_tag, Reference = boost::ptr_container_detail::ref_pair<int, const int* const>, Difference = int]()' make: ** [teste] Erro 1 My guess it that boost::ptr_container_detail::ref_pair should have an equality operator (and a less than also) somehow. In fact... by adding the following operator== to it the code compiles and behaves as expected: // #include <boost/utility/compare_pointees.hpp at top of the file bool operator==(const ref_pair &that) const { return first==that.first && boost::equal_pointees(second, that.second); } Something equivalent should be done for operator< Any objections? Regards, rod

2 1

[boost::test] running unit tests from DLL with automatic registration facility
by Ovanes Markarian 03 Jun '08

03 Jun '08

Hello Gennadiy, using boost 1.35 & console runner and MSVC 2005 we did the following test: FunctionalityDLL ^ | UnitTestDLL ^ | ConsoleTestRunner.EXE (as distributed in the example directory) UnitTest.DLL had the following defines: #ifndef BOOST_TEST_DYN_LINK # define BOOST_TEST_DYN_LINK #endif #ifndef BOOST_TEST_MODULE # define BOOST_TEST_MODULE example #endif #include <boost/test/unit_test.hpp> We spent some time to figure out how to cause ConsoleTestRunner.exe to find the init_unit_test function. Finally we end up looking to preprocessed code: There we found the function: bool init_unit_test() { using namespace ::boost::unit_test; assign_op( framework::master_test_suite().p_name.value, boost::unit_test::const_string( "example", sizeof( "example" ) - 1 ).trim( "\"" ), 0 ); return true; } Only after taking the function over to the cpp module (removing BOOST_TEST_MODULE define) adding the DLL export statement and adding the C-Linkage ConsoleTestRunner was able to find the address of this function. So the compilation unit looked like: #ifndef BOOST_TEST_DYN_LINK # define BOOST_TEST_DYN_LINK #endif #include <boost/test/unit_test.hpp> #include "TestsDLL.h" extern "C" __declspec(dllexport) bool init_unit_test() { using namespace ::boost::unit_test; assign_op( framework::master_test_suite().p_name.value, boost::unit_test::const_string( "example", sizeof( "example" ) - 1 ).trim( "\"" ), 0 ); return true; } BOOST_AUTO_TEST_SUITE( FunctionalityDLLTest ) BOOST_AUTO_TEST_CASE( test1 ) { BOOST_WARN( sizeof(int) < 4 ); } BOOST_AUTO_TEST_CASE( test2 ) { BOOST_REQUIRE_EQUAL( 1, 2 ); BOOST_FAIL( "Should never reach this line" ); } BOOST_AUTO_TEST_SUITE_END() Only after these modifications we were able to compile and run unit tests. Is there any chance to either update the docs with this example so that it becomes a sort of guarantee how it might work, or let the combination of BOOST_TEST_MODULE and BOOST_TEST_DYN_LINK generate the right linkage specification and export the symbol, otherwise this symbol was not found with GetProcAddress. With Kind Regards, Ovanes

2 1

Re: [Boost-users] serialization, 1.34/1.35 difference: serializing a vector
by Filip Peters 03 Jun '08

03 Jun '08

----- Original Message ---- From: "Pfligersdorffer, Christian" <Christian.Pfligersdorffer(a)eos.info> To: boost-users(a)lists.boost.org Sent: Friday, April 4, 2008 6:30:00 AM Subject: Re: [Boost-users] serialization, 1.34/1.35 difference: serializing a vector Hello everybody, we at EOS too rely heavily on backwards compatibility of boost::serialization so I read this topic with great interest. And found Filip's discovery quite shocking I must say! I tried the testing program on my machine using boost-1.33.1 and boost-1.35 and all of the boost::archives. And guess what: It worked!! Soooo, this is quite confusing! Either version 1.34 breaks the compatibility or Filip's observation is wrong. Or mine, as everything might depend on the machine architecture or some other nasty detail... Is think this has to be investigated closer. Very interesting. I dug up the zip file for 1.33.1 and ran a couple of tests. (mostly to prove myself that I'm not seeing ghosts.) result: from 1.33.1 to 1.34: no problem from 1.33.1 to 1.35: no problem from 1.34.1 to 1.35: problem. problem seems to be a new "optimization" in the vector serializing. I normally use the binary files, but for debugging purposes the xml output is fantastic. It shows exactly why it works/doesn't work. xml from 1.33.1: <?xml version="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE boost_serialization> <boost_serialization signature="serialization::archive" version="3"> <vectorToSave> <count>2</count> <item>0</item> <item>1</item> </vectorToSave> </boost_serialization> xml from 1.34.1: <?xml version="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE boost_serialization> <boost_serialization signature="serialization::archive" version="4"> <vectorToSave> <count>2</count> <item_version>0</item_version> <item>0</item> <item>1</item> </vectorToSave> </boost_serialization> xml from 1.35: <?xml version="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE boost_serialization> <boost_serialization signature="serialization::archive" version="4"> <vectorToSave> <count>2</count> <item>0</item> <item>1</item> </vectorToSave> </boost_serialization> As you can clearly see, from 1.34.1 to 1.35, the archive version stayed at 4, while item_version went missing in 1.35. Btw: my portable binary archive which really does nothing more than endian handling on top of the binary archive was broken by version 1.34 too (wrt reading files generated by 1.33.1). It has to do with the serialization of std collections as well, that much I found out. I did not do extensive investigation then but simply did not migrate. I may have time to take a closer look in the near future though. Regards, -- Christian Pfligersdorffer Software Engineering http://www.eos.info PS: Jens wrote: > Have you reported these issues? Please remember that once you > start testing development versions, especially release > snapshots, you could ensure that such issues are fixed before the > release! Never forget that there users who are not so involved in the boost community, be it for the lack of interest or simply because they are engaged in projects that leave no room for testing of upcoming library versions. Between the announcement and the official release there was less than a laughable day... Exactly. I didn't check the boost list in the previous two weeks (busy with an internal release myself), so wasn't really aware that there was rc2 and rc3.(heck, was rc1 even announced on the users list?) Monday I checked and saw that the boost website was updated, and a nice new version. When I got everything to compile, I noticed that I couldn't read in some older files. So I made a sample program, and reported the output. I also opened a trac ticket for this. Filip Peters. ____________________________________________________________________________________ You rock. That's why Blockbuster's offering you one month of Blockbuster Total Access, No Cost. http://tc.deals.yahoo.com/tc/blockbuster/text5.com

3 3

Re: [Boost-users] boost & c-style calback issue
by Anders Dalvander 02 Jun '08

02 Jun '08

> > Your approach is exactly what we don't want, since it makes it an object-based > based solution rather than instance-based. > We are using a functor (therefor the oprerator()) so that the solution remains > instance-based. > Is your application threaded? If so couldn't you use thread_specific_ptr for emulating a instance-based functor? class A; class B; class C { public: C() { instance().push_back(this); } ~C() { instance().pop_back(); } int operator()(A* a, B* b) { // Do your stuff. } static int wrapper(A* a, B* b) { return instance().back()->operator()(a, b); } private: C(const C& rhs); C& operator=(const C& rhs); static this_vector& instance() { this_vector* t = ptr.get(); if (!t) { t = new this_vector(); ptr.reset(t); } return *t; } typedef std::vector<C*> this_vector; static boost::thread_specific_ptr<this_vector> ptr; }; typedef int (*foobar)(A*,B*); void do_something(foobar cb); C obj; do_something(C::wrapper); Regards, Anders

1 0

Parsing URI's using boost regex
by Reena Dass 02 Jun '08

02 Jun '08

I'm trying to write a parser for URIs. I tried the grammar straight from RFC 2396 boost::regex parser("^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?", boost::regex::perl|boost::regex::icase); This does not work since it I get an exception due to invalid regular expression. I found another regular expression, which does not throw an exception but does not give me the parts of the URI using boost::match_result boost::regex parser("(([a-zA-Z][0-9a-zA-Z+\\-\\.]*:)?/{0,2}[0-9a-zA-Z;/?:@&=+$\\.\\-_ !~*'()%]+)?/{0,1}([0-9a-zA-Z;/?:@&=+$\\.\\-_!~*'()%]+)?(#[0-9a-zA-Z;/?:@ &=+$\\.\\-_!~*'()%]+)?", boost::regex::perl|boost::regex::icase); What am I doing wrong here? Reena

2 1

Is there any class that can promote a type to the complex type?
by Peng Yu 02 Jun '08

02 Jun '08

Hi, To promote a type, no matter complex or not, to the complex type can be done by the following code. RET would give me a complex type. The code is simple. I'm wonder if boost has already supplied such facility so that I don't have to include the following snippet in my own code. Thanks, Peng #include <complex> #include <iostream> template <typename T> struct convert { typedef std::complex<T> RET; }; template <typename T> struct convert<std::complex<T> > { typedef std::complex<T> RET; }; int main() { std::cout << convert<double>::RET() << std::endl; std::cout << convert<std::complex<double> >::RET() << std::endl; }

4 8

[boost.function] evaluating needed arguments only
by Olaf Krzikalla 02 Jun '08

02 Jun '08

Hi, is it possible to check a function bound to boost::function, if a particular argument is actually needed? Example: ---- struct bar {}; typedef boost::function<void(int, bar&)> tSlot; void foo(tSlot& s); void barNeeded(int, bar&); void barNeeded(int); void callfoo() { foo(boost::bind(&barNeeded, _1, _2)); foo(boost::bind(&barNotNeeded, _1)); } void foo(tSlot& s) { bar b; expensive_calculation(b); s(1, b); } ---- How can I prevent expensive_calculation if the function barNotNeeded was bound to tSlot? Best regards Olaf Krzikalla

2 3

Help with deduced template parameters
by Chris Ross 02 Jun '08

02 Jun '08

Forgive me if the answer ends up being strictly C++, rather than Boost. But, I have a template class that I would rather be able to deduce it's single [current] template parameter rather than having to specify it. I'm wondering if boost::function or boost::functional might make this easier for me to accomplish. My functor class is passed to a method similar to std::for_each, and that template functor class is: /*! Detach all members of the EntityList via stop_messaging. * \todo I suspect there's a way to make this not require template * arguments at instantiation, but at the moment it's being used * as: * mutexed_for_each(detach_all<typeof(arg)>(arg)); */ template <typename List> class detach_all { public: detach_all(const List &e) : list(e) {} template <class T> void operator ()(T*ent) const { if (dynamic_cast<Messageable*>(ent)) { std::for_each(list.begin(), list.end(), stop_messaging(dynamic_cast<Messageable*>(ent))); } } private: const List &list; }; If I make the class not be a template, but instead make the constructor a template member, it would be able to determine the template parameter List from the argument to the constructor. However, I'm then unable to keep a reference to the List argument to the constructor, which I need during the operator() invocation. Am I missing something about templates that I don't know? Or, is this just not possible with C++ templates, and in which case, can I use one of the boost libraries to help? Thanks much. - Chris

3 2

Re: [Boost-users] Problems using pre-compiled boost libraries with Visual Studio 2005
by Fanara, Mark A (GE Infra, Transportation) 02 Jun '08

02 Jun '08

Trigve - Installing SP1 corrected the problem. Thanks! Mark

1 0

Re: [Boost-users] Boost interprocess map in shared memory
by Jan Suchý 02 Jun '08

02 Jun '08

Hallo Ion, it's perfect. It's working now very well. I have used the v 1.35 solution. I am planning to port the application to the HP-UX from Linux. Do you see any problems to use interprocess module 1.35 on that platform? Thank you, have a nice day, Jan Suchy Date: Fri, 30 May 2008 21:22:45 +0200 From: Ion Gazta?aga <igaztanaga(a)gmail.com> Subject: Re: [Boost-users] Boost interprocess map in shared memory To: Boost User List <boost-users(a)lists.boost.org> Message-ID: <48405405.7010801(a)gmail.com> Content-Type: text/plain; charset=UTF-8; format=flowed Jan Such? wrote: > Now, I want to have library, where i want to find previously created shared segment and read it more-times. > But I do not want to open it on every runQ function call. > I need to open this segment only once and use it more-times when runQ in lib is called. > > I tried this: The problem is that managed_shared_memory is not copyable. With current trunk code you could move it from localsegment to segment: managed_shared_memory localsegment(open_only, "MySharedMemory"); segment = move(localsegment); //localsegment is now in default-constructed state For 1.35 code allocate it through new (not compiled, this might contain errors): std::auto_ptr<managed_shared_memory> segment_ptr; int init() { segment_ptr.reset (new managed_shared_memory (open_only, "MySharedMemory") ); } Regards, Ion

1 0

[Test] Documentation confusion
by Maik Beckmann 02 Jun '08

02 Jun '08

Hello, I am confused about the boost.test documentation. There are docs in trunk and after some hours of digging via gmane if found this one http://www.patmedia.net/~rogeeff/html/index.html index.html in trunk : ... Revised: 28 February, 2006 index.html external link: Copyright © 2001-2007 Gennadiy Rozental ... My impression is that the 2007 version is much more comprehensive. Why isn't the most recent one in trunk? Do the trunk docs got a link the the external docs? -- Maik

2 2

[test] Using the Unit Test Framework in Xcode 3 howto
by Richard Dingwall 02 Jun '08

02 Jun '08

Hi guys, I've written a short tutorial on getting up and running with the Boost Unit Test Framework in Xcode 3, including getting the IDE to parse test results. You can read it here: http://richarddingwall.name/2008/06/01/using-the-boost-unit-test-framework-… Richard

3 2

Boost.Test: missing include file
by Jarl Friis 02 Jun '08

02 Jun '08

I am using BOOST_CHECK_CLOSE which is defined in test_tools.hpp. I my test I have #include <boost/test/unit_test.hpp> Aparently this requires that I also include #include <boost/test/floating_point_comparison.hpp> wouldn't it be more correct to have this include in the test_tools.hpp file? Jarl

2 1

Test tree is empty
by Stephen Torri 02 Jun '08

02 Jun '08

I am sure this is a silly common answer but I am not sure how to resolve it. Here is my test suite below. I am following http://www.hpfsc.de/boosttest/ as my guide. #define BOOST_TEST_DYN_LINK #define BOOST_TEST_MODULE Statistical algorithm test suite #include <boost/test/unit_test.hpp> #include "Statistical_Test.h" using boost::unit_test_framework::test_suite; using boost::unit_test_framework::test_case; using namespace libreverse::classifier; class statistical_test_suite : public test_suite { public: statistical_test_suite () : test_suite ( "statistical_test_suite" ) { // create an instance of the test cases class boost::shared_ptr<Statistical_Test> instance ( new Statistical_Test() ); test_case* test_Constructor_test_case = BOOST_CLASS_TEST_CASE( &Statistical_Test::test_Constructor, instance ); add ( test_Constructor_test_case ); } }; test_suite* init_unit_test_suite ( int, char** ) { test_suite* top_test_suite = BOOST_TEST_SUITE("Statistical algorithm test suite"); // add test suites to the top test suite top_test_suite->add ( new statistical_test_suite() ); return top_test_suite; }

2 1

variant<> number of template arguments
by Hicham Mouline 01 Jun '08

01 Jun '08

Hello, I have a use of the visitor pattern (this was an answer provided to me on c++-moderated) as shown below. Question1: is the maximum number of variant<> arguments extendable? Question2: Below, some of the variants I have are full template specs. Is there a way to write the variant<> typedef with templates instead and have the operator() of the static visitor a template member. Rds, --------------- #include <iostream> #include <boost/variant.hpp> struct derived1 {}; struct derived2 {}; struct derived3 {}; typedef boost::variant<derived1, derived2, derived3> base; // function object struct operation : boost::static_visitor<int> { int operator()(const derived1&) const { return 1; } int operator()(const derived2&) const { return 2; } int operator()(const derived3&) const { return 3; } }; int main() { base a = derived1(); operation op; std::cout << a.apply_visitor(op) << std::endl; // prints "1" base a = derived2(); std::cout << a.apply_visitor(op) << std::endl; // prints "2" }

2 7

Having trouble using lapack bindings with boost
by Omar Alvi 01 Jun '08

01 Jun '08

Hello People, I am new to boost and I am trying to use use lapack bindings with boost as I need to use svd and some other advanced operations to solve linear systems. I have downloaded boost-bindings-20080501 from http://mathema.tician.de/dl/software/boost-bindings but cant figure out how to compile it on windows using VC8. Could someone please tell me how I can do that ? I saw somewhere that people have compiled it on linux using ./configure, make and make install but I want to compile it on windows. Thanks, Omar

1 0

Neighborhood of a subgraph
by Imran A. Pirwani 01 Jun '08

01 Jun '08

Hi, I am new to BGL and wanted to find out if there's an easy way to do the following in BoostGL: Given a subset of vertices (specified as a parameter to the function), a graph g, how does one query the neighbors of the subset that are not in the subset? Is there a function already written that I can use or an easy way to implement the above using simple building-blocks? I could maybe write my own...but I need to be making this query many many times for different subgraphs and I'd prefer to not have to check the membership within the subset many times over and over again. Any ideas will be greatly appreciated! Thanks! -Imran.

1 0

Trouble building statecharts with gcc 4.3.0
by Abhinandan Jain 01 Jun '08

01 Jun '08

Hi I am having trouble compiling boost statechart code using gcc 4.3.0. I tried this on the +sample program on statechart site and get the same errors. The errors are of the type error: value-initialization of reference for each process_event call in the example: myWatch.process_event( EvStartStop() ); myWatch.process_event( EvStartStop() ); myWatch.process_event( EvStartStop() ); Does anyone have a suggestion on how to fix this? Thanks Abhi

2 2

How to open a file with its associated application?
by Zdravko Monov 01 Jun '08

01 Jun '08

Hi! Is there a C++ library that can, given a filename, start the default application for its type to open it? -- Zdravko Monov

2 1

Re: [Boost-users] Boost-users Digest, Vol 1645, Issue 3
by Fox Fossati 31 May '08

31 May '08

Message: 8 Date: Wed, 28 May 2008 09:18:22 +0900 From: "Kurata Sayuri" <magicalsayuri(a)gmail.com> Subject: Re: [Boost-users] Visual C++ 2008- from system::String to boost::any ? To: boost-users(a)lists.boost.org Message-ID: <b09a4ea0805271718vffa0c18m2d2565947fd159e0(a)mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1 On Wed, May 28, 2008 at 12:19 AM, Fox Fossati <stonfofish(a)yahoo.it> wrote: > Excuse me, > I' m reading a DataGridView with VisualC++2008 with nested loops.... > ..... > Object^ val=dataGridView1[x,y]->Value; //Object is a Visual class > ..... > I convert every value in a System::String variable: > String^ str=val->ToString(); > > Then I'have to insert this data in a container (for example) this: > vector<boost::any>* vec=new vector<boost::any>(n) ; > //0<=x<n > (*vec)[x]=str; //ERROR! > (*vec)[x]=*str; //ERROR! > > Is there another method to transfer a value from a dataGridView to an STL > container? > Now i wrote only a vector to explain simply but in fact i will use a table > of this type vector<vector<boost:.any>*>. Your code is C++/CLI. So, you can use System::Object instead of boost::any. And also, cliext::vector instead of std::vector. Because native class (such as boost::any, std::vector) can't have managed instance. ---- #include <cliext/vector> Object^ val = dataGridView1[x,y]->Value; String^ str = val->ToString(); cliext::vector<Object^>^ vec = gcnew cliext::vector<Object^>(n) ; // or cliext::vector<String^> (*vec)[x]=str; ------------------------------ Is it possible to convert a System::String to an std:string? With a conversion then i can use my boost classes...but i need to convert from std::string to System::String and vice versa.... I should use visual c++ only for interface input/output unless using visual classes....do you think i will have a lot of troubles? It's very difficult to find a good ide rad.... Thanks a lot! ___________________________________ Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e la tua opinione! http://www.ymailblogit.com/blog/

1 0

[multi_array] how to modify the dimension of an array?
by r0d 31 May '08

31 May '08

Hello everybody, i'm looking for the best way to modify the dimensions of an array. For example, i got this one: boost::multi_array<int, 2> array ( boost::extents[3][3] ); and i want to change it in an array of 5 dimensions. Now, what i'm doing is to create a new array of 5 dimensions and copying the first one into it, then delete the first one. But i'm sure that a more efficient way exists. No? -- View this message in context: http://www.nabble.com/-multi_array--how-to-modify-the-dimension-of-an-array… Sent from the Boost - Users mailing list archive at Nabble.com.

2 3

Re: [Boost-users] Problems using pre-compiled boost libraries with Visual Studio 2005
by Trigve Siver 31 May '08

31 May '08

Hi, > From: "Fanara, Mark A (GE Infra, Transportation)" <Mark.Fanara(a)ge.com> > > I downloaded the current pre-built boost library files, but am having > some difficulty with them. > ... > > Next error message received is "This application has failed to start > because the application configuration is incorrect. Review the manifest > file for possible errors. Reinstalling the application may fix this > problem. For more details, please see the application event log." > > ... Do you have VS 2005 SP1 installed ? I think that precompiled packages are built on 2005 SP1 Trigve

1 0

[boost-users] boost::function object alignment question
by Neil Fang 31 May '08

31 May '08

Hi I have a win32 static library with a class like this : typedef boost::function<void (int)> TimerCallbackType; class TimerManager { public: int addTimer(int interval, TimerCallbackType callback); }; I use this lib in a Win32 exe, but it crashed on calling TimerManager::addTimer(), the error is: "Run-time check failure #0 - The value of ESP was not properly saved across a function call ... ". Both the exe and lib are compiled with "Struct Alignment" set to "4 Bytes". Finally, I find out a way make it work -- include the boost like this : #pragma pack(push,8) #include <boost/xxxx> #pragma pop() But, I don't know why? Could anybody here help to clear this question? I'm using MS Viusal C++ 2005. Thanks! -- Neil.Fang.CN

1 0

async_XXX() functions
by Longyu Mei 31 May '08

31 May '08

I noticed that there are couple of async_XXX methods, such as async_wait, async_connect, async_write and async_fill. Just wondering why we need those async_XXX methods? without the "async", those methods will block the execution and next statements will be deferred. with the "async", those methods will return immediately and the work will be deferred a little time but will still be executed in the same thread. I feel it confuses people. If we need next statements being executed immediately, we can use another thread to do the wait, connect, write or fill. It is more clear and clean. And the user/programmer has fully control on this. Maybe I have missed something here. please give me your comments. thanks, James

2 1

Re: [Boost-users] shared memory overhead
by Jason Sachs 30 May '08

30 May '08

> > Modern systems use on demand allocation. I.e. you can allocate a (f.ex.) > 32 MB > SHM chunk, but the actual resource usage (RAM) will correspond to what you > actually use. For example: > > 0 1M 32M > [****|..................] > | | > | +- unused part > | > +- used part of the SHM segment > As long as your program does not touch (neither reads nor writes) the > unused part, > the actual physical memory usage will be 1M + small amount for page tables > (worst > case: 4kB of page tables for 4MB of virtual address space). This is at > least how > SYSV SHM works on Solaris 10 (look up DISM - dynamic intimate shared > memory); I > would expect it to work in the same way on new linux kernels too. I'm not > sufficiently acquainted with NT kernel to be able to comment on it. Interesting. I hadn't thought about that. I tried a test program (running on Windows XP), and had a number of separate processes each allocate a new managed_windows_shared_memory object (with a different name for each process) with size 2^29 bytes (=512MB). I'm not exactly sure what resources it allocates; using TaskInfo to view resource usage, each process's "virtual KB" usage goes up by 512MB, but its "working set KB" usage doesn't increase until I actually allocate memory within the shared memory segment. Sounds good, but the 6th one of these failed and I got a warning saying my system was low on virtual memory. So it sounds like there is a 4GB total system limit for WinXP even for just reserving virtual address space -- which seems silly since each process should have its own address space and therefore as long as I don't actually allocate the memory, and each process's reserved address space doesn't exceed 2^32 (or 2^31 or whatever the per-process limit is), I should be able to reserve an unlimited amount of total address space. No can do. :( So strategy #1 of being profligate in choosing shared memory segment size fails on WinXP; there's a significant resource cost even if you don't actually allocate any memory. Drat.

3 4