- Boost-users - lists.boost.org

Re: [Boost-users] [msm] Acceptance into boost?
by christophe henry 01 Sep '09

01 Sep '09

>I really admire the Boost.MSM library currently in the vault. I like it a lot better than the original FSM. Are there any plans to integrate the msm library into >an official Boost release?> >--------- >Robert Dailey Thanks for the feedback! I'm happy that you like it! Msm is right now in the review queue but I need a review manager and I have been very busy with msm 2.0 and therefore had no time to look for one. So I'd like to use the chance to ask if some experienced boost'ler would accept to be the review manager for msm and help me speed up the process a little. If someone would also be interested in playing with msm 2 as beta tester please contact me too :) Christophe _________________________________________________________________ More than messages–check out the rest of the Windows Live™. http://www.microsoft.com/windows/windowslive/

1 0

[Test] data_access_test && common_layer::const_string && missing std::out_of_range exception
by Dmitry V. Krivenok 01 Sep '09

01 Sep '09

Hello! I'm trying to build boost-1.40.0 and have problems with Boost::Test. Please take a look at build log: ------------------------------------------------------------------------ common.mkdir /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test common.mkdir /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4 common.mkdir /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug gcc.compile.c++ /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.o gcc.link /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test testing.capture-output /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.run ====== BEGIN OUTPUT ====== Running 6 test cases... ../../libs/test/example/const_string_test.cpp(71): error in "data_access_test": exception std::out_of_range is expected *** 1 failure detected in test suite "const_string test" EXIT STATUS: 201 ====== END OUTPUT ====== LD_LIBRARY_PATH="/usr/local/bin:/usr/local/lib:/usr/local/lib32:/usr/local/lib64:$LD_LIBRARY_PATH" export LD_LIBRARY_PATH "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test" > "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.output" 2>&1 status=$? echo >> "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.output" echo EXIT STATUS: $status >> "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.output" if test $status -eq 0 ; then cp "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.output" "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.run" fi verbose=0 if test $status -ne 0 ; then verbose=1 fi if test $verbose -eq 1 ; then echo ====== BEGIN OUTPUT ====== cat "/home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.output" echo ====== END OUTPUT ====== fi exit $status ...failed testing.capture-output /home/krivenok/dev_builds/boost_1_40_0_build/boost/bin.v2/libs/test/example/const_string_test.test/gcc-4.4/debug/const_string_test.run... ...failed updating 1 target... ...updated 174 targets... ------------------------------------------------------------------------ The problem is: ../../libs/test/example/const_string_test.cpp(71): error in "data_access_test": exception std::out_of_range is expected I looked at line 71 in const_string_test.cpp: BOOST_CHECK_THROW( cs1.at( cs1.length() ), std::out_of_range ); This code is used to ensure that std::out_of_range is thrown in case of out of bound access to the string. However, common_layer::const_string doesn't provide such validation. operator[] and at() are identical in const_string: char operator[]( size_t index ) const { return m_begin[index]; } char at( size_t index ) const { return m_begin[index]; } I commented BOOST_CHECK_THROW out and now Boost::Test compiles w/o any errors. I think implementation of const_string::at() should be fixed.

2 1

[serialization] loading a map having a const mapped value type
by Marco Canini 01 Sep '09

01 Sep '09

Hi all, I'm unable to address this use case of the serialization library. I have a standard map in which the mapped value is const: map<string, const Protocol *> _table; My serialize is very simple: ar & _table; (the test case source code is attached) However, this can't be compiled as I get many errors related to the loading code inside boost::serialization: /usr/include/boost/archive/detail/iserializer.hpp:488: error: reinterpret_cast from type 'const Protocol**' to type 'void**' casts away constness /usr/include/boost/archive/detail/iserializer.hpp:491: error: no matching function for call to 'boost::archive::detail::load_pointer_type<boost::archive::text_iarchive, const Protocol*>::pointer_tweak(const boost::serialization::extended_type_info&, const Protocol*&, const Protocol&)' /usr/include/boost/serialization/access.hpp:109: error: passing 'const Protocol' as 'this' argument of 'void Protocol::serialize(Archive&, unsigned int) [with Archive = boost::archive::text_iarchive]' discards qualifiers /usr/include/boost/archive/detail/iserializer.hpp: In member function 'void boost::archive::detail::pointer_iserializer<T, Archive>::load_object_ptr(boost::archive::detail::basic_iarchive&, void*&, unsigned int) const [with T = const Protocol, Archive = boost::archive::text_iarchive]': /usr/include/boost/archive/detail/iserializer.hpp:299: error: invalid conversion from 'const void*' to 'void*' /usr/include/boost/archive/detail/iserializer.hpp:307: error: invalid conversion from 'const void*' to 'void*' /usr/include/boost/archive/detail/iserializer.hpp:307: error: initializing argument 1 of 'void boost::archive::detail::basic_iarchive::next_object_pointer(void*)' /usr/include/boost/serialization/access.hpp:123: error: invalid conversion from 'const void*' to 'void*' /usr/include/boost/serialization/access.hpp:123: error: initializing argument 2 of 'void* operator new(size_t, void*)' /usr/include/boost/serialization/access.hpp:123: error: no matching function for call to 'Protocol::Protocol()' /usr/include/boost/archive/detail/iserializer.hpp:555: error: invalid application of 'sizeof' to incomplete type 'boost::STATIC_ASSERTION_FAILURE<false>' /usr/include/boost/archive/detail/iserializer.hpp:555: error: invalid application of 'sizeof' to incomplete type 'boost::STATIC_ASSERTION_FAILURE<false>' Can somebody share his wisdom on this please? Thanks! -- Marco Canini

3 6

[mpl::print] Making mpl::print<> function
by Robert Ramey 01 Sep '09

01 Sep '09

I have the following which I expect to produce a compile time warning which includes the type name but it fails to do so. I found this in the C++ Template Metaprogramming book - but not in the MPL documentation. What am I missing here? Robert Ramey #include <boost/mpl/print.hpp> template<class T> int f(){ boost::mpl::print<T>; //static int x = g<T::value>(); return x; } struct A { BOOST_STATIC_CONSTANT(bool, value = false); }; ... int x = f<A>();

3 2

[msm] Acceptance into boost?
by Robert Dailey 01 Sep '09

01 Sep '09

I really admire the Boost.MSM library currently in the vault. I like it a lot better than the original FSM. Are there any plans to integrate the msm library into an official Boost release? --------- Robert Dailey

1 0

Iterator behaving like an indirect_iterator and a filter_iterator
by Teto 01 Sep '09

01 Sep '09

Hi everyone, I wanted to know if it was possible to have an iterator behaving like both an indirect_iterator and a filter_iterator. I've tried to without success. Here is a minimal code that compiles but crashes. ===================================== class CTest { public: CTest(int Value,int Id) : value(Value),id(Id) {}; int id; int value; }; std::vector<CTest*> vec; vec.push_back(new CTest(4,7)); vec.push_back(new CTest(4,1)); //I want an iterator that I could use like iter->id / iter->value and not (*iter)->id / (*iter)->value struct SPredicate { bool operator()( C& x) { return (-1 < x.id); } }; typedef boost::filter_iterator<SPredicate, CTest*> FilterIterator; FilteredIterator end(*vec.end()); for(FilteredIterator iter(*vec.begin());start != end; iter++){ std::cout << "id" << iter->id << "\tValue" << iter->m << std::endl; } ===================================== I'm new to boost and is very impressed by its possibilites. Hope I can understand the code better later. thanx matt

3 2

[mem_fn] return type
by Lowell McLinskey 01 Sep '09

01 Sep '09

I am having trouble understanding the return type of mem_fn. Is it possible to create a vector of this type, analogous to how one can create a vector of type mem_fun_t to hold the return values of mem_fun()? Here is an example of what I'd like to do. Thanks in advance. #include <vector> #include <functional> #include "boost/shared_ptr.hpp" #include "boost/mem_fn.hpp" class A { public: A(const int a) : a_(a) {} const int get() const {return a_;} private: int a_; }; int main(int argc, char** argv) { // I can do this std::vector<std::const_mem_fun_t<const int, A> > v; v.push_back(std::mem_fun(&A::get)); int i = v[0](new A(4)); // and this boost::shared_ptr<A> p(new A(7)); int j = boost::mem_fn(&A::get)(p); // but how do I do this? // std::vector< WHAT GOES HERE > w; // w.push_back(boost::mem_fn(&A::get)); // int k = w[0](p); return 0; }

3 3

[asio] What happens when a deadline_timer is deleted
by Bjarne Laursen 01 Sep '09

01 Sep '09

If I have a deadline_timer in my class and my class is deleted, the compleetion function will be called (with error code). But as the destructor have already been run I guess that the 'this' pointer will be invalid! Is this safe? What if my compleetion function is virtual? timer.async_wait(boost::bind(&MyClass::OnTimer, this, boost::asio::placeholders::error)); --

2 1

[Units] Runtime dimension check
by Thomas Taylor 01 Sep '09

01 Sep '09

Hi all! Is there an easy way to check if two quantities have the same dimension? E. g. a user enters a weight in tons and asks for a conversion to kilograms. So now at run time I would like to know if tons and kilograms have the same dimensions (i. e. are convertable) or not (illegal conversion). Thanks, Thomas

2 1

Problem detecting python
by Philippe Fremy 01 Sep '09

01 Sep '09

Hi, bjam does not seem to pick my python installation on windows. This is running on Windows XP, toolset gcc, boost 1.39, compiling from source with no changes. D:\program\boost_1_39_0>bjam --debug-configuration -d+2 --debug-building [...] notice: [python-cfg] running command '"C:\Python25\python" -c "from sys import *; print 'version=%d.%d\nplatform=%s\nprefix=%s\nexec_prefix=%s\nexecutable=%s' % (version_info[0],version_info[1],platform,prefix,exec_prefix,executable)" 2>&1'notice: [python-cfg] ...does not invoke a working interpreter notice: [python-cfg] No working Python interpreter found. The strange thing is that if I run the same command line, it works perfectly: D:\program\boost_1_39_0>"C:\Python25\python" -c "from sys import *; print 'version=%d.%d\nplatform=%s\nprefix=%s\nexec_prefix=%s\nexecutable=%s' % (version_info[0],version_info[1],platform,prefix,exec_prefix,executable)" 2>&1 version=2.5 platform=win32 prefix=C:\Python25 exec_prefix=C:\Python25 executable=C:\Python25\python.exe Any idea what confuses bjam ? I could not find any log file to see more details of what is going wrong... cheers, Philippe

5 8