March 2007 - Boost-users - lists.preview.boost.org

[fusion][mpl] How to make a vector of a vector of types
by Christian Henning 15 Mar '07

15 Mar '07

Hi there, subject probably sounds a bit strange but here is what I'm trying to accomplish. The problem description goes like this: using namespace boost; using namespace boost::fusion; struct A; struct B; struct C; int main() { typedef vector< A, B, C > types; // how to translates typed into the following?? typedef vector< std::vector<A>, std::vector<B>, std::vector<C> > vector_of_types; return 0; } I'm looking for a generic approach to do this. To me it seems like it should be possible by taking parts of the MPL lib, like mpl::push_back, and parts of the fusion lib, like fusion::for_each. Any ideas? Thanks ahead, Christian

2 2

[statechart] - problem with sc::transition<> -> invalid use of 'boost::statechart::detail::no_context::no_function [with Event = jm::ffcp::ev_init]' to form a pointer-to-member-func
by Oliver.Kowalke＠qimonda.com 15 Mar '07

15 Mar '07

Hello, I get following error: /opt/boost/include/boost-1_33_1/boost/statechart/transition.hpp:33: error: invalid use of 'boost::statechart::detail::no_context::no_function [with Event = jm::ffcp::ev_init]' to form a pointer-to-member-function /opt/boost/include/boost-1_33_1/boost/statechart/transition.hpp:33: note: a qualified-id is required if I try to compile: #ifndef JM_FFCP_UNINITIALIZED_H #define JM_FFCP_UNINITIALIZED_H #include <boost/mpl/list.hpp> //#include <boost/statechart/custom_reaction.hpp> #include <boost/statechart/simple_state.hpp> #include <boost/statechart/transition.hpp> #include <jm/ffcp/events.h> #include <jm/ffcp/proto.h> namespace jm { namespace ffcp { namespace mpl = boost::mpl; namespace sc = boost::statechart; template< typename SocketDeviceT > struct initialized; template< typename SocketDeviceT > struct uninitialized : public sc::simple_state< uninitialized< SocketDeviceT >, proto< SocketDeviceT > > { typedef mpl::list< // sc::custom_reaction< ev_init > sc::transition< ev_init, initialized< SocketDeviceT > > > reactions; uninitialized() { uninitialized< SocketDeviceT >::custom_static_type_ptr("uninitialized"); } // sc::result react( ev_init const& ev) // { return this->template transit< initialized< SocketDeviceT > >(); } }; } } #endif // JM_FFCP_UNINITIALIZED_H Any hints? regards, Oliver

6 10

Problem with fs::path initialization in windows
by Timo 15 Mar '07

15 Mar '07

Hello, I have a little problem with initialization of boost filesystem path. I'm working on Cygwin in Windows XP. The fs::initial_path() and fs::current_path() both work ok and show something like: C:\cygwin\boost But when I try to use a path of my own, say: fs::path full_path( "C:\\cygwin\\boost" ); I get thrown with an exception. Same deal with the rest of the combinations I can think of: C:/cygwin/boost c:/cygwin/boost. c/cygwin and /cygwin don't throw but the paths are not found. So how do initialize the fs::path with a Windows type of directory path? Timo

4 6

serialization gives memory leaks due to throwing constructors
by gast128 15 Mar '07

15 Mar '07

Dear all, I ran into an exception thrown by the constructor. But there is no way to call 'delete_created_pointers' on the object since the object is never fully created (see below) Is there something to do about? void Foo() { std::vector<Bus> v(2); //from the examples std::stringstream sstr; //save try { _ASSERT(sstr.good()); boost::archive::xml_oarchive oa(sstr); oa << BOOST_SERIALIZATION_NVP(v); } catch (boost::archive::archive_exception& re) { DBG_TRACEN(re.what()); } //load v.clear(); //sstr.seekg(0); sstr.seekg(1); //generate dileberate error boost::archive::xml_iarchive* p = NULL; try { _ASSERT(sstr.good()); boost::archive::xml_iarchive ia(sstr); p = &ia; ia >> BOOST_SERIALIZATION_NVP(v); } catch (boost::archive::archive_exception& re) { DBG_TRACEN(re.what()); //won't work: p is always NULL if (p) { p->delete_created_pointers(); } } } This throwing c++ constructor stuff is btw the way to transport errors according to a lot of c++ gurus. But my persnonal opinion is that it sucks. Every created object must be guarded with exception safety. For large applications, this exception mecahnism is very tricky and sources for a lot unwanted application terminations or memory leaks. Wkr me

5 24

Compatible question with STLPORT
by yunxig 15 Mar '07

15 Mar '07

Dear sir, When I include STLPORT in VS 2005 as follows: VC++ Directories -> Include files ->STLPORT-5.1.2\stlport This "Options" is set in the first line. However there's compiling errors: boost\include\boost-1_33_1\boost\config\auto_link.hpp(175) : fatal error C1189: #error : "Build options aren't compatible with pre-built libraries". When I delete STLPORT-5.1.2 from "Options", there's no compiling error. But why? Thanks.

2 3

help replace_all_regex and format_literal
by llwaeva＠21cn.com 15 Mar '07

15 Mar '07

Hi there, I found that the behavior of replace_all_regex is weird while specifing format flag as format_literal. For example, I would like to replace %f and %g in a source with some text respectively. The code is string source = "#f = %f \r\n #g = %g \r\n"; string text1("<footi(a)tv.com>"); string text2("""name""<name(a)email.101.pi>"); replace_regex(source, boost::regex("%f"), text1, boost::regex_constants::format_literal); replace_regex(source, boost::regex("%g"), text2, boost::regex_constants::format_literal); The result is fine. However, I would like to use replace_all_regex instead. i.e. replace_all_regex( source, boost::regex("(%f|%g)"), string("(?1 loofon(a)56.com)(?2 ""name""<name(a)email.101.pi>"), boost::regex_constants::format_literal); However, because of format_literal, what I get is #f = (?1 loofon(a)56.com)(?2 "name"<name(a)email.101.pi>") #g = (?1 loofon(a)56.com)(?2 "name"<name(a)email.101.pi>") rather than #f = loofon(a)56.com #g = "name"<name(a)email.101.pi>" In my case, I need to handle a heap of patterns and using replace_all_regex is convenient. Any way to use both replace_all_regex and format_literal without going into the trouble mentioned above? Thanks in advance.

3 2

[Date_Time] and [Filesystem] - compile error when including headers in wrong order
by KLINIK Markus SDI2-G (AREVA NP GmbH) 15 Mar '07

15 Mar '07

Hello, including the header files "boost/date_time/local_time/local_time.hpp" and "boost/filesystem/operations.hpp" in this order results in the following compile error: 'boost::equivalent' : use of class template requires template argument list Swapping lines 1 and 2 of the following sample code works. I'm using MSVC 7.0, M$ Windows XP and boost-1.33.1 <code> #include <boost/date_time/local_time/local_time.hpp> // 1 // #include <boost/filesystem/operations.hpp> // 2 // int main(void) { return 0; } </code> Best Regards Markus

1 0

Boost.Numeric.Interval: Please specify rounding control mechanism.
by Andreas Fabri 15 Mar '07

15 Mar '07

Hello, even with the release candidate for 1_34 I get this error message c:/Boost/include/boost-1_34\boost/numeric/interval/hw_rounding.hpp(38) : fatal error C1189: #error : Boost.Numeric.Interval: Please specify rounding control mechanism. on AMD Turion with XP 64 and VC 2005. Can anybody fix this or give a hint what I have to find out on mt platform to fix it. thanks, andreas

3 5

[mutable_queue] Mutable Queue problem
by Anton A. Patrushev 15 Mar '07

15 Mar '07

Hi, I wonder if there is a way to put edge descriptors into Mutable Queue? So far it looks like it can handle only vertex descriptors. Am I right? Thanks, Anton.

2 1

Re: [Boost-users] SQL library with full embedded SQL-syntax
by Thorsten Ottosen 15 Mar '07

15 Mar '07

Maciej Sobczak wrote: > Hi, > > Jeff Garland wrote: > > >>Well, I'll be brutally honest -- I don't think I would go that >>direction. We didn't pick a database project last year because a >>majority of the mentors agreed that the 'dsl-based' approach was more of >>a toy than reality. The main reason for this is that in 'real-world' >>applications queries are often dynamic and must be built at run-time. > > > The other reason is that in 'real-world' you might want to call stored > procedures instead of manually glued pieces of ad-hoc SQL queries, in > which case the support in the area of query composition is probably not > very useful anyway. In the 'real-world' both things are used. Choosing C++ for a projects where databases are involved is probably rare and Java/C#/Python/Ruby is more likely. C++ doesn't exactly boost productivity, so it has to offer something else to be a condidate. And what can C++ offer? Type safety. -Thorsten

1 0