- Boost-users - lists.boost.org

[BGL] Dijkstra algorithm help!!!
by Alejandro Marcos Aragón 09 Mar '07

09 Mar '07

Hi, I would like to know if anyone knows about a simplified version of the Dijkstra algorithm code. The one in the library is too difficult for me to understand and modify since my knowledge about templates is limited. I need to modify it somehow to make it do something I want and I don't think I can use the visitor concept to do this. I posted a message before, where I explained that I was trying to use a visitor to solve my problem but I didn't get anywhere. This is what I'm trying to do: I want to store all the distances for a vertex i, not only the smallest distance. In the current algorithm, when an edge is relaxed (meaning that the distance from the source is less than the one before), the distance is overwritten in the DistanceMap. I would like to keep all the distances with their respective vertices. Therefore, I thought that a priority queue (min heap) would be the right data structure to use. However, I don't know how to access to this information using the visitor. Imagine a graph like this one: 3 B------C | / | / 2| /6 | / | / A Imagine that the A is the source and it finds first that the distance to C is 6 (A-C edge). Then, when it finds that a lower distance is obtained through A-B and B-C (5), I would like to keep the previous result as well in a priority queue. For vertex C the priority queue looks then like this: (5,B) / / / (6,A) I would appreciate any suggestions. Alejandro

2 2

[state_saver] An extra constructor and a swapping destructor?
by Niels Dekker - mail address until 2007-11-30 08 Mar '07

08 Mar '07

The state_saver class at www.boost.org/boost/state_saver.hpp appears to be very useful for me. But I would like to propose two changes: First of all, I would like to have an extra constructor that would allow specifying a new value for the object (whose previous value will be restored when state_saver is destructed). Such a constructor would be very convenient, and it would also be conforming to io_state_saver's classes, by Daryle Walker. www.boost.org/libs/io/doc/ios_state.html A possible implementation would be: state_saver( T & object, const T & new_value ) : previous_value(object), previous_ref(object) { previous_ref = new_value; } Secondly I would suggest using a swap instead of an assignment, when restoring the previous value (when state_saver is destructed). A swap is much more likely to be fail-safe and efficient than an assignment. So I would propose the destructor to look as follows: ~state_saver() { // Doing "using std::swap", as suggested by Scott Meyers, // Effective C++ 3rd Edition, item 25, "Consider support // for a non-throwing swap" using std::swap; swap(previous_ref, previous_value); } Will these changes generally be considered an improvement? I already wrote Robert Ramey, the owner of state_saver.hpp. He also recommanded posting a message on this subject. So please let me know what you think! -- Niels Dekker http://www.xs4all.nl/~nd/dekkerware C++ programmer at LKEB, Leiden University Medical Center

1 0

BOOST_CHECK_PREDICATE vs BOOST_WARN_PREDICATE
by Alex Flint 08 Mar '07

08 Mar '07

The documentation page for the PREDICATE testing macros gives quite different syntax for the WARN vs CHECK methods: BOOST_WARN_PREDICATE( moo, (12)(i)(j) ); BOOST_CHECK_PREDICATE( std::not_equal_to<int>(), 2, (i,17) ); Since BOOST_WARN_PREDICATE and BOOST_CHECK_PREDICATE are only slight variants of one another, I'm confused as to why the syntax differs so much. Can anyone shed some light on this for me? Thanks, Alex

2 3

program_option error
by Aljaz 08 Mar '07

08 Mar '07

What am I doing wrong? I get error 'boost::program_options::typed_value<T> *boost::program_options::value(T *)' : expects 1 arguments - 0 provided class options { public: unsigned int _option1; int _option2; void parse_options() { boost::program_options::options_description config; config.add_options() ("set1", boost::program_options::value<&_option1>()->default_value(200), "seting 1") ("set2", boost::program_options::value<&_option2>()->default_value(300), "seting 2"); } }; Thanks for help

4 8

Re: [Boost-users] [boost::wave] Finding out if a macro is defined
by Ovanes Markarian 08 Mar '07

08 Mar '07

Hartmut, thanks for your help, but this issue is still alive ;) Just to be sure I did everything right, I took the entire head revision of boost and compiled it. I had no problems in the relevant libs. I also saw your changes in the cpp_context.hpp. The only problems I still get compiler errors using the flexi_string as described before. May usage scenario: token_type::string_type str("some string"); if(ctx.is_defined_macro(str)) { ... } There is still the iterator derefernece compiler error in the cpp_macromap.hpp: e:\boost\include\boost-1_35\boost\wave\util\cpp_macromap.hpp(396) : error C2228: left of '.get_value' must have class/struct/union type is 'const char' e:\boost\include\boost-1_35\boost\wave\cpp_context.hpp(198) : see reference to function template instantiation 'bool boost::wave::util::macromap<ContextT>::is_defined<const char*>(const IteratorT &,const IteratorT &)' being compiled with [ ContextT=boost::wave::context<token_groups::str_iter_type,token_groups::lex_iter_type>, IteratorT=const char * ] e:\projects\inpecting_wave\inpecting_wave\header_processor.cpp(178) : see reference to function template instantiation 'bool boost::wave::context<IteratorT,LexIteratorT>::is_defined_macro<boost::wave::cpplexer::lex_token<>::string_type>(const StringT &)' being compiled with [ IteratorT=token_groups::str_iter_type, LexIteratorT=token_groups::lex_iter_type, StringT=boost::wave::cpplexer::lex_token<>::string_type ] e:\boost\include\boost-1_35\boost\wave\util\cpp_macromap.hpp(396) : error C2228: left of '.c_str' must have class/struct/union the line: string_type name ((*it).get_value().c_str()); Still tries to call get_value on const char type. With Kind Regards, Ovanes

2 2

Re: [Boost-users] [enable_if] using enable_if to specialize std::numeric_limits?
by james.jones＠firstinvestors.com 08 Mar '07

08 Mar '07

From: "Ovanes Markarian" <om_boost(a)keywallet.com> > As I wrote before, you can specialize numeric_limits class with your custom > type: > > > template<class T> > struct my_numeric_limits_impl > { > static T min() { return boost::decimal::decimal_limits<T>::min(); } > }; > > namespace std > { > template<> > class numeric_limits<your_custom_type1> : public > my_numeric_limits_impl<your_custom_type1> > {}; > > template<> > class numeric_limits<your_custom_type2> : public > my_numeric_limits_impl<your_custom_type2> > {}; > > } > > > What is wrong on this approach? The only problem I see here is the the > customer must include the specialization of numeric limits. But this can be > easily solved if you put in the same header where your_custom_type is > defined. So everyone using yyour_custom_type will automatically include > numeric_limits and if needed use the correct version. I have no problem putting this in a standard include file. The problem I have is that I'm trying to make it as easy as possible for users to add their own models of my decimal type concept. I'd like for the steps to be as simple as: (a) define a new model and then (b) specialize is_decimal_type. So far I've been able to use (b) to drive lots of other functionality, and I'm frustrated that I can't get this to help with std::numeric_limits. But it sounds like there may be no perfect solution here. Here's an example of how enable_if *has* helped me in this project. I'm defining a bunch of operators, e.g. operator+, which need to take arguments only of is_decimal_type objects. I can easily do this: template<typename LHSType, typename RHSType> typename enable_if_c< is_decimal_type<LHSType>::value || is_decimal_type<RHSType>::value, typename result_type_2<LHSType, RHSType>::type>::type operator+(const LHSType& lhs, const RHSType& rhs) { typedef typename result_type_2<LHSType, RHSType>::type result_type; return result_type(operation_t::ADD, lhs, rhs); } Works great, and doesn't interfere with other specializations/overloads of operator+. - James Jones Administrative Data Mgmt. Webmaster 375 Raritan Center Pkwy, Suite A Data Architect Edison, NJ 08837

1 0

[bind] [lambda] and [function] in combination with each other
by Christoph Borgolte 08 Mar '07

08 Mar '07

I'm using a map of comparing functions, created with Boost.Lambda and Boost.Function. These are of type boost::function<bool(double, double)> f = ( _1 <= _2 ) Now I need one of these functions bound to a specific value to send this function object as a callback function. Things are compiling fine, but the result is wrong. I've boiled it down to a minimal example: Any hints? TIA Chris --------8<--------------8<------------ #include <iostream> #include <map> #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/function.hpp> #include <boost/bind.hpp> typedef boost::function<bool(const double& ,const double& )> CMP_FNC_DBL_DBL; typedef boost::function<bool(const double& )> CMP_FNC_DBL; typedef std::map<std::string, CMP_FNC_DBL_DBL> FUNCTION_MAP; int main() { FUNCTION_MAP fm; fm["=="] = (boost::lambda::_1 == boost::lambda::_2); fm["!="] = (boost::lambda::_1 != boost::lambda::_2); //... CMP_FNC_DBL_DBL fdd = fm["=="]; std::cout << fdd(11., 10.) << std::endl; //works //Getting a function object via bind CMP_FNC_DBL_DBL fdd2 = boost::bind(&FUNCTION_MAP::operator[] , fm, std::string("==")); std::cout << fdd2(11., 10.) << std::endl; //works not //that is what I really need: a CMP_FNC_DBL_DBL bound to a constant value CMP_FNC_DBL f = boost::bind(boost::bind(&FUNCTION_MAP::operator[] , fm, std::string("==")), 4.); //doesnt compile std::cout << f(3) << std::endl; //works not } --------8<--------------8<------------ result: 0 1 1

2 2

Compilation issue on MSVC 8 for mobile devices
by Silvatar Flood 08 Mar '07

08 Mar '07

When I try to compile boost 1.33.1 using MSVC 8.0 (VS.net 2005) for a mobile device, I get the following error: H:\PROJECTS\boost_1_33_1\boost/config/compiler/visualc.hpp(133) : fatal error C1189: #error : unknown CE compiler I tried to work around it by changing: # if defined(UNDER_CE) to # if 0 and it seems to work for the most part, except that it looks for certain intrinsic functions (_InterlockedIncrement for example) that apparently aren't available on Windows CE (though the platform API does have them, but not the compiler intrinsic versions). I could probably hack my way around this problem too, but I decided to wait and first ask if there's a more "proper" way to solve this issue. Any ideas? Yoni Lavi ____________________________________________________________________________________ Sucker-punch spam with award-winning protection. Try the free Yahoo! Mail Beta. http://advision.webevents.yahoo.com/mailbeta/features_spam.html

2 1

Question about ifstream ranges
by Pfligersdorffer, Christian 08 Mar '07

08 Mar '07

Hello all! I face the problem of partitioning a file that i read with ifstream into portions that are to be processed by some clients. I want to ensure that a client only views a specific part (window, range) of the stream. So far I store 2 streamoffsets posStart and posEnd, call seekg on the stream, pass it to the client by ref and rely on the client to check when it reached posEnd. However, I certainly cannot copy the whole range into a new stream object and pass that one. Efficiency does matter. Any suggestions? Maybe some iterator-magic can be done? Thanks in advance, -- Christian Pfligersdorffer Dipl.-Ing. Bakk.rer.nat. Software Engineering EOS GmbH - Electro Optical Systems Robert-Stirling-Ring 1 D-82152 Krailling / Munich Phone +49 89 893 36-229 Fax +49 89 893 36-287 christian.pfligersdorffer(a)eos.info http://www.eos.info

2 1

Re: [Boost-users] [enable_if] using enable_if to extend std::numeric_limits?
by james.jones＠firstinvestors.com 07 Mar '07

07 Mar '07

From: "Ovanes Markarian" <om_boost(a)keywallet.com> > Ok, I see. This is a bit more complex. ;) > Can you influence a type passed to numeric_limits? If yes you can do this: Thought about that, but it's suboptimal - see my response to me22 for more on this. Basically, I'd like any existing usage of numeric_limits<double> to be able to be replaced by numeric_limits<DecimalType> without any other changes. It wouldn't be *too* bad to do this: template< class T, class Enable = void > decimal_limits {}; template <> decimal_limits <double> : public std::numeric_limits <double> {}; template <> decimal_limits <float> : public std::numeric_limits <float> {}; template <class T> decimal_limits <T, enable_if<is_decimal_type<T> > > { // my decimal_limits implementation }; However, this would require users of my library to replace all references to numeric_limits with decimal_limits, so this is also less than ideal. Suppose the user has the following code: template<typename T = double> int f(const T& x) { x = abs(x); int i = 0; while(x > std::numeric_limits<T>::epsilon()) ++i; return i; } Ideally, the user should be able to do this: f<new_decimal_type> (some_new_decimal_value); and it should work with no modifications to the original function. - James Jones Administrative Data Mgmt. Webmaster 375 Raritan Center Pkwy, Suite A Data Architect Edison, NJ 08837

1 0