July 2003 - Boost-users - lists.boost.org

boost::shared_ptr and stl algorithms
by Bill Eidson 15 Apr '05

15 Apr '05

std::vector< boost::shared_ptr<T> > TVec; I want to write a function object to run std::for_each. to do somework against members of T and accumulate results in functor. My attempts have failed in VC7 with memory errors in my operator() method. struct fo { bool operator()( boost::shared_ptr<T>& t ) { return true; } }; Thanks in advance bille

5 9

BGL and reversing / moving edges
by joshrose 17 Dec '03

17 Dec '03

G'day, I was hoping someone might be able to tell me if they've run into any situations using BGL where they needed to move (or reverse (i.e. flip the source and target)) an edge within a graph implemented as an adjacency_list. I'm working on an implementation of an algorithm for max. weight bipartite matching ("the assignment problem"), and one of the assumptions in the algorithm is that non-matched edges are directed from one subgraph, A, to the other subgraph, B, and matched edges are directed the other way. Over the course of growing the matching, a simple way to add and remove edges from the candidate set is just to reverse the edge, thus the need for some way to reverse the edge within BGL. I don't think I can just remove and then re-add the edge in the other direction, primarily because doing so will (I believe) invalidate any edge_descriptors referring to that edge. In particular, there are property maps keeping track of the weight of edges, the predecessors of certain nodes, etc. that use edge descriptors as keys. I've spent a fair bit of time reading the documentation and looking through the implementation of adjacency_lists (incl. edge_list_impl), and I'm not really sure of a way to either move an edge directly within a graph, or even to create a new edge from another s.t. they have the same edge_descriptor (in general a bad thing, probably). Anyone have any thoughts on this? Thanks, -Josh

4 10

tuples: operator ==
by Konstantin M. Litvinenko 12 Aug '03

12 Aug '03

Hello, All! I can't compile this simple test: --------------------------------------------- #include <boost/tuple/tuple.hpp> #include <iostream> using namespace boost::tuples; using namespace std; int main() { typedef tuple<int, tuple<int, int> > Tuple; Tuple t1(1, make_tuple(1, 1)); Tuple t2(2, make_tuple(2, 2)); if (get<1>(t1) == get<1>(t2)) { cout << "equal"; } else { cout << " not equal"; } return 0; } ----------------------------------------- What am I doing wrong? Compilers: Borland Builder 6 (bcc32.exe 5.6.4) and VC7 With best regards, Konstantin M. Litvinenko aka Dark Angel.

2 1

Linking with BGL adjacency_list_io.hpp
by Solomon Gibbs 05 Aug '03

05 Aug '03

Hi, I'm having trouble linking code that uses adjacency_list_io.hpp. If the header is included is included in more than one compilation unit, I get "multiple definition of function" errors for the non-template functions in adjacency_list_io.hpp: std::istream& operator >> ( std::istream& in, no_property& ) and void getSubset( no_property& p, const no_property& s ) I'm sure that there is some C+ subtlety that I'm missing here, but I'm not familiar enough with the library to understand what it is. Shouldn't non-template function code be in a .cpp file to keep this sort of thing from happening? - Solomon

3 5

A Quick Question about Threads
by Paul Grenyer 01 Aug '03

01 Aug '03

Hi All I'm planning on replacing all our (Win32 API Wrapped) threads with the boost threads. I was looking at the thread class interface and there isn't a method to ask if the thread is still running. This is something we use quite a lot in out current wrapper. To get around this I was thinking of using a member variable in the function object that is passed to the thread constructor to indicated whether the thread had finished. Is there a better way? I've not had a really good look at the docs yet, but there doesn't appear to be a Critical Section class. Should I be using the Mutex for synchronisation instead? -- Regards Paul Paul Grenyer Email: pjgrenyer(a)iee.org Web: www.paulgrenyer.co.uk EvaneScence: http://www.evanescence.com

6 15

Graph Library: Ordered Vertex List?
by dmatt001 01 Aug '03

01 Aug '03

Hi Guys, I would like to use an ordered vertex list (ie based on property type) in an implementation that I am working on such that duplication can be avoided. I have read and understand how to implement an ordered edge list thanks to the previous posts but am finding it hard to get started! Perhaps I have missed the key example? It appears that from the way config::VertList is constructed that it may be a similar process in nature? adj_list_impl::add_vertex appears to discard the boolean return value from the associative insertion. Would such a construction allow me to avoid the duplication I seek to avoid? I am guessing so, but some advice would give me a confidence "boost". ( <-- It's a lame pun: I really do need help ;-) ) My first steps are below rehashing the excellent ordered edges example with a simple search/replace and maching my target container base. I'll keep plugging along with it anyway and it may just compile one day. Suggestions/Ideas/Refs/Ptrs are appreciated. Matt --- first guess follows --- template <class VertDesc> struct order_by_name : public std::binary_function<VertDesc,VertDesc,bool> { bool operator()(const VertDesc& v1, const VertDesc& v2) const { // Order by vertex name. return boost::get(boost::vertex_name, v1) < boost::get(boost::vertex_name, v2); } }; #if !defined BOOST_NO_TEMPLATE_PARTIAL_SPECIALIZATION struct ordered_set_by_nameS { }; namespace boost { template <class ValueType> struct container_gen<ordered_set_by_nameS, ValueType> { typedef std::set<ValueType, order_by_name<ValueType> > type; }; } #else struct ordered_set_by_nameS { template <class T> struct bind_ { typedef std::multiset<T, order_by_name<T> > type; }; }; namespace boost { template <> struct container_selector<ordered_set_by_nameS> { typedef ordered_set_by_nameS type; }; } #endif

3 3

Compile time Problem with boost::type_traits library
by Amit Bhatia 01 Aug '03

01 Aug '03

Hello, I am trying to use the boost::function library which in turn includes many of the headers from other boost libraries. I am using g++ 3.2 with: g++ -c -g -fwritable-strings -Wno-deprecated -static -ansi -pedantic -DANSI_HEADERS -D_G_HAVE_BOOL as the options. The file is_integral.hpp and the file is_convertible.hpp in the type_traits library contain the declarations of the type long long which the ISO C++ seemingly doesn't accept. The compiler gives the warning: ISO C++ does not support `long long' . I don't think there is a new version of this on boost. I downloaded the boost lib sometime in June Version 1.30.0. Can this problem be because of any of the compiler options that I am using? thanking you, regards, amit.

3 2

Spirit Sequential-or
by Bruno Martínez 01 Aug '03

01 Aug '03

Hi boosters, In operators.html (a || b) is said to be equivalent to (a | b | a >> b), but, if I understand precedence correctly, it should be equivalent to (a >> b | a | b). I'm just learning Spirit, so this may probably be my mistake, right? Cheers, Bruno Martinez ---------------------------------------------------- ATENCION Su servicio de correo tiene antivirus ?? @internet URUGUAY: http://www.internet.com.uy

3 3

FC++ in Boost, Spirit, Preprocessing
by Jan Kybic 31 Jul '03

31 Jul '03

Hello, I want to write in favor of Boostifying the functional library FC++, which I find very useful. Would a merger with the Phoenix library from Spirit (already in Boost) be feasible? Phoenix seems to have a slightly nicer syntax but lacks (as far as I know) the Haskell prelude functions and lazy list implementation, also the closures do not seem very intuitive to me. My main concern with both FC++ and Phoenix is that when defining new functiods, the user needs to write a lot of often repetitive code. This could be simplified using a suitable preprocessor. Would that be acceptable and are there other people interested ? Yours, Jan -- ------------------------------------------------------------------------- Jan Kybic <kybic(a)ieee.org> tel. +420 2 2435 7264 or <kybic(a)fel.cvut.cz>, http://cmp.felk.cvut.cz/~kybic

3 3

date_parsing.hpp.patch
by Tonal 31 Jul '03

31 Jul '03

This is simple refactore and improved date_parsing.hpp begin 666 date_parsing.hpp.patch M*BHJ(&1A=&5?<&%R<VEN9RYH<' N;W)I9PE4:'4(a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a)9&5T86EL.CIP87)S95]D E871E/&1A=&5?='EP93XH=&]K*3L-"B @("!]#0H@( T*(" -"@`` ` end

1 0

Spirit Parsing Problem
by sps22 31 Jul '03

31 Jul '03

Hi, I am trying to use Spirit (in the latest Boost 1_30_0) for a project. I want to parse a file whose structure is the following: section1 { something; } I am using the file_iterator to iterate on the input file and then define comment as well as the grammar. I am at my wits end why my parser keeps failing. (When this simple grammar works, I want to make the parser more complex.) My code is attached below. (Note: I used the example in the distribution c_grammar.cpp as the starting point for it.) Thanks. -surinder PS: This input file structure is going to be, of course, more complicated than the example shown above. Inside the setions, I would like to pass on flags and statements to the code. If it is better for me to use another approach to parse this file structure, please suggest me. #include <iostream> #include <fstream> #include <vector> #include <boost/spirit.hpp> /* Input file for this looks like: file { something; } */ ////////////////////////////////////////////////////////////////////// ///////// // used namespaces using namespace boost::spirit; using namespace std; ////////////////////////////////////////////////////////////////////// ///////// // parsing helper function // Here's our comment rule struct skip_grammar : public grammar<skip_grammar> { template <typename ScannerT> struct definition { definition(skip_grammar const& /*self*/) { skip = space_p | comment_p("//") // C++ comment | comment_p("/*", "*/") // C comment | comment_p("#") // added for correctly ; } rule<ScannerT> skip; rule<ScannerT> const& start() const { return skip; } }; }; ////////////////////////////////////////////////////////////////////// ///////// // typedef typedef char CharT; typedef file_iterator <CharT> IteratorT; typedef scanner <IteratorT> ScannerT; typedef rule <ScannerT> RuleT; template<typename GrammarT> void parse(GrammarT const& g, char const* filename) { ifstream in(filename); if (!in) { cerr << "Could not open input file: " << filename << endl; return; } IteratorT start(filename); if (!start) { cerr << "Could not open the file " << endl; return ; } IteratorT end = start.make_end(); skip_grammar skip; parse_info<IteratorT> result = parse(start, end, g, skip); if (result.full) cerr << filename << " Parses OK" << endl; else { cerr << filename << " Fails Parsing" << endl; for (int i = 0; i < 50; i++) { if (result.stop == end) break; cerr << *result.stop++; } cerr << endl; } } struct c_grammar : public grammar<c_grammar> { template <typename ScannerT> struct definition { definition(c_grammar const& /*self*/) : SEMICOLON(';'), COMMA(','), COLON(':'), ASSIGN('='), LEFT_PAREN('('), RIGHT_PAREN(')') { // C keywords keywords = "file", "physics", "math"; // C operators LEFT_BRACE = chlit<>('{') | strlit<>("<%"); RIGHT_BRACE = chlit<>('}') | strlit<>("%>"); LEFT_BRACKET = chlit<>('[') | strlit<>("<:"); RIGHT_BRACKET = chlit<>(']') | strlit<>(":>"); // Tokens FILE = strlit<>("file"); PHYSICS = strlit<>("physics"); MATH = strlit<>("math"); // string literals STRING_LITERAL_PART = lexeme_d[ !chlit<>('L') >> chlit<>('\"') >> *( strlit<>("\\\"") | anychar_p - chlit<>('\"') ) >> chlit<>('\"') ] ; STRING_LITERAL = +STRING_LITERAL_PART; // Rules file_statement = FILE >> LEFT_BRACE >> *(anychar_p) >> RIGHT_BRACE; // does not work // file_statement = *(anychar_p); // parses OK translation_unit = file_statement; } symbols<> keywords; chlit<> SEMICOLON, COMMA, COLON, ASSIGN, LEFT_PAREN, RIGHT_PAREN; rule<ScannerT> LEFT_BRACE, RIGHT_BRACE, LEFT_BRACKET, RIGHT_BRACKET; rule<ScannerT> FILE, PHYSICS, MATH, STRING_LITERAL, STRING_LITERAL_PART; rule<ScannerT> file_statement, translation_unit; rule<ScannerT> const& start() const { return translation_unit; } }; }; ////////////////////////////////////////////////////////////////////// ///////// // main entrypoint int main(int argc, char* argv[]) { // Start grammar definition cerr << "C Grammar checker implemented with Spirit ..." << endl; // main driver code c_grammar g; if (2 == argc) parse(g, argv[1]); else cerr << "No filename given" << endl; return 0; }

1 0

Re: [Boost-Users] Newbie shared_ptr in a container question.
by Joshua Little 31 Jul '03

31 Jul '03

>From: "Duane Murphy" <duanemurphy(a)mac.com> >Reply-To: Boost-Users(a)yahoogroups.com >To: "Boost Users" <Boost-Users(a)yahoogroups.com> >Subject: Re: [Boost-Users] Newbie shared_ptr in a container question. >Date: Sun, 27 Jul 2003 09:17:48 -0700 > >--- At Sat, 26 Jul 2003 19:57:17 +0000, Joshua wrote: > > >I've been reading through posts on here and in the docs but I think I may > >have missed something. Heres what I am trying to do: > >lets say I have a class Foo* and I'm trying to use it in a > >std::vector<shared_ptr<Foo> >. I can add them all day long and it works > >fine. The problem is that when I try to remove the shared_ptr from the > >vector. I expected it to delete the memory for the Foo* then delete > >itself (to completely free the memory that the shared_ptr used and the > >memory that the Foo was using. I think this is not happening though since > >the destructor is not being called when the elements are removed from the > >vector. Heres how I am currently inserting the shared_ptr's into the > >vector (yes I know this is the wrong way now that I've read a few of the > >posts here): > > > >typedef std::vector<boost::shared_ptr<Foo> > FooVector; > > > >FooVector Foobar; > >Foo* blah = new Foo(); > >boost::shared_ptr<Foo> * NewPointer = new boost::shared_ptr<Foo>(blah); > >Foobar.insert(*NewPointer); //Yes I know this looks horrible. And its > >probably wrong. > > > > > >When I call Foobar.erase() I thought that the shared_ptr would be removed > >and the memory freed (and Foo's destructor called). I tried it like this >: > >FooVector::iterator CurrentFoo = Foobar.begin(); > >Foobar.erase(CurrentFoo); > > > >That did remove the shared_ptr but the Foo* did not seem to get > >deallocated, its destructor was never called. Some of the posts seem to > >indicate that this is a very bad way to let shared_ptr manager the > >memory, is there a better way to do what I want to do using shared_ptr or > >is shared_ptr the wrong tool to use? > >There have been many good suggestions as to what the correct sequence for >inserting the shared_ptr into the vector. I'm not sure the real problem >has been adequately explained. Why isnt the foo* being deleted. > >Shared_ptr<> counts the number of instances of the pointer that are >created. When you did: > >boost::shared_ptr<Foo> * NewPointer = new boost::shared_ptr<Foo>(blah); > >You created the first instance; count == 1. > >When you did: > >Foobar.insert(*NewPointer); //Yes I know this looks horrible. And its > >A copy of the shared_ptr<> was made. The count is now 2. > >When you did: > >Foobar.erase(CurrentFoo); > >One copy (of two) was "deleted". In this sense, the shared_ptr<> was >deleted. The destructor for the shared counter mearly decrements the >counter. Count == 1. > >The problem is that the original count was leaked by using "new >shared_ptr<>". This is why all the suggestions show that you should be >using the stack based method for shared_ptr<>. > >Shared_ptr is almost never allocated from the heap. There really is not >much point. It's use is as a value representation of pointer. > >I hope this helps you understand where your code went astray. > >Good luck, > ...Duane > > > > > ...Duane > Reading many of these replys I now see why what I was doing wasn't working, so I'm going to change my code around and hopefully get this cleared up. Thanks to everyone for your help :) _________________________________________________________________ Add photos to your e-mail with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail

2 1

Newbie: Way to convert "pointer-to-member-function" to a "pointer-to-function".
by Amit Bhatia 31 Jul '03

31 Jul '03

Hello All!, I am trying to figure out a way of converting "pointer-to-member-function" to a "pointer-to-function" so that I can pass it as a callback function pointer to constructor of some third party class. I know there is NO direct way of doing this in c++. I am further constrained because of the following facts: 1. The fact that I am using a third party class for whose constructor I need to pass pointers to functions having some specific signature. 2. I don't want to be using static declaration for the method that I wish to pass to the third party constructor as a callback function(due to the fact that static methods can access only access static data members). So if I am not mistaken, then, I suppose, that boost function library might be of help in here by somehow using some elegant tricks' to do this task. (There is the fact also that every instance of my class has lot of data which is different and I need access to the data for the method that I wish to pass). Am I right, or am I looking at the wrong place? If I am indeed looking at the wrong place then where will it be best to look for it. I remember some notes on binding in the documentation of the boost function library though. Thanks in advance, regards, amit.

1 0

Detecting None in Boost.Python
by ahakim777 31 Jul '03

31 Jul '03

Hi, This is a Boost.Python question. How do I detect that a Python function has passed a 'None' to a wrapped C++ function? Cheers, Ammar

2 1

filesystem::path & wide char strings?
by Cody Batt 30 Jul '03

30 Jul '03

Hello, I want to use the filesystem::path class for a project I'm working on, but I need to support wide character strings in the path names. I might be missing something, but it seems that path only supports std::string. Is this because of a portability issue or what is the rationale for not using something like basic_string<CharType>? Are there any plans to provide support for wide character strings in the future? Thanks, Cody Batt

2 1

howto map c++ namespaces using boost::python?
by cavok＠filibusta.crema.unimi.it 30 Jul '03

30 Jul '03

hi all, i'm a newbie user of Boost. i need some suggestions to expose c++ namespaces in python. in python i use classes as namespaces, but using class_ construct to build python classes seems to not be applicable to c++ namesapces. moreover, i have a global c++ object which i'd like to map onto python. it is not that clear to me how i can achieve this result. thanks domenico -----[ Domenico Andreoli, aka cavok --[ http://filibusta.crema.unimi.it/~cavok/gpgkey.asc ---[ 3A0F 2F80 F79C 678A 8936 4FEE 0677 9033 A20E BC50

2 1

Boost Bind on MSVC 7.1
by Dale Peakall 30 Jul '03

30 Jul '03

The following example is a very simple, and contrived, example that demonstrates the problem I'm seeing. This fails to compile with 'term does not evaluate to a function taking 1 arguments'. Is this a problem with bind on MSVC 7.1, or have I just forgotton how to use it/what it can do. - Dale. #include <iostream> #include <string> #include <boost/bind.hpp> int my_function(std::string const& value) { return static_cast<int>(value.length()); } int main() { std::string my_string("hello world!"); std::cout << boost::bind<int>(_1, my_string)(my_function); }

2 1

boost.python exception
by newwz2003 29 Jul '03

29 Jul '03

Can I raise exception in C++ module can catch it in python ? or where can I get detailed doc about boost.python? Thank you

2 1

Deferred Function Forwarding
by vivek_arora 29 Jul '03

29 Jul '03

I am looking for pointers on how to use boost to implement a set of template classes/helper functions which can be used for deferred function/method call forwarding. Here's an example of illustrating what I wish to accomplish. I can live with restictions on the function argument types. //Define a template type which can capture arguments of any (or most) functions. I believe boost tuple could suit my requirements. //Define a template helper function save_args which constructs a arguments tuple based on the signature of the function passed as the first argument. e.g . R f(T1, T2, T3); boost:tuple<R, T1, T2, T3> args = save_args(f, t1, t2, t3); //Don't want to explicitly specify types like save_args<R,T1,T2,T3> (t1,t2,t3) //Define another template helper function call_with_args which given another function "g" with a signature conforming with "f" and the argument tuple invokes "g" e.g. R g(T1, T2, T3) //This should call g(args.get<0>, args.get<1>, args.get<2>) call_with_args(g, args); Could someone provide ideas, pointers/sample code for implementing save_args, call_with_args helper template functions. Thanks -Vivek Arora

3 2

Feature Suggestion
by everwicked 29 Jul '03

29 Jul '03

Hello guys, I have just been introduced to the library by some people at Efnet's #c++ and I have to say there is quite a lot of useful work in there. I hope boost makes it into the standard. It's about time to get a feature rich std library. Want I wanted to ask you is to provide a few more functions in library interval. I was comparing boost with blitz++ as the parts I need at the moment are mainly of scientific nature. Anyways, I found boost more to my likeness because it has other features I might want to use later, because it supports more compilers and lastly because it's API is better organised in my opinion. The functions that I find that are missing from library 'interval' are the ones listed here: http://www.oonumerics.org/blitz/manual/blitz03.html#l77 Is there any chance you could implement them in boost? I wouldn't ask but I think that they are not a lot of work. Thanks in advance. Kind regards, George

3 2

boost_date_time: ftime ==>gettimeofday for microsec_clock?
by dick.bridges＠tais.com 28 Jul '03

28 Jul '03

I need a microsec_clock on Windows XP. My first impulse is to define gettimeofday() implemented using ftime(). Should this work ok or am I missing something obvious? TIA

1 0

Newbie - stack overflow in connected components
by Frank H 28 Jul '03

28 Jul '03

Hello: Environment: Windows 2K, MSVC 6 SP5, Boost 1.30.0 I'm new to BGL ( 2 days experience ) and am trying to use the connected components graph. My implementation seems to be OK for my small test case ( 10 verticies) but when I run my "real data" through it ( 15000 verticies) I get a stack overflow error. I note that there is a patch in the sourceforge CVS for the dfs algoritm which implements a non-recursive solution from dgregor that I understand will avoid the stack overflow. I naively downloaded the "depth_first_search.hpp" V1.35 from the CVS and of course it doesn't compile. The question: "How can I apply the patch to fix the stack problem( ie how to determine what and where are the other files I need or do I download the entire current version ( ie tag MAIN)?" Thanks for any help Frank

4 9

RE: [Boost-Users] Newbie missing a header for boost_date_time?
by dick.bridges＠tais.com 28 Jul '03

28 Jul '03

<history> <snip/> > ...which refers to lexical_cast.hpp > > template<typename Target, typename Source> > class lexical_stream > { > bool operator<<(const Source &input) > { > here ===> return stream << input; <==== here > } > } <snip/> I'm starting to think this may be related to the 1_30_0 lexical cast problem. Can you upgrade lexical_cast.hpp from cvs? You can download it from here: http://cvs.sourceforge.net/cgi- bin/viewcvs.cgi/boost/boost/boost/lexical_cast.hpp Jeff </history> I downloaded the head this (Monday) morning and replaced the 1_30_0 version. Compile error disappeared. (Sanity check: copied the 'old' version back and compile error re-appears.) Thanks for the help. That was driving me nuts.

1 0

link problems with precompiled headers and boost libs?
by co_tangens 28 Jul '03

28 Jul '03

Hi, I'm using the datetime library of boost 1_30_0 with VC7.1 with precompiled headers on. We're getting the following link warnings/errors: om.lib(bg_extra.obj) : warning LNK4006: ___@@_PchSym_@00@UwvevolknvmgUtnkEPozhgUhlfixvUytPzwwOlyq@ already defined in om.lib(bg_init.obj); second definition ignored om.lib(bg_extra.obj) : error LNK2005: ___@@_PchSym_@00@UwvevolknvmgUtnkEPozhgUhlfixvUytPzwwOlyq@ already defined in om.lib(bg_init.obj) We include boost headers in our own header files. Our own headerfiles are then used to create a precompiled header. It seems that all object files that use the precompiled header get these link errors... Does anyone know the reason/ a solution for this error? Thanks a lot!

1 1

spirit & tokenizer
by Andrey Sverdlichenko 28 Jul '03

28 Jul '03

Hello. Is there any "right way" to parse strings, already splitted with tokenizer? I wrote iterator that can work with this splitted strings, but i also need some "skip parser" to recognize token boundaries as whitespaces and can't design one. -- Best regards, Andrey mailto:blaze@lists.infosec.ru

4 7