- Boost-users - lists.boost.org

Usage of iterator_facade and iterator_adaptor
by Peng Yu 02 Oct '07

02 Oct '07

Hi, I'm trying to use Boost.Iterator library. I understand that iterator_adaptor specialize some part of iterator_facade. However, I can not tell for what application I should use iterator_facade or iterator_adaptor. I think the manual should summarize this for the user. Can somebody give me some hint on which to use them and under what condition, if you have the experience? Thanks, Peng

2 1

[boost::enable_shared_from_this] in inheritance tree throws bad_weak_ptr exception
by dariomt＠gmail.com 02 Oct '07

02 Oct '07

I need to create a class hierarchy where every level inherits from boost::enable_shared_from_this. I am having problems caused, I believe, by having more than one copy of enable_shared_from_this. I am using boost 1.34.1 with VS2003 and VS2005, with and without ICL9.1, always getting the same result: a boost::bad_weak_ptr exception. #include <boost/shared_ptr.hpp> #include <boost/enable_shared_from_this.hpp> using namespace boost; struct A : enable_shared_from_this<A> { shared_ptr<A> getA() {return enable_shared_from_this<A>::shared_from_this();} }; struct B : A, enable_shared_from_this<B> { shared_ptr<B> getB() {return enable_shared_from_this<B>::shared_from_this();} }; int main() { shared_ptr<A> pA (new A); pA->getA(); // this works OK shared_ptr<B> pB (new B); pB->getB(); // throws boost::bad_weak_ptr() !? return 0; } question 1: Is this expected or am I doing something wrong? question 2: If it is expected: is it possible to have it explained in the documentation? question 3: Is there a way to accomplish this (inheritance with shared_from_this in all levels) using enable_shared_from_this? Thanks in advance

3 3

[Process] getcwd memory leak in context.hpp
by Chris Fairles 02 Oct '07

02 Oct '07

Valgrinding the single_process_tutorial example, it seems getcwd in context.hpp is called with a NULL buffer and 0 size meaning that space is automagically malloc'd (according to POSIX.1-2001). ==10383== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 5 from 1) ==10383== malloc/free: in use at exit: 64 bytes in 1 blocks. ==10383== malloc/free: 32 allocs, 31 frees, 22,225 bytes allocated. ==10383== For counts of detected errors, rerun with: -v ==10383== searching for pointers to 1 not-freed blocks. ==10383== checked 162,760 bytes. ==10383== ==10383== 64 bytes in 1 blocks are definitely lost in loss record 1 of 1 ==10383== at 0x4A0590B: realloc (vg_replace_malloc.c:306) ==10383== by 0x39848C03FC: getcwd (in /lib64/libc-2.5.so) ==10383== by 0x407894: boost::process::basic_work_directory_context<std::string>::basic_work_directory_context() (context.hpp:85) ==10383== by 0x407A19: boost::process::basic_context<std::string>::basic_context() (context.hpp:142) ==10383== by 0x403730: main (single_process_tutorial.cpp:44) ==10383== ==10383== LEAK SUMMARY: ==10383== definitely lost: 64 bytes in 1 blocks. I think the memory should be free()'d if the returned buf isnt NULL (or create a buffer isntead of passing NULL,0). Chris

1 0

identical boost expression used in different parts of the code - looking for an alias
by e r 02 Oct '07

02 Oct '07

hi for example, i use boost::bind( &A::f, _1 ) in different parts of my code. how can i define an alias for this expression? i'm thinking of using something like some_type* fo = new boost::bind(&A::f,_1); and use *fo in place of boost::bind(...). is this crazy? what should some_type be?

3 3

boost::shared_ptr and "this"-pointer: how to use shared_from_this?
by Taneli R. 02 Oct '07

02 Oct '07

I'm converting pipes and filters architecture made with raw pointers to use shared_ptrs. My Pipe class has a static counter, which gives unique pipeID to every pipe. Since Pipe class has pure virtual functions, the pipes in my architecture are instances of derived classes. Pipe class looks like this: class Pipe { public: Pipe() : pipeID(++pipeCount) {} virtual bool lay(boost::shared_ptr<Pipeable>& source, string& sourceTag, boost::shared_ptr<Pipeable>& destination, string& destTag) = 0; /*...other pure virtual functions here...*/ protected: int pipeID; static int pipeCount; } My problem is that when I try to make a new pipe between source and destination, the pipe is made twice (with different pipeID-numbers), which is a problem because there should be only one pipe connecting my components. I believe it is a result from using "this"-pointer directly without enabling "shared_from_this". I know that I should use it somehow but the same time I should maintain the unique pipeID. If I create a new shared_ptr with "shared_from_this" enabled inside SimplePipe, it has a different pipeID than the one I should be using. I have to use shared_ptr because I'm storing pointers into containers. And finally I cannot instantiate Pipe directly from its base class with using "shared_from_this" because it has pure virtual functions. Any help for solving this problem or a decent tutorial on using shared_ptrs and enable_shared_from_this would be appreciated. The following code snippets illustrate the problem: // //Main // boost::shared_ptr<Pipe> pipe; pipe = inputPipes[tag]; //find a matching pipe from the map if (pipe.get() == 0) //if not found, create a new pipe { boost::shared_ptr<Pipeable> source; // no source boost::shared_ptr<Pipeable> destination (boost::dynamic_pointer_cast<Pipeable>(comp)); pipe = boost::shared_ptr<Pipe>(new SimplePipe(source, srcTag, destination, destTag)); //creating new pipe inputPipes[tag] = pipe; //put the pipe into a map } // //SimplePipe, derived from Pipe // SimplePipe::SimplePipe(boost::shared_ptr<Pipeable>& source, string& sourceTag, boost::shared_ptr<Pipeable>& destination, string& destTag) { lay(source, sourceTag, destination, destTag); } bool SimplePipe::lay(boost::shared_ptr<Pipeable>& source, string& sourceTag, boost::shared_ptr<Pipeable>& destination, string& destTag) { if (source.get() != 0 && destination.get() != 0 && !sourceTag.empty() && !destTag.empty()) //Both ends attached { if (source->makeSource(pipeID, sourceTag)) { this->source = source; //problem? if (destination->makeDestination(pipeID, destTag)) { this->destination = destination; //problem? return true; } else { source->detachPipe(pipeID); } } } else if (source.get() == 0 && destination.get() != 0 && !destTag.empty()) //only destination attached { if (destination->makeDestination(pipeID, destTag)) { this->destination = destination; //problem? return true; } } else if (source.get() != 0 && destination.get() == 0 && !sourceTag.empty()) { /*etc etc...*/ } return false; } -- View this message in context: http://www.nabble.com/boost%3A%3Ashared_ptr-and-%22this%22-pointer%3A-how-t… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

[boost::graph] creating an external map for depth_first_search
by v r 02 Oct '07

02 Oct '07

I cannot figure out how to create a color map for depth first search, i`m using bundled properties and the following type as graph. adjacency_list<listS,setS,bidirectionalS,WorldNode> world_graph_t; right now i`m trying to make it like this, but i get 50 errors and cannot deduce what my error is. template<class VisitorT> THOR_INLINE void DepthFirstSearch(const VisitorT& vis, const vertex_descriptor& start) { using namespace boost; std::vector<default_color_type> color_vec( num_vertices(graph) ); depth_first_search(graph, visitor(vis), color_map(make_iterator_property_map(color_vec.begin(),get(vertex_color,graph))), start); } --------------------------------- Вы уже с Yahoo!? Испытайте обновленную и улучшенную. Yahoo! Почту!

3 2

Bjam linking problem
by desai3＠cooper.edu 02 Oct '07

02 Oct '07

Hi folks: I have successfully wrapped up a C++ program in Python. Now I want to give my libboost_python.so and server.so files to my friend, so that he can use it too. However server.so makes use of libstdc++ and ssl library files, which I have to link statically to the server.so in order to be able to give server.so to someone. However Bjam is not linking libstdc++.a or libssl.a to server.so statically. I will really appreciate if someone could give me any direction about how to link these libraries statically to server.so . Thank you very much. Maulik P.S. My Jamfile looks like this ----------------------------------------------------------------------- project-root ; import python ; extension server : server.cpp <template>@boost/libs/python/build/extension : <include>. <include>../dqp <include>/export/apps/dq/dqweb/include <library-path>/home/o066144/work_in_progress/api/SunOS-5.8-g++ <library-path>/export/apps/dq/dqweb/lib <library-path>/usr/local/lib <find-library>ssl <find-library>crypto <find-library>dl <find-library>socket <find-library>nsl <find-library>gcc <find-library>stdc++ ; ---------------------------------------------------------------------------

2 2

BGL: traveling salesman problem
by Michael Held 02 Oct '07

02 Oct '07

hi bgl-guys, is there a 'traveling salesman' implementation for bgl or is anything in this direction planned? there are several algorithms available, e.g. http://www.tsp.gatech.edu/concorde.html which could be ported to bgl?! best regards, michael

1 0

dumb question about boost::any.
by chun ping wang 02 Oct '07

02 Oct '07

How would i go about implementing operator< for boost::any....

1 0

[program_options] Why is allow_unregistered not implemented for config file?
by Marty Fried 01 Oct '07

01 Oct '07

I want to use the program_options library for my app's, well, program options. There is one set of options, which will usually only be one or two lines, that will have an unknown name; this is only handled by the (boost::program_options::) basic_command_line_parser template, with, apparently, no equivalent functionality for config files. This seems to me to be an oversight. I would have thought there would at least be an option to ignore unknown parameters if they aren't supported like for the command line. For my needs, it looks like I might need to modify the Boost sources to disable the exception on unknown parameters, or force the user to create a special config file for, most likely, a single parameter. Or I can change the syntax to something much less natural. What I'm hoping for is either someone to inform me that I'm wrong, and there is a way to handle config files more elegantly than what I'm seeing, or a suggestion as to how best to implement what I need. Thanks for any help you can offer. -- Marty Fried - email: public.forums (at) gmail.com Left coast, USA

2 2