- Boost-users - lists.boost.org

Equality comparison of boost::function<>
by Zachary Turner 02 Sep '09

02 Sep '09

Is this possible? I always get ambiguities when I try it. Here's some sample code: class foo { public: typedef enum { lost, reset } device_event; typedef boost::function<void (window&, device_event)> device_listener; void add_listener(const device_listener& listener) { std::list<device_listener>::const_iterator iter = std::find(listeners_.begin(), listeners_.end(), listener); if (iter == listeners_.end()) listeners_.push_back(listener); } private: std::list<device_listener> listeners_; }; The above is slightly simplified from my actual code, but when I do this I get the following: I'm using Visual C++ 9.0 SP1 1>f:\dev\testing\window.h(42) : error C2666: 'operator ==' : 4 overloads have similar conversions 1> q:\include\boost\1.39.0\boost\function\function_template.hpp(995): could be 'void boost::operator ==<R,T0,T1>(const boost::function2<R,T0,T1> &,const boost::function2<R,T0,T1> &)' [found using argument-dependent lookup] 1> with 1> [ 1> R=void, 1> T0=zg::window &, 1> T1=zg::window::device_event 1> ] 1> q:\include\boost\1.39.0\boost\function\function_base.hpp(791): or 'bool boost::operator ==<zg::window::device_listener>(Functor,const boost::function_base &)' [found using argument-dependent lookup] 1> with 1> [ 1> Functor=zg::window::device_listener 1> ] 1> q:\include\boost\1.39.0\boost\function\function_base.hpp(782): or 'bool boost::operator ==<zg::window::device_listener>(const boost::function_base &,Functor)' [found using argument-dependent lookup] 1> with 1> [ 1> Functor=zg::window::device_listener 1> ] 1> or 'built-in C++ operator==(void (__thiscall boost::function2<R,T0,T1>::dummy::* )(void), void (__thiscall boost::function2<R,T0,T1>::dummy::* )(void))' 1> with 1> [ 1> R=void, 1> T0=zg::window &, 1> T1=zg::window::device_event 1> ] 1> while trying to match the argument list '(const zg::window::device_listener, const zg::window::device_listener)' 1

7 13

[BGL] reducing memory requirements for adjacency list
by Soeren Meyer-Eppler 02 Sep '09

02 Sep '09

Hi, I'm using adjacency_list< vecS, vecS, bidirectionalS ... > extensively. I have so many graphs loaded at once that memory becomes an issue. I'm doing static program analysis and store the callgraph and flowgraphs of the disassembled binary in boost graphs. Thus I can have several ten thausand functions==flowgraphs and one gigantic callgraph. I'd really like to reduce memory usage for my graphs while still using the BGL. Since my graphs are static after loading and edge and vertex counts are known beforehand I see huge potential for optimization. For example, I'd like to allocate a single buffer for all vertices/edges of a single graph and let the graph just store indices into that buffer. more questions: 1) what's the memory overhead of using vertex and edge properties? I have quite a few of them. 2) is it possible to convince the BGL to use the shrink to fit idiom? As I understand it the adjacency lists use push_back to add edges. Is it possible to reduce memory usage by swapping the resulting vector with a copy of itself? Maybe by copying the whole graph? 3) Is it possible to use boost pool allocators with the BGL? As far as I can tell the BGL currently performs lots of small allocations - I'd really like to avoid that for space and runtime efficiency reasons. Did anyone already build a BGL version optimized for memory usage? Should I try using the existing graph structures and augment it with custom allocators or somesuch or is it more fruitful to write my own implementation and try to stay interface compatible with the BGL so I may continue to use it's algorithms? best regards, Sören

4 8

Template mailing list / forum
by Ryan McConnehey 02 Sep '09

02 Sep '09

What is a good template forum or mailing list so that I don't clutter the boost mailing list with non boost questions? Ryan

2 2

how to know which libraries to link?
by pedro chaparro 02 Sep '09

02 Sep '09

HI, im newbie with boost, i have a doubt about how to know which libraries u have to link when u compile a cpp file, in the example below for example which would be the libraries to link? i do for example c++ -I boostdirectory myfile.cpp -o myfile [ and here? ]. Im working with ubuntu, and boost 4.0 thanks // // timer.cpp // ~~~~~~~~~ // // Copyright (c) 2003-2008 Christopher M. Kohlhoff (chris at kohlhoff dot com) // // Distributed under the Boost Software License, Version 1.0. (See accompanying // file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) // #include <iostream> #include <boost/asio.hpp> #include <boost/date_time/posix_time/posix_time.hpp> int main() { boost::asio::io_service io; boost::asio::deadline_timer t(io, boost::posix_time::seconds(5)); t.wait(); std::cout << "Hello, world!\n"; return 0; } -- Ing Pedro Alonso Chaparro Valero Ciudad Politecnica de la Innovación iTEAM - Mobile Communications Group Polytechnic University of Valencia C\ Camino de Vera S/N, Edificio 8G 46022 Valencia, Spain

3 2

[Interprocess] assertion while destroying inexistent named instance
by Rui Pedro Fernandes 02 Sep '09

02 Sep '09

Hi, During my tests I found that if I change the managed shared memory index from the default iset_index to map_index the test fails when a process tries to destroy an already inexistent named instance, using its name. With map_index, an assertion fails when the object is found not to be present. Shouldn't it behave like in the case of iset_index ? Thank you. Rui

2 1

Re: [Boost-users] Boost and autocompletion
by dhruva 02 Sep '09

02 Sep '09

Hello, > On Tue, Sep 1, 2009 at 11:24 AM, Germ?n Diagowrote: > > try eclipse with eclipse cdt plugin. In www.eclipse.org you can find a > > bundle of "eclipse for c/c++ developers". > > That's the one you need. > > Yea, that is the best I have seen so far, but nothing gets anywhere > close to the quality of Visual Assist + MSVC sadly, that is my only > main tie to still programming on Windows, the IDE is just so much > better. Might see if it works in wine... Slightly off topic, GNU Global is pretty good in creating the good cross reference. I use it in my project with full boost being indexed. Emacs + Global is my development environment that works on both windoze and gnu/linux. I like the references feature of global which lists all potential locations where the symbol is referred (using pattern match only, same named functions appear even if it is totally unrelated to the symbol you are trying to explore) -dhruva Love Cricket? Check out live scores, photos, video highlights and more. Click here http://cricket.yahoo.com

2 1

Re: [Boost-users] [Spirit] Creating a vector of structs
by Lindley M French 02 Sep '09

02 Sep '09

> On Tue, Sep 1, 2009 at 4:38 PM, Lindley M French<lfrench1(a)gmu.edu> > wrote:> I'm playing around with Spirit for a particular parsing > task. I could do this manually, but I figured this was a good > chance to learn what Spirit had to offer. > > > > The trouble is this. While Spirit offers ready-made semantic > actions for parsing primitives, it doesn't seem to offer anything > terribly useful for compound statements. As far as I can tell, you > can define your own semantic actions but they simply take in the > matched string as an argument----there's no way to get at the > values parsed on that level. > > > > For example, let's say I have a type > > struct Person > > { > > ? string name; > > ? double height; > > ? int age; > > ? Person(string n, double h, int a) > > ? : name(n), height(h), age(a) > > }; > > > > And I have a document containing > > > > 5.5 15 Henry > > 4.6 32 Jane > > 6.1 21 Bob > > > > I could parse this using > > > > rule<> personrule = strict_real_p >> int_p >> ~eol_p; > > parse(str, *(personrule[WhatGoesHere] >> eol_p) >> end_p); > > > > But how can I build a vector<Person> out of this? I can write my > own functor to do the push_back, but I need to get at the parsed > values of name, age, and height for each personrule somehow. > > > > I suppose I *could* just run another parse() on the input the > semantic action for that specific person....but that really seems > like it shouldn't be necessary. > > You are using Spirit.Classic, you should use Spirit.QI (Boost trunk, > docs are at: > http://svn.boost.org/svn/boost/trunk/libs/spirit/doc/html/index.html) then you can do it easy, like so: > > BOOST_FUSION_ADAPT_STRUCT(Person > (double,height) > (int,age) > (string,name) > ) > > rule<YourIteratorType> personrule; > personrule %= double_ >> int_ >> raw[lexeme[*(char_-eol)]]; > std::vector<Person> theVec; > parse(str.begin(),str.end(), *personrule, theVec); > > This is also better to ask on the spirit mailing list. :) Unfortunately for the moment I'm stuck with Boost 1.35, which I don't believe has Spirit 2. I'll try asking on the Spirit list.

2 1

[Spirit] Creating a vector of structs
by Lindley M French 01 Sep '09

01 Sep '09

I'm playing around with Spirit for a particular parsing task. I could do this manually, but I figured this was a good chance to learn what Spirit had to offer. The trouble is this. While Spirit offers ready-made semantic actions for parsing primitives, it doesn't seem to offer anything terribly useful for compound statements. As far as I can tell, you can define your own semantic actions but they simply take in the matched string as an argument----there's no way to get at the values parsed on that level. For example, let's say I have a type struct Person { string name; double height; int age; Person(string n, double h, int a) : name(n), height(h), age(a) }; And I have a document containing 5.5 15 Henry 4.6 32 Jane 6.1 21 Bob I could parse this using rule<> personrule = strict_real_p >> int_p >> ~eol_p; parse(str, *(personrule[WhatGoesHere] >> eol_p) >> end_p); But how can I build a vector<Person> out of this? I can write my own functor to do the push_back, but I need to get at the parsed values of name, age, and height for each personrule somehow. I suppose I *could* just run another parse() on the input the semantic action for that specific person....but that really seems like it shouldn't be necessary.

2 1

Boost and autocompletion
by Hassan Syed 01 Sep '09

01 Sep '09

Hello All, I've been trying to get autocompletion to work under Unix. I do not mind which editor I use really. I was hoping to get the quality of visual studio code completion. I have tried VIM/Ctags/Omnicompletion, Netbeans and Eclipse. The first option does not work well since ctags has a problem indexing boost, the smallest index I could build is 56 mb (after excluding the vector*.hpp files). Netbeans works well however it is dog slow. Eclipse doesn't work at all (perhaps I've done something wrong). Is there a combination for a posix system that will work well with boost (or other template heavy libraries) or do I need to code in visual studio and just recompile under a posix system ?. I guess what I could do is hide the boost code I use behind my own objects in visual studio and then resume development on Posix where it can index my objects nicely. It's depressing how bad the code completion engines are in non-Microsoft editors for C++ :(. Are there any mechanisms in c++0x to make the template code easier to parse for naive code-completion engines ? Regards Hassan

3 4

lock-free shared_ptr on solaris
by Frank Meerkötter 01 Sep '09

01 Sep '09

Hello, from looking at [1] i understood that there is no lock-free implementation of shared_ptrs on solaris. When snooping around in the source i noticed that there is a ./smart_ptr/detail/atomic_count_solaris.hpp It is using the solaris specific <atomic.h> to implement an atomic counter. frank@dexter:~/boost_1_40_0/boost$ find -type f | xargs grep atomic_count_solaris.hpp ./smart_ptr/detail/atomic_count_solaris.hpp:// boost/detail/atomic_count_solaris.hpp But apparently it is not in use. I am currently wondering why? KR, Frank [1] http://www.boost.org/doc/libs/1_40_0/libs/smart_ptr/shared_ptr.htm

2 2