June 2007 - Boost-users - lists.boost.org

derivative work
by Ruediger Berlich 14 Jun '07

14 Jun '07

Hi there, just a short question about the meaning of "derivative work" in the Boost software license (and the associated requirement to list the Copyright and license terms in such work). My understanding of the term "derivative work" is that it applies to modifications of the original code only, not to software _using_ the code. I.e., if I were to modify the code in shared_ptr.hpp, then of course I would have to include the original Copyright notices and license terms. Whereas, if I use shared_ptr in my software, I do not need to list the original Copyright notices and license terms. Plus, the Copyright of my own cose is solely with me. Is this interpretation correct ? Best Regards and thanks for your help, Ruediger Berlich

10 19

Getting the argument types from a function type
by Kirit Sælensminde 14 Jun '07

14 Jun '07

I've been playing around with a few functional programming idioms in C++ using Boost.function and Boost.lambda and was wondering if it was possible to fetch out the argument type (i.e. int) from a type like this: void (*somefunc)( int ) It looks to me like boost::function<> might be doing something along these lines. I've been working on a version of 'with' for use in initialising lists etc. At the moment you need to do this: std::list< int > list1; list1.push_back( 3 ); list1.push_back( 1 ); list1.push_back( 4 ); But using 'with' you can do this: with( list2, &std::list< int >::push_back )( 3 )( 1 )( 4 )( 1 )( 5 ); The implementation I have right now is fairly simple: template< typename F > struct with_binder { with_binder( F f ) : m_f( f ) { } const with_binder &operator()( int i ) const { m_f( i ); return *this; } private: F m_f; }; template< typename O, typename F > inline with_binder< boost::function< void ( int ) > > with( O &o, F f ) { return with_binder< boost::function< void ( int ) > >( boost::lambda::bind( f, &o, boost::lambda::_1 ) ); } But the 'int' argument type is hard coded. I could have it as the first template parameter to 'with' and simply require that it is given when using 'with', but it would neater if it could be derived from the type 'F' in 'with'. Clearly to be more generally useful it also needs to be extended for multiple parameters. K

2 2

chdir
by ivan＠0x4849.net 14 Jun '07

14 Jun '07

Hi, I am trying to build boost on a debian system using bash shell. I am getting erros: /bin/sh: line 1: chdir: command not found Is this a common problem? Should i just alias chdir to cd? Thanks, Ivan --------------------- Ivan Novick http://www.0x4849.net

2 2

graph library : isomorphism.hpp vertex invariant parameter
by Patrick A. La Fratta 14 Jun '07

14 Jun '07

Setting the "vertex_invariant" parameter to the isomorphism function does not seem to have any affect on the behavior of the function. In isomorphism.hpp, it appears that the function "isomorphism_impl" is supposed to read the parameter at the return statement: return isomorphism(G1, G2, f, choose_param(get_param(params, vertex_invariant1_t()), invariant1), choose_param(get_param(params, vertex_invariant2_t()), invariant2), choose_param(get_param(params, vertex_max_invariant_t()), ( invariant2.max )()), index_map1, index_map2 ); If "vertex_invariant1_t()" and "vertex_invariant2_t()" are both changed to "vertex_invariant_t()", the setting of the parameter then seems to have an affect. Has anyone been able to successfully set the vertex_invariant parameter using the named parameter version of the isomorphism function? I have had to resort to calling one of the internal isomorphism functions to set this parameter. any help is appreciated, Patrick

3 5

some libraries missing
by Fulvio Esposito 13 Jun '07

13 Jun '07

Hi all boost users, recently I downloaded source package for version 1.34 from sf.net, but i discovered some library that was accepted were missing (like asio, gil and some others). I think at least they should be in CVS but i've never worked with a source control system. How can I obtain a full distribution from CVS? I'm working on Microsoft Windows XP/Visual C++ Express 2005! Thank you all guys, Fulvio Esposito

2 1

[boost::program_options] optional arguments
by Bryan Green 13 Jun '07

13 Jun '07

Hello, Is there a mechanism in boost::program_options for supporting optional arguments? If not, is there a suggested program_options-based alternative? This question was posed by someone else about a year ago, but did not get a response. Here is the quote: BEGIN QUOTE From: Dale McCoy (dalestan_at_[hidden]) Date: 2006-04-22 21:15:51 I'm trying to convert from getopt (GNU's getopt_long, to be more specific) to boost::program_options, but have run into trouble converting "D::" and {"data",optional_argument,NULL,'D'} (If you're not familiar with getopt, that means that --data is equivalent to -D, and both take an optional argument, so "-D", "--data", "-Darg", and "--data=arg" are all valid.) I see that the first argument to add_option should be "data,D", but then things get fuzzy. I saw nothing in the documentation concerning optional arguments, either directly, or using additional parsers. What is the appropriate method for implementing such options? I'm using currently using Boost 1.33.1. Dale END QUOTE Bryan

2 4

limits.hpp defective
by Daniel Krügler 13 Jun '07

13 Jun '07

Hello, ~3 weeks ago I reported a bug in limits.hpp: May, 21th: "Wrong member definitions in limits.hpp" and also put it into the bug tracking system: http://svn.boost.org/trac/boost/ticket/972 Since no feedback occurred, I repost this now. Another, related question: I found that both SVN and CVS repositories have rather different contents. Could someone explain to me the reasons for these deviations? Thanks in advance, Daniel Krügler

3 2

[intrusive] lock-free algorithms
by Zeljko Vrba 13 Jun '07

13 Jun '07

Are there any plans to extend the intrusive library with lock-free data structures (stacks, FIFO queues)? I'm about to start writing my own lock-free classes and have thought that it'd be nice to include them in the intrusive library, if possible [I've gotten used to the interface, and I don't want to have different interfaces in my code]. My goals: - lock-free stack and queue - SMR (both described in "Safe Memory Reclamation for Dynamic Lock-Free Objects Using Atomic Reads and Writes" by Maged M. Michael) - target platform: Solaris using atomic_ops (3C); possibly also inline ASM for SunCC I don't have neither time nor resources (nor will, for that matter) to make a full implementation that works on all platforms supported by Boost. Also, I do *not* want to use Boost.Thread. The code is parametrized by the following (all related to SMR): - number of participating threads (every "thread" must be associated with a unique integer starting from 0) - number of hazard pointers - batch size SMR also needs static storage accessed by all threads. Any suggestions are welcome.

2 6

astar help
by bringiton bringiton 13 Jun '07

13 Jun '07

I have a number of questions RE the astar-cities example. Why is the distance_heuristic class needed. Is it used to find the weight for a test vertex to the goal vertex? I put a breakpoint in distance_heuristic::operator() and it never gets called. Is it ever used? astar_search (... distance_heuristic<mygraph_t, cost, location*> ...) Do I have to define my own or supply an astar_heuristic? Or can I pass a default heuristic? astar_goal_visitor. Do I need to use this or can I pass a default default_astar_visitor? Are there other ways of reaching a goal without throwing an exception? Thanks. Any advice appreciated

1 1

multi_index_container and templates
by Andrej van der Zee 13 Jun '07

13 Jun '07

Hi, I am trying to set up a multi_index_container-type with a template argument like this: template <typename event_type> struct mi_container { typedef typename mi::composite_key< event_type, mi::const_mem_fun<fields::field_ts, ts_type, &fields::field_ts::ts>, mi::const_mem_fun<fields::field_addr, addr_type, &fields::field_addr::addr>, mi::const_mem_fun<fields::field_port, port_type, &fields::field_port::port> > self_key_type; typedef typename mi::composite_key< event_type, mi::const_mem_fun<fields::field_ts, ts_type, &fields::field_ts::ts>, mi::const_mem_fun<fields::field_peer_addr, addr_type, &fields::field_peer_addr::peer_addr>, mi::const_mem_fun<fields::field_peer_port, port_type, &fields::field_peer_port::peer_port> > peer_key_type; typedef typename mi::multi_index_container< event_type, mi::indexed_by< mi::ordered_non_unique< mi::tag<self_key_tag>, self_key_type>, mi::ordered_non_unique< mi::tag<peer_key_tag>, peer_key_type> > > type; }; mi_container<ev_connect>::type connected; This gives me the following compile-error: main.cpp: In instantiation of mi_container<event::ev_connect>: main.cpp:78: instantiated from here main.cpp:57: error: invalid use of event::fields::field_ts::ts to form a pointer-to-member-function main.cpp:57: note: a qualified-id is required main.cpp:64: error: invalid use of event::fields::field_ts::ts to form a pointer-to-member-function main.cpp:64: note: a qualified-id is required main.cpp:75: error: invalid use of event::fields::field_ts::ts to form a pointer-to-member-function main.cpp:75: note: a qualified-id is required Probably I doing something silly. Note that if I remove the class template and directly specialize it, there is no problem at all. Can anybody help me out? Thanks, Andrej ___________________________________________________________ Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today http://uk.rd.yahoo.com/evt=44106/*http://uk.docs.yahoo.com/mail/winter07.ht…

2 5