July 2010 - Boost-users - lists.boost.org

Boost::interprocess::string problems
by this_thread 26 Jul '10

26 Jul '10

Good morning all! Many thanks in advance to anyone that can help me out with this matter. I've been doing some work with Boost::Interprocess shared memory lately, and finding it incredibly useful. I've successfully put interprocess mutexes (muti?) and vectors to work, but for whatever reason, I can't seem to get the hang of interprocess strings. I'm sure I'm just missing something stupidly obvious, and to a degree I already see it but don't know what the solution is. For a vector, I know I've got a create a new interprocess allocator which allocates whatever I want to store using the interprocess segment manager: typedef boost::interprocess::allocator<Foo, boost::interprocess::managed_shared_memory::segment_manager> FooAllocator; Then create a special vector which uses that allocator: typedef boost::interprocess::vector<Foo, FooAllocator> FooInterprocessVector; This works great, provided I remember to find<FooInterprocessVector> on the opposite process before accessing the vector during periods when it may have been modified. I can't seem to get the hang of strings though. I've looked around online and can't seem to find an example of how this should be set up, and I can't get it to accept an allocator. The only way I can seem to typedef one is something like: typedef boost::interprocess::string BoostString - or - typedef boost::interprocess::basic_string<char> BoostString; Using this, so long as I only write a short string to the shared memory, things work fine, but anything over a few characters quickly clobbers other data. Has anyone used Interprocess::Strings before and could possibly point me in the direction I'm going wrong? Thank you all! -- View this message in context: http://old.nabble.com/Boost%3A%3Ainterprocess%3A%3Astring-problems-tp292061… Sent from the Boost - Users mailing list archive at Nabble.com.

3 8

base64 encoding ?
by Timothy Madden 26 Jul '10

26 Jul '10

Hello I keep looking in the online docs, and I can find old discussions on the issue, but I can not find the function for base64 encoding in boost. Can anyone please direct me to it ? Is it still there in boost ? Thank you, Timothy Madden

2 3

boost basic_file open
by Afshin Amighi 26 Jul '10

26 Jul '10

Hello , I am very new in boost file system and simply I wrote a sample to use boost file system. I need a file descriptor for both read and write. When I implement below code for "sink" and "source" files, they work. But I don't understand why I couldn't open a basic_file object to read and write. Any comment is appreciated. Thanks in advance. void test_boost2(){ long long size=5*1024*1024; char buffer [L_tmpnam]; const char* prefix = "c1t"; string tempPath(archive::tmpdir()); filesystem::path path(tempPath); filesystem::space_info spi(space(path)); strcpy(buffer,prefix); if( spi.free > size * 3/2 && tmpnam (buffer+sizeof(prefix)-1) != NULL){ buffer[sizeof(prefix)-1]='_'; path/=buffer; boost::iostreams::basic_file<char> *tmpFile= new basic_file<char>(path.string()); tmpFile->open(path.string()); if(tmpFile->is_open()){ long long p=size-DATA_SIZE-2; tmpFile->seek(p,ios_base::beg); tmpFile->close(); } delete tmpFile; } }

1 0

boost matching \b and match_not_null flag
by Gavrilov, Roman 26 Jul '10

26 Jul '10

Hello, I'm trying to find word boundaries with regex_search and "\b" pattern. I have the following code: string buff = "aaaa*******\\b"; boost::match_flag_type flags = boost::match_default | boost::match_not_null; boost::regex re("\\b"); boost::match_results<string::const_iterator> what; string::const_iterator start = buff.begin(); string::const_iterator end = buff.end(); if ( regex_search(start, end, what, re, flags)) { string out(start,what[0].second); cout << "match : " << out << endl; } I would expect it to print "aaaa" in out string and to match the second boundary, but in fact it didn't match anything. If I remove the match_not_null flag then it will match the first boundary at the beginning of the string as expected. I don't understand this behavior, am I missing something? Any help would be highly appreciated. Thanks, Roman

2 2

[move] move emulation with bind?
by Tim Blechmann 26 Jul '10

26 Jul '10

hi all, i'm trying to make one of my classes movable. however i have some issues when using it with boost::bind: class my_movable_class; void my_function(my_movable_class const &); this doesn't compile: boost::bind(my_function, boost::move(my_movable_class)) the reason is, that the members of boost::rv are private ... is this my fault or an issue with the move emulation? thanks, tim -- tim(a)klingt.org http://tim.klingt.org Just what the hell is the experimental tradition? Morton Feldman

3 4

[statechart] template state derived from simple_state compile error
by Nicolas Emiliani 26 Jul '10

26 Jul '10

Hi, I'm experiencing a compiling problem with boost.statechart. I created a template class called request<TCallBackType> derived from state_machine, and two template states called init<TCallBackType> and post<TCallBackType>, the idea is that request's initial states is init, and when recating to an event (initialize) it transits to post. The error i get is : sc_test.cpp: In member function ‘boost::statechart::result init<TCallBackType>::react(const initialize&)’: sc_test.cpp:38: error: expected primary-expression before ‘>’ token sc_test.cpp:38: error: expected primary-expression before ‘)’ token I'm pretty sure this has to be with the fact that the template method transit is here dependant of the template parameter, thus the compiler can't solve it, but I can't seem to figure it out. Tryed using typename but did not work. I'm using gcc (Ubuntu 4.4.3-4ubuntu5) 4.4.3 Any help will be aprecciated. Here's the code, there is a comment where the error occurs. #include <boost/statechart/event.hpp> #include <boost/statechart/state_machine.hpp> #include <boost/statechart/simple_state.hpp> #include <boost/statechart/state.hpp> #include <boost/statechart/transition.hpp> #include <boost/statechart/custom_reaction.hpp> #include <boost/mpl/list.hpp> namespace sc = boost::statechart; //FORWARD DECLARATION OF THE INITIAL STATE template<typename TCallBackType> struct init; template<typename TCallBackType> struct post; //STATE MACHINE template<typename TCallBackType> struct request : sc::state_machine<request<TCallBackType> , init<TCallBackType> > { TCallBackType call_backs; }; //EVENT struct initialize : sc::event<initialize> { }; //INITIAL STATE template<typename TCallBackType> struct init : sc::simple_state< init<TCallBackType>, request<TCallBackType> > { //reactions typedef boost::mpl::list< sc::custom_reaction<initialize> > reactions; //on initialize sc::result react(const initialize& e){ //HERE I GET THE COMPILING ERROR //IF I COMMENT THIS IT ALL GOES SMOOTHLY return this->transit< post<TCallBackType> > (); } typedef post<TCallBackType> p; }; //SECOND STATE template<typename TCallBackType> struct post : sc::simple_state< post<TCallBackType>, request<TCallBackType> > { //reactions typedef boost::mpl::list< sc::custom_reaction<initialize> > reactions; //on initialize sc::result react(const initialize& e){ return this->discard_event(); } }; int main(int argc, char* argv[]){ request<int> r; r.initiate(); return 0; } Nicolas Emiliani.

3 4

Re: [Boost-users] Multi_indexcomposite_key
by isjk 26 Jul '10

26 Jul '10

there are one hundred million defferent manufacturer, all have "A4", "A8", "A12", "A16"... model car. manufacturer and model is many-to-many relationship.</:includetail> Particularly large data.</:includetail> </:includetail></:includetail> struct _Modify_price</:includetail> {</:includetail> _Modify_price(int price_):m_price_(price_){}</:includetail> void operator()(car_model& cm_)</:includetail> {</:includetail> cm_.price = m_price_;</:includetail> }</:includetail> private:</:includetail> int m_price_;</:includetail> }; ------------------ 原始邮件 ------------------ 发件人: "Joaquin M Lopez Munoz"<joaquin(a)tid.es>; 发送时间: 2010年7月26日(星期一) 凌晨2:49 收件人: "boost-users"<boost-users(a)lists.boost.org>; 主题: Re: [Boost-users] Multi_indexcomposite_key isjk <isjk <at> foxmail.com> writes: > Oh, sorry. Nothing to be sorry about. On another topic, please don't top-post and try to follow the discussion policies of the list as summarized at http://www.boost.org/community/policy.html > Assume "cadillac" have "A4" model, "bmw" have "A4" model too. > now, if I will querying "A4" 's price, I must kown the manufacturer. > > ordered_no_unique< > tag<model>,BOOST_MULTI_INDEX_MEMBER(car_model,std::string,model), Correct, you must change from the example's original ordered_unique to ordered_no_unique so as to allow cars with from different manufacturers yet the same model name. > ct.insert(car_model("A4",cadillac,78901)); > ct.insert(car_model("A4",bmw,65412)); > > index_iterator<car_table,model>::type im0,im1; > boost::tuples::tie(im0,im1)=get<model>(ct).equal_range(model); > > car_table_manufacturer_view ctmv; > copy(im0,im1,back_inserter(ctmv)); > car_table_manufacturer_view::iterator ictmv = ctmv.find(cm); This looks right (haven't compiled it), although a little contrived: unless you expect to have *many* cars with the same model name, you can just traverse [im0,im1) and locate the car you're after. > ctmv.modify(cctmv, _Modify_price(10000)); // right or wrong??? I don't know how you've defined _Modify_price. Care to show? Joaquín M López Muoz Telefónica, Investigación y Desarrollo _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users </:includetail>

2 1

[serialization] Reading back old archives
by David Raulo 26 Jul '10

26 Jul '10

Hello, I have a problem reading back serialization archives generated by older version of boost. I used to have a YYYYMMDD class versioning scheme. This broke with boost 1.42, which truncates the class version numbers to 16 bits, causing my code to generated corrupted archives. I fixed my code by implementing my own class versioning when compiling with boost 1.42, so I am again able to create archives that I can read back. Unfortunately this fixes only half of the problem. How can I read back archives that were saved using older releases of the serialization library? I saw that someone else opened ticket 3990 for the same problem. Here you suggested using the function get_library_version() in the following way: template<class Archive> void serialization(Archive & ar, unsigned int version){ library_version_type library_version = get_library_version if(library_version < 6) // the library version is a date my_date = version; else{ // don't use the boost serialization number, use ours instead ar & my_date; } ar && old_data_items if(ar.is_loading) if(my_date > "jan 10, 2010") ar & new_data_item; but I can't get this to work: the version argument is already truncated, and can't be interpreted as a date. As the bug reporter said, this is a case of data loss. Should I reopen ticket 3990, or is this code supposed to work with svn (I admit I did not test the trunk)? Thanks for your help, David.

2 11

[Serialization][BGL] problem serializing an edge
by Ireneusz Szcześniak 25 Jul '10

25 Jul '10

Hi, I have a problem serializing an edge in a graph. This is my test case: #include <iostream> #include <boost/archive/text_oarchive.hpp> #include <boost/graph/adj_list_serialize.hpp> using namespace std; using namespace boost; int main() { typedef adjacency_list<listS, vecS, undirectedS> Graph; typedef graph_traits<Graph>::edge_descriptor Edge; Graph g; Edge e; boost::archive::text_oarchive oa(cout); oa << g; oa << e; return 0; } Compilation with gcc version 4.4.1 returns this error: /usr/include/boost/serialization/access.hpp:109: error: ‘class boost::detail::edge_desc_impl<boost::undirected_tag, unsigned int>’ has no member named ‘serialize’ I would appreciate it if someone could point out what the problem is and how to fix it. Thanks, Irek -- Ireneusz (Irek) Szczesniak http://www.irkos.org

3 6

Re: [Boost-users] ¤Multi_indexcomposite_key~~
by isjk 25 Jul '10

25 Jul '10

Oh, sorry. Assume "cadillac" have "A4" model, "bmw" have "A4" model too. now, if I will querying "A4" 's price, I must kown the manufacturer. ordered_no_unique< tag<model>,BOOST_MULTI_INDEX_MEMBER(car_model,std::string,model) >,ct.insert(car_model("A4",cadillac,78901));ct.insert(car_model("A4",bmw,65412)); index_iterator<car_table,model>::type im0,im1; boost::tuples::tie(im0,im1)=get<model>(ct).equal_range(model);car_table_manufacturer_view ctmv;copy(im0,im1,back_inserter(ctmv));car_table_manufacturer_view::iterator ictmv = ctmv.find(cm); ctmv.modify(cctmv, _Modify_price(10000)); // right or wrong??? ------------------ 原始邮件 ------------------ 发件人: "Joaquin M Lopez Munoz"<joaquin(a)tid.es>; 发送时间: 2010年7月25日(星期天) 晚上7:17 收件人: "boost-users"<boost-users(a)lists.boost.org>; 主题: Re: [Boost-users]¤Multi_indexcomposite_key~~ isjk <isjk <at> foxmail.com> writes: > > > thanks a lot! > > now, I have a new problem: > > Multi_index example 6: complex searches and forergn keys, > struct car_manufacturer > { > std::string name; > }; > > struct car_model > { > std::string model; > car_manufacturer* manufacturer; > int price; > }; > > I want use "model, name" to query "price", and I maybe to modify "price". > > How to do it? thanks. Sorry but your question is not clear to me: what do you mean by querying "price" using "model, name"? If you know the model of a car then you just can use find() in the car_table to locate the element and access the price. As for modifying, there's a tutorial section on updating elements of a multi_index_container at http://www.boost.org/libs/multi_index/doc/tutorial/basics.html#ord_updating Joaquín M López Muoz Telefónica, Investigación y Desarrollo _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1