October 2007 - Boost-users - lists.preview.boost.org

Re: [Boost-users] regex_search in std::string
by Troy Alexander 10 Oct '07

10 Oct '07

Thanks. That did it. I was focusing on the string to match not the return type. Makes since that they need to be the same type. Troy -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of John Maddock Sent: Wednesday, October 10, 2007 2:44 AM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] regex_search in std::string Troy Alexander wrote: > The 1st regex_search looking at input_line char[] works Ok it's the > 2nd one looking at date_and_time_s std::string that gets the error. > If I remark out the 2nd one the program builds and I watch the > input_line doing what it should in the debugger. > > Also this code used to work before I upgraded from Visual C++6 to > Visual C++8 and from bost_1_33_1 to boost_1_34_0. I don't know if it > was the change in compiler or the boost lib that broke this. In VC6 this worked "by accident" because const char* and std::string::const_iterator were the same type. That's not true in general, and you should use boost::smatch as the match_results type for use with std::string's. HTH, John. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

[BGL] maximum flow through push_relabel_max_flow
by "Alejandro M. Aragón" 10 Oct '07

10 Oct '07

Hello all, I've been working with the Edmunds & Karp maximum flow algorithm in the BGL for some time without any problems. However, I realized that I could do better in time complexity by moving to the push_relabel_max_flow algorithm so I gave it a try. I've been trying to figure out what the problem with my code is for the entire day without success. I was wondering if someone could tell me what is my mistake. It is a very simple problem, four vertices, four edges, the result from the algorithm should be two: 2--------1 | | | | | | 0--------3 Source is 0, target is 1, all edges have unit capacity. The code is: void compute_max_flow(Graph* g_) { // obtain number of vertices in undirected graph size_t n = num_vertices(*g_); typedef adjacency_list_traits<vecS, vecS, directedS> Traits; typedef adjacency_list<listS, vecS, directedS, property<vertex_index_t, int>, property<edge_capacity_t, int, property<edge_residual_capacity_t, int, property<edge_reverse_t, Traits::edge_descriptor> > > > FlowGraph; typedef graph_traits<FlowGraph>::edge_descriptor EdgeFlow; FlowGraph g(num_vertices(*this->g_)); property_map<FlowGraph, vertex_index_t>::type indexmap = get(vertex_index, g); property_map<FlowGraph, edge_capacity_t>::type capacity = get(edge_capacity, g); property_map<FlowGraph, edge_reverse_t>::type rev = get(edge_reverse, g); property_map<FlowGraph, edge_residual_capacity_t>::type residual_capacity = get(edge_residual_capacity, g); Traits::vertex_descriptor s, t; EdgeFlow e1,e2; bool in1,in2; // add edges, reverse edges and capacities to new graph for flow computation edge_iter ei,ei_end; for (tie(ei,ei_end) = edges(*g_); ei != ei_end; ++ei) { // obtain vertices in the original graph vertex u = source(*ei, *g_); vertex v = target(*ei, *g_); // insert edges in new graph and add unit capacities tie(e1, in1) = add_edge(u,v,g); tie(e2, in2) = add_edge(v,u,g); if(in1 && in2) { // assign reverse edges to property map rev[e1] = e2; rev[e2] = e1; capacity[e1] = 1; capacity[e2] = 0; } } graph_traits<FlowGraph>::vertex_iterator vvi,vvi_end; for(tie(vvi,vvi_end) = vertices(g); vvi!=vvi_end; ++vvi) { cout<<"index -> "<<indexmap[*vvi]<<endl; } graph_traits<FlowGraph>::edge_iterator eei,eei_end; for(tie(eei,eei_end) = edges(g); eei!=eei_end; ++eei) { cout<<"edge -> "<<*eei<<endl; cout<<"capacity -> "<<capacity[*eei]<<endl; cout<<"rev -> "<<rev[*eei]<<endl; } int flow = push_relabel_max_flow(g, 0, 1, capacity, residual_capacity, rev, indexmap); cout<<"flow -> "<<flow<<endl; } Results in: index -> 0 index -> 1 index -> 2 index -> 3 edge -> (0,2) capacity -> 1 rev -> (2,0) edge -> (0,3) capacity -> 1 rev -> (3,0) edge -> (1,2) capacity -> 1 rev -> (2,1) edge -> (1,3) capacity -> 1 rev -> (3,1) edge -> (2,0) capacity -> 0 rev -> (0,2) edge -> (2,1) capacity -> 0 rev -> (1,2) edge -> (3,0) capacity -> 0 rev -> (0,3) edge -> (3,1) capacity -> 0 rev -> (1,3) flow -> 0 I would appreciate any inputs on this. Thank you, a²

2 5

Usage of Transitive_closure().
by Ning Ruan 10 Oct '07

10 Oct '07

my code to use transitive_closure(): Graph tc; VertexIndexMap index_map = get(&VertexProp::id, g); vector<Vertex> to_tc_vec(num_vertices(g)); iterator_property_map<Vertex*, VertexIndexMap, Vertex, Vertex&> g_to_tc_map(&to_tc_vec[0], index_map); transitive_closure(graph, tc, g_to_tc_map, (index_map)); I use print_graph(tc, get(&VertexProp::id, tc) to print the transitive graph, then it shows: 0 --> 0 0 0 0 0 0 0 0 0 --> 0 0 0 --> 0 --> 0 0 0 0 0 0 0 0 0 --> 0 --> 0 --> 0 0 0 0 --> 0 0 0 0 0 0 0 0 the result miss the id of vertex, what 's the problem. my original graph g is correcet: 0 --> 1 2 3 1 --> 4 5 2 --> 3 --> 6 7 4 --> 5 --> 6 --> 1 7 --> 0

3 4

regex_search in std::string
by Troy Alexander 10 Oct '07

10 Oct '07

I am struggling with a regex in some code of mine. I want to use regex_search to look for a match in a std::string, but my code will not build. This is a snippet of what I am trying to do. using namespace std; string in_file_name_s; string date_and_time_s; string day_s; char input_line[5000]; boost::regex re_date_time("(.*-.*-.*:.*:.*),.*,.*,.*,.*,.*,.*,.*,.*,"); boost::regex re_pices_date_time("(.*)\\-(.*)\\-(.*) (.*)"); boost::cmatch match; while (inFile.getline(input_line , 5000)) { if (boost::regex_search(input_line, match , re_date_time )) { date_and_time_s = match[1]; if (boost::regex_search(date_and_time_s , match , re_pices_date_time )) { day_s = match[1]; } } } When I compile I get the following error error C2784: 'bool boost::regex_search(const std::basic_string<charT,ST,SA> &,const boost::basic_regex<charT,traits> &,boost::regex_constants::match_flag_type)' : could not deduce template argument for 'const boost::basic_regex<charT,traits> &' from 'boost::cmatch' The 1st regex_search looking at input_line char[] works Ok it's the 2nd one looking at date_and_time_s std::string that gets the error. If I remark out the 2nd one the program builds and I watch the input_line doing what it should in the debugger. Also this code used to work before I upgraded from Visual C++6 to Visual C++8 and from bost_1_33_1 to boost_1_34_0. I don't know if it was the change in compiler or the boost lib that broke this. Thanks in advance for your help. Troy

2 1

Re: [Boost-users] Boost::regex causes runtime errors with CLR
by Hansi Hansi 10 Oct '07

10 Oct '07

Hello, In VS2005 is a problem with the Initializing of the CRT when you have a .net application. Normally it helps when you use as entry point '?mainCRTStartupStrArray@@$$FYMHP$01AP$AAVString@System@@@Z'. I hope this helps. Best regards Hansjörg _____________________________________________________________________ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=000000000066

2 1

Boost::regex causes runtime errors with CLR
by Rajib 09 Oct '07

09 Oct '07

Hi, I'm trying to use boost::regex with a "Windows Forms Application" in Visual Studio 2005. Everything works fine with normal C++, and even using a "CLR Console Application." But when I use the "Windows Form Application" template the program crashes while initializing variables. Basically I take code like: #include <boost/regex.hpp> using boost::regex; using boost::regex_search; bool h() { return regex_search("asdfasdf",regex("sd")); } put it in a separate file, turn off precompiled headers for that file, and never actually call h(). The program crashes instantly with a debug assertion failed at dbgheap.c/Line 1414/_CrtIsValidHeapPointer. The call stack looks like: ... msvcr80d.dll!__dllonexit(int (void)* func=0x00446110, void (void)* * * pbegin=0x0012eab8, void (void)* * * pend=0x0012eab0) Line 273 + 0x11 bytes C temp.exe!_onexit(int (void)* func=0x00446110) Line 110 + 0x11 bytes C temp.exe!atexit(void (void)* func=0x00446110) Line 127 + 0x9 bytes C temp.exe!boost::re_detail::`dynamic initializer for 'block_cache''() Line 189 + 0x2b bytes C++ ... It does this whether I compile the file with regex with the /clr flag or not (although the /clr flag gives a lot of errors, one of which isn't about converting fastcall to stdcall: include\boost/shared_ptr.hpp(92) : warning C4793: 'vararg' : causes native code generation for function 'void boost::detail::sp_enable_shared_from_this(const boost::detail::shared_count &,...)' Thanks, Rajib

1 0

Boost Spirit and C++/CLI
by Jason Stewart 09 Oct '07

09 Oct '07

Has anyone tried using boost spirit with C++/CLI? I have a library that uses Spirit internally to parse a file. The library is built as an unmanaged C++ static library. I then link that to a managed C++/CLI application but when I try to parse the file I get heap corruption. The corruption seems to happen near an #ifdef for BOOST_SPIRIT_THREADSAFE so I tried to define that for the build of the library and the application. Unfortunately it now fails to initialize correctly and dies on startup. Has anyone done this before? Thanks, Jason Stewart

1 0

boost::serialization XSD support
by Juhasz, Zoltan (IT) 09 Oct '07

09 Oct '07

Hi, I've got a pretty strange problem. We are using boost::serialization library to serialize arbitrary structs into XML. That is working pretty good. Our problems are: - we also need to generate schema files (XSD) for the archive file, since this XML archive file is going to be used by other programs (e.g. written in Java etc.), and - the unserialization part should be validated through the schema file. Currently, the boost::serialization library can serialize / deserialize our structs to XML just fine, but: How can we add XSD generation support? How can we 'force' boost::serialization to validate the XML file through an schema file (XSD)? PS: Unfortunately, the automatic generation of the XSD file is a strict requirement, that is why we cannot use regular XML <> C++ object mapping tools, where the C++ codes are generated by an external tool from a schema (XSD) file. PS2: If you know other third party tools that can solve the above mentioned problem, please feel free to write. Thank you in advance! Best Regards, Zoltan Juhasz Morgan Stanley | Technology | Budapest Zoltan dot Juhasz at MorganStanley dot com -------------------------------------------------------- NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

2 1

how to schedule nodes in graph
by Suresh Gupta 09 Oct '07

09 Oct '07

Hi, I want to do the following: I have a graph in boost. Now I want to schedule the graph nodes in topological order as follows: 1. The nodes which do not have any parents are to be scheduled in first cycle. Record the list of these nodes. 2. Remove already recorded nodes (in step 1) from the graph. 3. Repeat 1 and 2 on the remaining graph and so on. How to do the above? I need functions: 1. Get all nodes which do not have parents. 2. To remove the nodes from graph. Please suggest me about the functions in boost so do the same. Thanks, Suresh

2 3

Boost::Regex question
by Dima Kagan 09 Oct '07

09 Oct '07

Hi, I'm currently evaluating the Regex package and encountered a minor issue. I need to match a string that can contain some special characters (.[{()\*+?|^$) as well. My question is if there is a way to tell Regex to treat this string "as is", including the special characters? I need to do this without escaping the special characters in the string or the match expression. For those of you familiar with Tcl Expect, I'm looking for an option similar to "expect -ex". Thank you! Regards, Dima Kagan

2 2

defining a function according to a signature given as template parameter
by Stefan Funck 09 Oct '07

09 Oct '07

Hi everybody, I'd like to have a generic wrapper that can be put "in between" a signal and a slot to provide some additional functionality to the slot. So where I normally have something like void myslot(int i) { std::cout << "slot(" << i << ")\n"; } int main(int argc, char *argv[]) { boost::signal<void (int)> sig; sig.connect(boost::bind(myslot, _1)); sig(123); return 0; } I want something like int main(int argc, char *argv[]) { boost::signal<void (int)> sig; SlotWrapper<void (int)> slotWrap(boost::bind(myslot, _1)); sig.connect(boost::bind(SlotWrapper<void (int)>::slot, &slotWrap, _1)); sig(123); return 0; } The signature of SlotWrapper<Signature>'s member function slot() is defined by Signature. slot()'s implementation should have some common functionality (that's what I want the wrapper for) and then call the actual slot that was given at construction (myslot() in the example above). Is that doable? So far I have two questions: 1) How can I define a member function according to the signature that is given as a template parameter to the class? 2) If I can do 1), how can I then bind the arguments given and call another function, i.e. the original slot? A skeleton of the wrapper might look like: template <typename Signature> class SlotWrapper { public: SlotWrapper(const typename boost::signal<Signature>::slot_function_type& slot) : m_slot(slot) {} // define a function according to Signature typename boost::signal<Signature>::result_type slot(...arguments...) { // do something common // call somehow the actual slot (m_slot) with the given ...arguments... } private: const typename boost::function<Signature> m_slot; }; How can I write the slot() function? Thanks for any help, Stefan

2 2

boost::basic_match::size_type
by Bill Webster 09 Oct '07

09 Oct '07

The basic_match access methods have int index parameters, whereas basic_match::size() returns basic_match::size_type. This means that the following results in compiler warnings (under MSVC 8.0) - boost::smatch match; // ... for ( boost::smatch::size_type i = 0; i < match.size(); i++ ) { cout << match.str( i ).c_str() << endl; }

1 0

Dimensional analysis
by Joel C. Salomon 09 Oct '07

09 Oct '07

Have the dimensional analysis templates suggested in <http://www.boost.org/libs/mpl/doc/tutorial/dimensional-analysis.html> ever been developed into readily usable form? --Joel

3 3

serialization and object tracking question...
by Graham Reitz 09 Oct '07

09 Oct '07

I am having trouble wrapping my head around object tracking for serialization. I see how to change the default behavior: BOOST_CLASS_TRACKING(my_class, boost::serialization::track_never) ...but based on the error below how do I determine what kind of object tracking is needed? oserializer.hpp:566: error: invalid application of 'sizeof' to incomplete type 'boost::STATIC_ASSERTION_FAILURE<false>' inline void save(Archive & ar, T &t){ // if your program traps here, it indicates taht your doing one of the following: // a) serializing an object of a type marked "track_never" through a pointer. // b) saving an non-const object of a type not markd "track_never) // Either of these conditions may be an indicator of an error usage of the // serialization library and should be double checked. See documentation on // object tracking. BOOST_STATIC_ASSERT(check_tracking<T>::value); The following is the class to be serialized: class client_i { private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & m_client; } clients m_client; public: client_i() {}; client_i(clients c) : m_client(c) {}; ~client_i() {}; }; clients is a mysql++ specialized sql structure that is created via one of the mysql++ macros, i.e. sql_create_N(). Any help understanding this would be greatly appreciated. Thanks, Graham

2 4

thread creating just hangs
by Eyon Land 08 Oct '07

08 Oct '07

Hey folks, I'm new to boost. I've used pthread and wanted to try out boost::thread. After installing the latest release I found that creation of the thread instance simply hangs. I use CentOS 4.3 on a single CPU machine. Has this happened to anyone else? Thanks in advance! Eyon ____________________________________________________________________________________ Moody friends. Drama queens. Your life? Nope! - their life, your story. Play Sims Stories at Yahoo! Games. http://sims.yahoo.com/

1 0

Threads, condition variables, and signals
by Scott Gifford 08 Oct '07

08 Oct '07

Hello, I'm writing a multithreaded server on Unix using boost::thread. The basic operation is fairly straightforward; a master thread starts up worker threads, and coordinates various management tasks using condition variables. As with most Unix servers, the server is shut down with an OS signal telling the server to shut down. The server catches this signal, then simply calls exit(1). Any important cleanup happens in the destructors of global variables, so it will happen properly no matter how the server exits. One object which is always cleaned up is the object for the master thread. The master thread sets a flag, then does a notify_all() on some condition variables to tell the other threads to shut down, The problem I'm having is that if the master thread is waiting on a condition variable when the signal arrives, and the destructor does a notify_all() on that condition variable, then the destructor for the boost::condition variable hangs. For example, if I run the included code and hit CTRL-C after a short bit, I get this output: Waiting... ^C Caught signal 2 Destroyed ThreadTest Running destructor 1 [ ... hangs ... ] Here is the program: #include <iostream> #include <signal.h> #include <boost/thread.hpp> #include <boost/thread/mutex.hpp> #include <boost/thread/condition.hpp> template<int n> struct PrintOnDestruct { ~PrintOnDestruct() { std::cout << "Running destructor " << n<< std::endl; } }; class ThreadTest { public: void doSomething() { boost::mutex::scoped_lock lock(myMutex); while(true) { std::cout << "Waiting..." << std::endl; myCond.wait(lock); } } ~ThreadTest() { boost::mutex::scoped_lock lock(myMutex); myCond.notify_all(); std::cout << "Destroyed ThreadTest" << std::endl; } private: PrintOnDestruct<2> d2; boost::mutex myMutex; boost::condition myCond; PrintOnDestruct<1> d1; }; // Exit properly on signal void sighandle(int sig) { std::cerr << "Caught signal " << sig << std::endl; exit(1); } ThreadTest tt; int main() { // Get ready to handle signals signal(SIGINT,sighandle); tt.doSomething(); } Does anybody have any suggestions for a straightforward way to handle this properly? Thanks! ----ScottG.

2 3

[preprocessor] using __FILE__ with BOOST_PP_FILENAME_x
by Server Levent Yilmaz 08 Oct '07

08 Oct '07

Hi all, Please take a look as this trivial example of file iteration: // trivial.h #if !BOOST_PP_IS_ITERATING #ifndef TRIVIAL_H #define TRIVIAL_H #include <boost/preprocessor/iteration/iterate.hpp> #define BOOST_PP_FILENAME_1 "trivial.h" //!!! #define BOOST_PP_ITERATION_LIMITS (1, 3) ??=include BOOST_PP_ITERATE() #endif // TRIVIAL_H #else BOOST_PP_ITERATION() #endif Perhaps not so obvious many (wasn't to me), that #define BOOST_PP_FILENAME_1 __FILE__ doesn't work. One has to specify (kind of like hardcode) the file's name in there.. Is there some, possibly contrived, way to achieve automation of this? regards, - Levent PS: What an AMAZING library this Boost.Preprocessor is!? -- Server Levent Yilmaz Mechanical Engineering University of Pittsburgh

2 2

exp functions with interval
by Mark Beinker 08 Oct '07

08 Oct '07

Hi, I was trying to use the boost interval class using g++ under cygwin and ubuntu, but unfortunately, I have trouble to use the exp function. The following code: #include <iostream> #include <cmath> #include <boost/numeric/interval.hpp> #define DOUBLE boost::numeric::interval<double> int main(int argc, char** argv) { DOUBLE x(4,5); std::cout.precision(19); DOUBLE res = exp(x); std::cout<<'['<<res.lower()<<", "<<res.upper()<<']'<<std::endl; return 0; } generates the following compile error: /usr/include/boost/numeric/interval/transc.hpp: In function `boost::numeric::interval<T, Policies> boost::numeric::exp(const boost::numeric::interval<T, Policies>&) [with T = double, Policies = boost::numeric::interval_lib::policies<boost::numeric::interval_lib::rounded_math<double>, boost::numeric::interval_lib::checking_strict<double> >]': itest.cpp:14: instantiated from here /usr/include/boost/numeric/interval/transc.hpp:34: error: 'struct boost::numeric::interval_lib::rounded_math<double>' has no member named 'exp_down' itest.cpp:14: instantiated from here /usr/include/boost/numeric/interval/transc.hpp:34: error: 'struct boost::numeric::interval_lib::rounded_math<double>' has no member named 'exp_up' I use boost version 1.34.1. Any idea why this doesn't work? Regards, Mark

2 1

[string_algo] Case-insensitive keys in map with is_iless?
by Boris 07 Oct '07

07 Oct '07

Am I supposed to be able to use boost::algorithm::is_iless (see http://www.boost.org/doc/html/boost/algorithm/is_iless.html) to create a map with case-insensitive keys? I wonder as either it is broken in VS2005 or I do something wrong: #include <boost/algorithm/string.hpp> #include <string> #include <map> int main() { std::map<std::string, int, boost::algorithm::is_iless> m; m["test"] = 1; } C:\Program Files\Microsoft Visual Studio 8\VC\include\xlocale(583) : error C2440: 'type cast' : cannot convert from 'unsigned char' to 'std::basic_string<_Elem,_Traits,_Ax>' ... C:\Program Files\Microsoft Visual Studio 8\VC\include\xlocale(561) : error C2440: 'type cast' : cannot convert from 'std::basic_string<_Elem,_Traits,_Ax>' to 'unsigned char' The assignment causes the problem. If you use insert() you get the same compiler error. Without the assignment the code compiles fine. Boris

2 3

"Missing" lib for 1_34_1 using MSVC-71
by Larry 07 Oct '07

07 Oct '07

I had a program that I had built with Boost 1_34 and it worked well (built with MSVC 7.1). I downloaded 1_34_1 and built 1_34_1 for VC8 and VC7.1 using bjam 3.1.14. bjam --toolset=msvc-7.1 install I switched the program build to use Boost 1_34_1 and the build fails looking for libboost_date_time-vc71-mt-sgd-1_34_1.lib as part of the link. Going back to the Boost build I find msvc.archive bin.v2\libs\date_time\build\msvc-71\debug\link-static\runtime-link-static\libboost_date_time-vc-sgd-1_34_1.lib common.copy C:\Boost\lib\libboost_date_time-vc-sgd-1_34_1.lib 1 file(s) copied. The only thing I changed was the include list entry from c:\boost\include\boost_1_34 to c:\boost\include\boost_1_34_1. The lib directory is the same in both cases. I didn't find anything looking through the boost.user group or in the doc but perhaps I looked at it and just didn't realize the answer was there. Perhaps I didn't do something I should have or did something I shouldn't have. Any ideas? I got surprised (probably again) when building Boost because I had run vcvars32.bat for MSVC 7.1 and then did bjam --toolset=msvc install and the surprise was it built for MSVC 8.0. The doc does say the build will find the VC compiler but I had sort of assumed it would use what was in the environment before searching elsewhere for a compiler version. But after that I ran bjam --toolset=msvc-71 install and it ran without complaint. I did several iterations of the 7.1 build including a build with -a and also did a --clean followed by a Boost build and neither altered the outcome of the Boost build (that I could see) and did not alter the outcome of the program build with VC 7.1. The program build was done via VS. Moving the include back to 1_34 works. Larry

2 2

[thread] example to delegtate jobs to threads
by Torsten Mohr 07 Oct '07

07 Oct '07

Hi, i wrote a small test project that makes use of boost::thread to split jobs into parts that are delegated to threads. A class "Job" contains a state machine (IDLE, EXECUTING, ABORTED, ...). That class can be derived from and methods "execute" and "report" can be overwritten. A class JobProvider can be derived from to feed new data to a Job. A Template "Foreman" can then start several threads and delegate Jobs that get their data from "JobProvider" to the Jobs and they report when they are done. In the example, MyJobProvider provides numbers from 3 to 100. MyJob::execute calculates if these numbers are prime. MyJob::report prints out the numbers that are prime. All the methods "report" are protected, only one of them at a time is ever called, methods "execute" are run in parallel. The performance is poor, can anybody give me a hint why? Best regards, Torsten. ######## job.h: #ifndef JOB_H #define JOB_H 1 #include <boost/thread.hpp> extern boost::mutex io_mutex; #define QWE { boost::mutex::scoped_lock lck(io_mutex); printf("file %s line %i\n", __FILE__, __LINE__); } class Job { boost::mutex job_mutex; boost::mutex cond_mutex; boost::mutex* m_report_mutex; boost::condition* m_condition; boost::thread* thr; bool m_stopped; int m_state; void setState(int state); protected: bool is_stopped(void); public: enum { CMD_IDLE = 212, CMD_EXEC, CMD_KILL }; enum { STATE_IDLE = 37, STATE_EXEC_PREPARE, STATE_EXEC, STATE_EXEC_ABORTED, STATE_REPORT_PREPARE, STATE_REPORT, STATE_DEAD }; void set_foreman_data(boost::mutex* report_mutex, boost::condition* condition); void init(void); virtual void execute_handler(void); void report_handler(void); void fork(void); int getState(void); void doCommand(int command); virtual void execute(void); virtual void report(void); }; class JobProvider { public: bool provide(Job& j); }; #endif ######## job.cc: #include <boost/thread.hpp> #include <boost/bind.hpp> #include "job.h" int Job::getState(void) { boost::mutex::scoped_lock lock(job_mutex); return m_state; } void Job::setState(int state) { boost::mutex::scoped_lock lock(job_mutex); m_state = state; } void Job::doCommand(int cmd) { boost::mutex::scoped_lock lock(job_mutex); // sparse transitions if((m_state == STATE_EXEC || m_state == STATE_EXEC_PREPARE) && cmd == CMD_KILL) { m_state = STATE_EXEC_ABORTED; m_stopped = true; } else if(m_state == STATE_IDLE && cmd == CMD_EXEC) { m_state = STATE_EXEC_PREPARE; } else if(m_state == STATE_IDLE && cmd == CMD_KILL) { m_state = STATE_DEAD; m_stopped = true; } } bool Job::is_stopped(void) { boost::mutex::scoped_lock lock(job_mutex); return m_stopped; } void Job::execute_handler(void) { int state; do { state = getState(); switch(state) { case STATE_EXEC_PREPARE: execute(); state = getState(); if(state != STATE_EXEC_ABORTED) { boost::mutex::scoped_lock lock(*m_report_mutex); report(); m_condition->notify_one(); setState(STATE_IDLE); } break; case STATE_DEAD: case STATE_EXEC_ABORTED: return; } } while(1); } void Job::init(void) { m_state = STATE_IDLE; } void Job::fork(void) { thr = new boost::thread(boost::bind(execute_handler, this)); } void Job::report_handler(void) { boost::mutex::scoped_lock lock(*m_report_mutex); report(); } void Job::set_foreman_data(boost::mutex* report_mutex, boost::condition* condition) { m_report_mutex = report_mutex; m_condition = condition; } void Job::execute(void) { } void Job::report(void) { } ######## foreman.h: #ifndef FOREMAN_H #define FOREMAN_H 1 #include <boost/thread.hpp> #include <boost/bind.hpp> #include "job.h" template<class JOBB, class JOBB_PROVIDER> class Foreman { JOBB* jb; JOBB_PROVIDER* m_provider; int m_threads; bool m_finished; boost::mutex report_mutex; boost::mutex m_cond_mutex; boost::condition m_condition; boost::thread* me; void handler(void); public: Foreman(int threads, JOBB_PROVIDER* provider); ~Foreman(); void spawn(void); void cleanup(void); void abort(void); bool is_finished(void); }; template<class JOBB, class JOBB_PROVIDER> Foreman<JOBB, JOBB_PROVIDER>::Foreman(int threads, JOBB_PROVIDER* provider) : m_threads(threads), m_provider(provider) { jb = new JOBB[m_threads]; } template<class JOBB, class JOBB_PROVIDER> Foreman<JOBB, JOBB_PROVIDER>::~Foreman() { delete[] jb; } template<class JOBB, class JOBB_PROVIDER> void Foreman<JOBB, JOBB_PROVIDER>::handler(void) { bool all_done; do { boost::mutex::scoped_lock lock(m_cond_mutex); all_done = true; for(int i = 0; i < m_threads; ++i) { int state = jb[i].getState(); if(state == Job::STATE_IDLE) { if(m_provider->provide(jb[i])) { jb[i].doCommand(Job::CMD_EXEC); all_done = false; } else { m_finished = true; } } else { all_done = false; } } if(m_finished || all_done) { return; } m_condition.wait(lock); } while(1); // (!m_finished || !all_done); } template<class JOBB, class JOBB_PROVIDER> void Foreman<JOBB, JOBB_PROVIDER>::spawn(void) { for(int i = 0; i < m_threads; ++i) { jb[i].set_foreman_data(&report_mutex, &m_condition); jb[i].init(); jb[i].fork(); } m_finished = false; // me = new boost::thread(boost::bind(handler, this)); handler(); } template<class JOBB, class JOBB_PROVIDER> void Foreman<JOBB, JOBB_PROVIDER>::cleanup(void) { // me->join(); } template<class JOBB, class JOBB_PROVIDER> void Foreman<JOBB, JOBB_PROVIDER>::abort(void) { } template<class JOBB, class JOBB_PROVIDER> bool Foreman<JOBB, JOBB_PROVIDER>::is_finished(void) { return m_finished; } #endif ######## my_job.h: #ifndef MY_JOB_H #define MY_JOB_H 1 #include "job.h" class MyJob : public Job { int m_suspect; bool is_prime; public: MyJob(); virtual void execute(void); virtual void report(void); void setJobStuff(int suspect); }; class MyJobProvider : public JobProvider { int m_actual; public: MyJobProvider(); bool provide(MyJob& j); }; #endif ######## my_job.cc: #include <iostream> #include <math.h> #include "my_job.h" #include "job.h" MyJob::MyJob() { m_suspect = 0; } void MyJob::setJobStuff(int suspect) { m_suspect = suspect; } void MyJob::execute(void) { int root = (int)sqrt((double)m_suspect); for(int j = 2; j <= root; j++) { if(is_stopped()) { return; } if((m_suspect % j) == 0) { is_prime = false; return; } } is_prime = true; } void MyJob::report(void) { if(is_prime) { std::cout << m_suspect << " is prime" << std::endl; } } MyJobProvider::MyJobProvider() : m_actual(3) {} #define NN 100 bool MyJobProvider::provide(MyJob& j) { j.setJobStuff(m_actual); m_actual++; if(m_actual > NN) return false; return true; } ######## qwe.cc: #include <iostream> #include "job.h" #include "my_job.h" #include "foreman.h" boost::mutex io_mutex; void test1(void) { MyJobProvider p; MyJob j; while(p.provide(j)) { j.execute(); j.report(); } } void test2(void) { MyJobProvider p; Foreman<MyJob, MyJobProvider> fm(7, &p); fm.spawn(); do { } while(!fm.is_finished()); fm.cleanup(); } int main(int argc, char** argv) { test2(); return 0; }

1 0

Serializing a Method
by at forum 07 Oct '07

07 Oct '07

Hi to all I am new to the boost library, I downloaded it and I am doing some tests. I was using MFC serialization. I want to do something like this : template<class _CReceiver, class _CMemberType> class CASetMemberCommand { friend class boost::serialization::access; public: typedef void (_CReceiver::* Action)(_CMemberType) ; protected: _CReceiver *m_pReceiver; _CMemberType m_Property; Action m_fnAction; public: CASetMemberCommand() { m_pReceiver = NULL; } CASetMemberCommand(_CReceiver *pReceiver, Action fnAction, _CMemberType memberProperty) { m_pReceiver = pReceiver; m_fnAction = fnAction; m_Property = memberProperty; } ~CASetMemberCommand() { } virtual void Execute() { if(m_pReceiver) (m_pReceiver->*m_fnAction)(m_Property); } template<class Archive> void serialize(Archive & ar, const unsigned int /* file_version */){ ar & BOOST_SERIALIZATION_NVP(m_pReceiver) //& BOOST_SERIALIZATION_NVP(m_fnAction) & BOOST_SERIALIZATION_NVP(m_Property); } void Save(const char * filename) { std::ofstream ofs(filename); assert(ofs.good()); boost::archive::xml_oarchive oa(ofs); oa << BOOST_SERIALIZATION_NVP(this); } void Load( const char * filename) { std::ifstream ifs(filename); assert(ifs.good()); boost::archive::xml_iarchive ia(ifs); ia >> BOOST_SERIALIZATION_NVP(*this); } }; int main(int argc, char *argv[]) { gps_position pos(32,52,25.0f); CASetMemberCommand<gps_position,int> cmd(&pos,&gps_position::SetMinutes,25); CASetMemberCommand<gps_position,int> cmd4(&pos,&gps_position::SetDegrees,25); cmd.Execute(); SaveCommand(cmd,"D:/test.xml"); cmd.Save("D:/Test2.xml"); CASetMemberCommand<gps_position,int> cmd2; cmd2.Load("D:/Test2.xml"); return 0; } the problem is that BOOST_SERIALIZATION_NVP(m_fnAction) is not working (compile errors); Is there any way to save a methode of w class, so it can be loaded and used later? Thank you

2 3

Allocator and smart_ptrs
by Seweryn Habdank-Wojewódzki 06 Oct '07

06 Oct '07

Dears I found that only shared_ptr can use allocator. Is there any possibility to join the existing allocator and the other smart_ptrs. Why the rest smart_ptrs can not use allocators? Regards. -- |\/\/| Seweryn Habdank-Wojewódzki \/\/

2 4

Need a help about boost.thread
by Steven Woody 06 Oct '07

06 Oct '07

hi, i am doing a job and considering if boost.thread can help. what i am doing is writing some code for other one's use. my code will be called by many threads, and accessing some shared data. my problem is how to keep my code thread-safe. i have no way to control how those other threads were created ( actually, those had already done by another team using windows MFC, and, for some reasons, my code can not use MFC, is has to be portable bwteen Linux and Windows ). in breif, you can think my code in a way of producer/consumer pattern. the producer is a rouine which will be called by many threads ( which are actaully connected sockets created by a listening socket ), and the consumer routine will be called by another thread ( which is actually GUI's main thread ). what i expected is: 1, only one producer can write to the shared data; 2, when a producer is writting, other producers have to busy waiting; and the GUI thread should immediately exit; 3, when GUI thread is reading, no producer can write to the shared data; to my problem, does the boost.thread can help? thanks in advance. -- woody then sun rose thinly from the sea and the old man could see the other boats, low on the water and well in toward the shore, spread out across the current.

3 7

Boost/C++ group on facebook
by Jan Stetka 06 Oct '07

06 Oct '07

Boost/C++ group on facebook: Already have 47 members! Show your support for boost (unofficially) http://www.facebook.com/group.php?gid=4200526117

2 2