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

Re: [Boost-users] [Boost][serialization] Serialization and DLLs
by Topi Mäenpää 04 Nov '05

04 Nov '05

> So the problem is already ratched up to another level of complexity. > > class X can be in a DLL. > class Y can be in an application > > everything is fine so far > > class Z :: public class Y creates a problem in that class Y gets > registered in the application even if its already in the DLL. This will > require some thought to know what the correct way to handle this is. I have a suggestion, please comment. The problem _could_ be in that guid_initializer isn't exported. If I register my base class in a DLL, the DLL interface does not expose the guid_initializer::instance I have created. Since my exe now doesn't know about it, it creates a new definition. The same applies to other DLLs. (This would also explain why the same code works on Linux: all symbols are automatically exported.) What if we exposed the guid_initializer struct to the DLL interface? Would it help? At least it shouldn't do any harm to export it from the exe. -Topi-

2 1

Threads (Not Mutex) Docs?
by Merrill Cornish 04 Nov '05

04 Nov '05

Where is the documentation of Boost Threads--not mutexes, threads? All of the documentation files I can find that say "threads" in the file name are really about mutexes. A lot of different mutexes. Theories and rationales about mutexes. Many, many variations on mutexes, most with a small notation that this variation isn't supported yet. OK, I understand about how to use mutexes to protect access by multiple threads to common variables, but I can't find any documentation--other than reading the source--on how to create, manipulate, and terminate any threads to use mutexes on. While I'm not particularly opposed to reading the source, there are enough subtleties in using threads (see the Boost treatise on mutexes) that I hesitate to go with Boost threads and will probably have to use pthreads-win32. Merrill

1 0

shared_ptr problems
by Filipe Sousa 04 Nov '05

04 Nov '05

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi! I'm trying to create a wrapper around mysql but I have a problem with shared_ptr. Look at the following code: #include <boost/shared_ptr.hpp> #include <iostream> #include <map> struct MYSQL {}; MYSQL* mysql_init() { std::clog << __PRETTY_FUNCTION__ << std::endl; return new MYSQL; } void mysql_close(MYSQL* m) { std::clog << __PRETTY_FUNCTION__ << std::endl; delete m; } class Connection { boost::shared_ptr<MYSQL> m_mysql; public: void connect() { m_mysql.reset(mysql_init(), mysql_close); } bool isConnected() const { return m_mysql.get() != 0; } }; class ConnectionManager { std::map<std::string, Connection> m_connections; public: static ConnectionManager* instance() { static ConnectionManager cm; return &cm; } Connection connection(const std::string& name = "default") { return m_connections[name]; } }; int main() { { Connection c = ConnectionManager::instance()->connection(); std::clog << "isConnected " << c.isConnected() << std::endl; c.connect(); std::clog << "isConnected " << c.isConnected() << std::endl; } Connection c = ConnectionManager::instance()->connection(); std::clog << "isConnected " << c.isConnected() << std::endl; return 0; } And the output is: isConnected 0 MYSQL* mysql_init() isConnected 1 void mysql_close(MYSQL*) isConnected 0 What I don't understand is why shared_ptr calls mysql_close too soon. After all still exists a reference to Connection. I mean, mysql_close should by called when the ConnectionManager singleton is destroyed or am I doing something wrong. - -- Filipe Sousa -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) iD8DBQFDalOEbQdNYqwwwCwRAsNLAJ43BSZTCuraPQ2ouFqGoW/2+JDPDgCeOoQ7 8ZoF4LDO7vQyrPM/Whtx/O8= =iil3 -----END PGP SIGNATURE-----

4 4

a newbie question
by Tuan A. Tran 04 Nov '05

04 Nov '05

Hi, I am a newbie. I would like to run a simple threaded program using boost. I got a sample file see below and compile with the following command g++ simple.cpp -lthread I got error //// In file included from /usr/include/boost/thread/detail/config.hpp:18, from /usr/include/boost/thread/thread.hpp:15, from /usr/include/boost/thread.hpp:15, from simple.cpp:9: /usr/include/boost/config/requires_threads.hpp:47:5: #error "Compiler threading support is not turned on. Please set the correct command line options for threading: -pthread (Linux), -pthreads (Solaris) or -mthreads (Mingw32)" In file included from /usr/include/boost/thread/thread.hpp:19, from /usr/include/boost/thread.hpp:15, from simple.cpp:9: /usr/include/boost/thread/mutex.hpp:58: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:58: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:58: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:58: error: (use `=' to initialize static data members) /usr/include/boost/thread/mutex.hpp:58: error: variable or field `do_lock' declared void /usr/include/boost/thread/mutex.hpp:58: error: declaration of `int boost::mutex::do_lock' /usr/include/boost/thread/mutex.hpp:56: error: conflicts with previous declaration `void boost::mutex::do_lock()' /usr/include/boost/thread/mutex.hpp:59: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:59: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:59: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:59: error: variable or field `do_unlock' declared void /usr/include/boost/thread/mutex.hpp:59: error: declaration of `int boost::mutex::do_unlock' /usr/include/boost/thread/mutex.hpp:57: error: conflicts with previous declaration `void boost::mutex::do_unlock()' /usr/include/boost/thread/mutex.hpp:100: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:100: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:100: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:100: error: variable or field `do_lock' declared void /usr/include/boost/thread/mutex.hpp:100: error: declaration of `int boost::try_mutex::do_lock' /usr/include/boost/thread/mutex.hpp:97: error: conflicts with previous declaration `void boost::try_mutex::do_lock()' /usr/include/boost/thread/mutex.hpp:101: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:101: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:101: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:101: error: variable or field `do_unlock' declared void /usr/include/boost/thread/mutex.hpp:101: error: declaration of `int boost::try_mutex::do_unlock' /usr/include/boost/thread/mutex.hpp:99: error: conflicts with previous declaration `void boost::try_mutex::do_unlock()' /usr/include/boost/thread/mutex.hpp:144: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:144: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:144: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:144: error: variable or field `do_lock' declared void /usr/include/boost/thread/mutex.hpp:144: error: declaration of `int boost::timed_mutex::do_lock' /usr/include/boost/thread/mutex.hpp:140: error: conflicts with previous declaration `void boost::timed_mutex::do_lock()' /usr/include/boost/thread/mutex.hpp:145: error: `cv_state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:145: error: `state' was not declared in this scope /usr/include/boost/thread/mutex.hpp:145: error: invalid data member initialization /usr/include/boost/thread/mutex.hpp:145: error: variable or field `do_unlock' declared void /usr/include/boost/thread/mutex.hpp:145: error: declaration of `int boost::timed_mutex::do_unlock' /usr/include/boost/thread/mutex.hpp:143: error: conflicts with previous declaration `void boost::timed_mutex::do_unlock()' In file included from /usr/include/boost/thread.hpp:19, from simple.cpp:9: /usr/include/boost/thread/once.hpp:35: error: type specifier omitted for parameter `once_flag' /usr/include/boost/thread/once.hpp:35: error: parse error before `&' token In file included from /usr/include/boost/thread.hpp:20, from simple.cpp:9: /usr/include/boost/thread/recursive_mutex.hpp:55: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:55: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:55: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:55: error: variable or field ` do_lock' declared void /usr/include/boost/thread/recursive_mutex.hpp:55: error: declaration of `int boost::recursive_mutex::do_lock' /usr/include/boost/thread/recursive_mutex.hpp:53: error: conflicts with previous declaration `void boost::recursive_mutex::do_lock()' /usr/include/boost/thread/recursive_mutex.hpp:56: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:56: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:56: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:56: error: variable or field ` do_unlock' declared void /usr/include/boost/thread/recursive_mutex.hpp:56: error: declaration of `int boost::recursive_mutex::do_unlock' /usr/include/boost/thread/recursive_mutex.hpp:54: error: conflicts with previous declaration `void boost::recursive_mutex::do_unlock()' /usr/include/boost/thread/recursive_mutex.hpp:103: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:103: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:103: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:103: error: variable or field ` do_lock' declared void /usr/include/boost/thread/recursive_mutex.hpp:103: error: declaration of `int boost::recursive_try_mutex::do_lock' /usr/include/boost/thread/recursive_mutex.hpp:100: error: conflicts with previous declaration `void boost::recursive_try_mutex::do_lock()' /usr/include/boost/thread/recursive_mutex.hpp:104: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:104: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:104: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:104: error: variable or field ` do_unlock' declared void /usr/include/boost/thread/recursive_mutex.hpp:104: error: declaration of `int boost::recursive_try_mutex::do_unlock' /usr/include/boost/thread/recursive_mutex.hpp:102: error: conflicts with previous declaration `void boost::recursive_try_mutex::do_unlock()' /usr/include/boost/thread/recursive_mutex.hpp:154: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:154: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:154: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:154: error: variable or field ` do_lock' declared void /usr/include/boost/thread/recursive_mutex.hpp:154: error: declaration of `int boost::recursive_timed_mutex::do_lock' /usr/include/boost/thread/recursive_mutex.hpp:150: error: conflicts with previous declaration `void boost::recursive_timed_mutex::do_lock()' /usr/include/boost/thread/recursive_mutex.hpp:155: error: `cv_state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:155: error: `state' was not declared in this scope /usr/include/boost/thread/recursive_mutex.hpp:155: error: invalid data member initialization /usr/include/boost/thread/recursive_mutex.hpp:155: error: variable or field ` do_unlock' declared void /usr/include/boost/thread/recursive_mutex.hpp:155: error: declaration of `int boost::recursive_timed_mutex::do_unlock' /usr/include/boost/thread/recursive_mutex.hpp:153: error: conflicts with previous declaration `void boost::recursive_timed_mutex::do_unlock()' ///////////////////////// If I use commend g++ simple.cpp -pthread I got error //////// /tmp/cchxK5CW.o(.text+0x66): In function `main': : undefined reference to `boost::thread::thread[in-charge](boost::function0<void, std::allocator<boost::function_base> > const&)' /tmp/cchxK5CW.o(.text+0x9d): In function `main': : undefined reference to `boost::thread::join()' /tmp/cchxK5CW.o(.text+0xa8): In function `main': : undefined reference to `boost::thread::~thread [in-charge]()' /tmp/cchxK5CW.o(.text+0xc2): In function `main': : undefined reference to `boost::thread::~thread [in-charge]()' collect2: ld returned 1 exit status /////////// I do not know what went wrong. Can anyone tell me what I did wrong? How should I correct so that I can compile this example? TIA Tuan ========== #include <boost/thread.hpp> #include <iostream> using namespace std; void ciao() { cout << "Ciao, thread using boost" << endl; } /******** this is the main thread's code */ int main(int argc,char *argv[]) { boost::thread thrd(&ciao); thrd.join(); return(0); }

2 2

serialization binary_oarchive without signature?
by relative＠gmx.net 03 Nov '05

03 Nov '05

i'd like to use boost serialization to serialize > 1 independent objects to a large file through a seekable stream, i.e. > 1 serialization archives in one file. can you turn off the archive signature("serialization::archiv") at the beginning of the output, or do I have to write my own archive class for that? and 2nd, is there a way(except serializing to a std::stringstream) to find out the archive size beforehand, e.g. an archive class which doesn't really serialize things but add up the size in bytes of the binary output? thanks, -- 10 GB Mailbox, 100 FreeSMS/Monat http://www.gmx.net/de/go/topmail +++ GMX - die erste Adresse f�r Mail, Message, More +++

2 1

A question about boost::graph iterators
by Jeffrey Holle 03 Nov '05

03 Nov '05

I'm using gcc 4.0.2 with boost 1.33.0. I'm getting a compiler warning about taking the address of a temporary with the following code: Edge* TransshipmentPivotingRule::enteringEdge(const FeasibleSpanningTree* pFST) { graph_traits<AntiTree>::edge_iterator ei,ei_out; tie(ei,ei_out)=edges(*m_pAntiTree); return (find_if(ei,ei_out, ProfitableEdge(*m_pAntiTree,m_c,m_y,*pFST))!=ei_out) ? &*ei : NULL; } Note the "&*ei". I know that the iterator is tempory, but what the iterator is pointing to isn't. Is this a legit warning that I need to be concerned about? In case it matters, AntiTree is a filtered_graph.

2 2

Thread Pools
by Brian Cox 03 Nov '05

03 Nov '05

Hello, I read a few years ago that Bill Kempf was going to develop a thread pool for Boost.Threads. Does anyone know the status of that? Does anyone have any good examples of a thread pool or worker queue developed using Boost.Threads? Thanks for any info. -Brian-

1 0

[Boost][serialization] link problems with MSVC.NET/no bjam
by Topi Mäenpää 03 Nov '05

03 Nov '05

Hi, I have struggled to get the boost libraries link into my application with no success so far. I was able to compile most of the demo programs with bjam, but it seems that only static linking is used there. I'm using the dynamic versions. The libraries compiled fine with MSVC.NET 7.1. However, whenever I try to link them to an application, every symbol seems to be undefined. I turned on the diagnostics to see that the automatic linking is happening (with the nonstandard preprocessor stuff). I also added the libraries to my project manually without success. The funny (?) thing is that the lib file has a definition for the "missing" functions, except for one minor difference. This is the mangled symbol the linker complains about: __imp_?save@?$xml_oarchive_impl@Vxml_oarchive@archive@boost@@@archive@boost@@QAEXABV?$basic_string@GU?$char_traits@G@std@@V?$allocator@G@2@@std@@@Z (Omitting namespaces this one decrypts to xml_oarchive_impl<xml_oarchive>::save(basic_string& const)) This is what seems to be in the library: __imp_?save@?$xml_oarchive_impl@Vxml_oarchive@archive@boost@@@archive@boost@@QAEXABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z The only difference is that the library symbol contains a "D" and the linker is looking for "G" in a couple of places. I have found no way of going any further. Since the demos are built in debug mode and with static libraries, I found no useful information by inspecting the commands bjam uses in building them. Everything is just like I have in my own project. I have defined BOOST_SERIALIZATION_DYN_LINK, if you happen to think about that. I even managed to link my app with mingw, but got strange run-time errors. Everything works fine with gcc on Linux. Any advice? -Topi-

3 4

[Boost][serialization] Serialization and DLLs
by Topi Mäenpää 03 Nov '05

03 Nov '05

Hi, Some time ago, we had a discussion about serialization of dynamically loaded classes. I got my code work on Linux, but still cannot do the same on Windows. I'd like someone to elaborate the stuff below a bit: Robert Ramey wrote: > Another post has me convinced that the problem is that there are multiple > copies of a base class extended_type_info record which I believe > shouldn't occur. I can confirm that this indeed happens. > If you trap your debugger the serialization.dll (shared_lib) at line 58 of > extened_type_info.cpp you should trap every time an extended_type_info > record is "self_registered". Examination of the stack should reveal > where this is being called from. I expect that the base class is being > self registered from both the DLL as well as the main executable. Whenever I register a class from a DLL, its base classes get registered as well, although they already have been registered. If I load multiple DLLs that all contain derived classes, the base classes are registered many times, but only once for each DLL. > My current thinking is that this behavior should be prohibited and > detected with an exception. Other users have addressed this > by tweaking the code in extended_type_info.cpp to permit > multiple "self registration" of the same extended_type_info type. > This eliminated their symptoms and everything seems to work. > But from my understanding of the library, I believe this just > hides a deeper problem. Can someone point out the details? How should this be done? Any other suggestions? Best regards, Topi

2 1

Newbie help needed
by Ricardo Kirkner 03 Nov '05

03 Nov '05

Hi: I am trying to write python bindings for an already existant c++ library. I started using pyste, because I felt this was the easiest way to do it. However I keep getting errors (mostly I think because of string and vector arguments conversion). I just wanted to ask if I am on the right track here. Maybe someone can give me some pointers on how to make this task easier? (I was tempted to switch to pyrex, swig, etc, but everything seems to have its own problems.) Should I use pyste, or should I rather write boost.python code directly? Thank you, Ricardo

2 1