February 2006 - Boost-users - lists.preview.boost.org

multiple implicit constructors (MPL?)
by Cory Nelson 10 Feb '06

10 Feb '06

Hello, I have a class, large_integer<BaseT>. I build 256bit integers by doing: typedef large_integer<unsigned long long> uint128; typedef large_integer<uint128> uint256; Unfortunately C++ only seems to allow one implicit conversion, so I'm unable to do uint256 a = 0, and have to do uint256 a = uint128(0); instead. I'm able to get around one of these by having a constructor which takes the base type, but that only gets me one free conversion, meaning a uint512 has the same issue. Is there a way around this (I'm guessing MPL would be my best bet?)? -- Cory Nelson http://www.int64.org

1 0

[thread] Mysterious deadlock
by David Greene 10 Feb '06

10 Feb '06

I have this simple testcase (below) which runs into deadlock (full run trace after the testcase): ... [Producer] Producer generated data 28 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 29 [Producer] push(): about to lock [Producer] push(): locked [Consumer] Signaled for data [Consumer] empty(): about to lock <<DEADLOCK>> The call to empty() by Consumer is after the first time it is woken up from full_condition. So the deadlock occurs right when the consumer first attempts to grab some data. Is there something I'm not understanding about how mutexes and locks work in Boost.Thread? I feel like I'm missing something obvious. This comes out of a larger and more complicated program but I was able to isolate the mysterious (to me) behavior I'm seeing. I used to use LOG for debugging but I thought there might be some interaction between the queue mutex locks and the print mutex locks so I switched to the non-locking LOG_NOLCK but observed no change in behavior. Any help is greatly appreciated. -Dave --------------------------------- Testcase #include <iostream> #include <queue> #include <map> #include <string> #include <boost/thread.hpp> #include <boost/thread/recursive_mutex.hpp> #include <boost/thread/condition.hpp> #include <boost/bind.hpp> #include <pthread.h> typedef boost::recursive_mutex mutex_type; typedef mutex_type::scoped_lock lock_type; // Atomic prints mutex_type print_mutex; mutex_type print_map_mutex; // For print/debug purposes typedef std::map<pthread_t, std::string> thread_map_type; thread_map_type thread_map; template<typename T> void register_thread_id(const T &entity) { lock_type lock(print_map_mutex); std::cout << "Registering " << pthread_self() << " to " << entity.name() << std::endl; thread_map[pthread_self()] = entity.name(); } #define LOG(msg) \ { \ lock_type lock(print_mutex); \ std::cout << "[" << thread_map[pthread_self()] \ << "] " << msg << std::endl; \ } #define LOG_NOLCK(msg) \ { \ std::cout << "[" << thread_map[pthread_self()] \ << "] " << msg << std::endl; \ } // Synchronize start (make sure all registered) mutex_type go_mutex; boost::condition go_condition; bool ready = false; class Consumer { private: boost::thread thread; std::queue<int> queue; typedef boost::recursive_mutex mutex_type; typedef mutex_type::scoped_lock lock_type; mutable mutex_type mutex; mutable mutex_type full_mutex; mutable boost::condition full_condition; void wait_until_ready(void) const { ::lock_type lock(go_mutex); while(!ready) { go_condition.wait(lock); } LOG_NOLCK("go!"); }; void wait_for_data(void) const { lock_type lock(full_mutex); while(empty()) { LOG_NOLCK("Waiting for data"); full_condition.wait(lock); LOG_NOLCK("Signaled for data"); } LOG_NOLCK("Wakeup for data"); }; void event_loop(void) { register_thread_id(*this); wait_until_ready(); while(true) { wait_for_data(); LOG_NOLCK("Consumer received data " << top()); pop(); } }; public: Consumer(void) : thread(boost::bind(&Consumer::event_loop, this)) {}; std::string name(void) const { return("Consumer"); }; void push(int val) { LOG_NOLCK("push(): about to lock"); { lock_type lock(mutex); LOG_NOLCK("push(): locked"); queue.push(val); lock_type full_lock(full_mutex); LOG_NOLCK("push(): about to unlock"); } full_condition.notify_all(); }; int top(void) const { LOG_NOLCK("top(): about to lock"); lock_type lock(mutex); LOG_NOLCK("top(): locked"); int item = queue.front(); LOG_NOLCK("top(): about to unlock"); return(item); }; void pop(void) { LOG_NOLCK("pop(): about to lock"); lock_type lock(mutex); LOG_NOLCK("pop(): locked"); queue.pop(); LOG_NOLCK("pop(): about to unlock"); }; bool empty (void) const { LOG_NOLCK("empty(): about to lock"); lock_type lock(mutex); LOG_NOLCK("empty(): locked"); bool val = queue.empty(); LOG_NOLCK("empty(): about to unlock"); return(val); }; }; class Producer { private: boost::thread thread; void wait_until_ready(void) const { lock_type lock(go_mutex); while(!ready) { go_condition.wait(lock); } LOG_NOLCK("go!"); }; void event_loop(void) { int i = 0; register_thread_id(*this); wait_until_ready(); while(true) { LOG_NOLCK("Producer generated data " << i); consumer->push(i); ++i; }; }; Consumer *consumer; public: Producer(Consumer *c) : consumer(c), thread(boost::bind(&Producer::event_loop, this)) {}; std::string name(void) const { return("Producer"); }; }; void go(void) { // Wait until everyone arrives lock_type lock(go_mutex); ready = true; // Notify everyone go_condition.notify_all(); } int main(void) { Consumer c; Producer p(&c); go(); while(true); return(0); } ----------------- Ouput Trace [Consumer] empty(): about to unlock [Consumer] Waiting for data Registering 48958384 to Producer [Producer] go! [Producer] Producer generated data 0 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 1 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 2 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 3 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 4 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 5 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 6 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 7 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 8 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 9 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 10 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 11 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 12 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 13 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 14 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 15 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 16 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 17 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 18 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 19 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 20 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 21 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 22 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 23 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 24 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 25 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 26 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 27 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 28 [Producer] push(): about to lock [Producer] push(): locked [Producer] push(): about to unlock [Producer] Producer generated data 29 [Producer] push(): about to lock [Producer] push(): locked [Consumer] Signaled for data [Consumer] empty(): about to lock <<DEADLOCK>>

4 5

unit test framework problems
by Niklas Wiberg 09 Feb '06

09 Feb '06

Hi, I'm having problems with the unit test framework when I try to run it in release mode. My compiler is VC++ 6 sp6. The unit test executable I'm building is "Using MFC as a shared DLL" and is set to use the "Multithreaded DLL" runtime library. The test subject code that is linked to this app is built with the same settings. I'm linking with libboost_unit_test_framework-vc6-mt-1_33.lib. The problem is that when the unit test starts to execute it fails with an access violation. In debug build it works fine. I enabled some debug info in the release build and found that what seems to be happening is a deallocation of "invalid" heap memory, which I'm guessing could be a result of different runtime libs being used. The stack trace at the point of the crash is included below. Has anyone had the same kind of problem and perhaps found a solution? Best regards, Niklas Wiberg NTDLL! 7c901230() NTDLL! 7c96cd80() NTDLL! 7c960af8() KERNEL32! 7c85e7af() _CrtIsValidHeapPointer(const void * 0x003b2418) line 1697 _free_dbg_lk(void * 0x003b2418, int 1) line 1044 + 9 bytes _free_dbg(void * 0x003b2418, int 1) line 1001 + 13 bytes free(void * 0x003b2418) line 956 + 11 bytes operator delete(void * 0x003b2418) line 7 + 10 bytes std::allocator<char>::deallocate(void * 0x003b2418, unsigned int 33) line 64 + 16 bytes std::basic_string<char,std::char_traits<char>,std::allocator<char> >::_Tidy(unsigned char 1) line 592 std::basic_string<char,std::char_traits<char>,std::allocator<char> >::~basic_string<char,std::char_traits<char>,std::allocator<char> >() line 59 + 17 bytes boost::filesystem::path::m_path_append(const std::basic_string<char,std::char_traits<char>,std::allocator<char> > & {0x0032f9f1 "D:\dev\UnitTest"}, unsigned char (const std::basic_string<char,std::char_traits<char>,std::allocator<char> > &)* 0x0054f9b3 boost::filesystem::native(const std::basic_string<char,std::char_traits<char>,std::allocator<char> > &)) line 351 boost::filesystem::path::path(const char * 0x0032f988, unsigned char (const std::basic_string<char,std::char_traits<char>,std::allocator<char> > &)* 0x0054f9b3 boost::filesystem::native(const std::basic_string<char,std::char_traits<char>,std::allocator<char> > &)) line 220 + 37 bytes boost::filesystem::current_path() line 803 + 22 bytes boost::filesystem::initial_path() line 810 + 28 bytes init_unit_test_suite(int 1, char * * 0x003b2da0) line 19 + 5 bytes UNITTEST! boost::unit_test::framework::init(int,char * * const) + 201 bytes UNITTEST! main + 54 bytes UNITTEST! mainCRTStartup + 227 bytes KERNEL32! 7c816d4f()

2 3

[iostreams] Hang when constructing an xml_iarchive with non-existent source
by Paul Giaccone 09 Feb '06

09 Feb '06

I have a function that uses a chain of filters that I am using to read encrypted, bzipped data from file. It is working fine. However, I have just discovered that if I call my function with a file that does not exist, the code hangs in the constructor for xml_iarchive. Of course, I have since edited my code to trap non-existent files before the code gets that far, but I am surprised that: * is_complete() returns true when the source is a non-existent file * the constructor for xml_iarchive hangs when it is given a stream that has a non-existent file as its source Is this correct behaviour or a bug? Of course, it is the programmer's responsibility to ensure the source exists, but a hang rather than an exception seems suggests to me that something has been overlooked in the library. The (edited) function exhibiting this behaviour is given below. I am using gcc 3.2.3 and Boost 1.33.1. Paul void load(const std::string& filename) { //my revised tests "filename" here and bails out if the file does not exist try { boost::iostreams::filtering_istream in; in.push(boost::iostreams::bzip2_decompressor()); in.push(MyInputEncryptionFilter()); in.push(boost::iostreams::bzip2_decompressor()); in.push(boost::iostreams::file_source(filename, BOOST_IOS::in | BOOST_IOS::binary)); if (!in.is_complete()) { std::cerr << "Error creating filtering input stream" << std::endl; return; } boost::archive::xml_iarchive input_archive(in); //deserialise file here return; } //various catches removed }

2 3

[bgl]
by Marcio Paim de Aquino 09 Feb '06

09 Feb '06

Hi, I'm trying to compile this example http://www.boost.org/libs/graph/example/johnson-eg.cpp. I have the following error from VC++ 6 SP2: --------------------Configuration: johnson-eg - Win32 Debug-------------------- Compiling... johnson-eg.cpp c:\arquivos de programas\microsoft visual studio\vc98\include\boost\graph\graph_concepts.hpp(479) : error C2664: 'const_constraints' : cannot convert parameter 1 from 'int [5][5]' to 'const int (&)[5][5]' Reason: cannot convert from 'int [5][5]' to 'const int [5][5]' There is no context in which this conversion is possible c:\arquivos de programas\microsoft visual studio\vc98\include\boost\graph\graph_concepts.hpp(477) : while compiling class-template member function 'void __thiscall boost::BasicMatrixConcept<int [5][5],unsigned int,int>::constraints(void)' Error executing cl.exe. johnson-eg.exe - 1 error(s), 0 warning(s) Does anyone know what's wrong? Regards Marcio Paim

2 1

Re: [Boost-users] [Serialization]: bug while serializing a std::string
by RIVASSEAU Jean Noel 09 Feb '06

09 Feb '06

I am assuming you meant : std::ofstream ifs("container.mob"); boost::archive::text_oarchive ia(ofs); oa << MyObject; ofs << std::endl; and not std::ifstream ifs. So if I understand well, this "patch" only adds a newline to the archive at the saving stage. I tried it but it did not help in any way; in fact I cannot see how it could help, since the problem is at the point of the loading of the std::string, not at the end of the load of the whole object. What I mean here is that MyObject is, in my case, a class containing a std::vector of pointers to class containing std::strings. One of these strings is too long, which causes the crash. Maybe what you meant is appending the std::endl after the serialization of the std::string, but this is not possible since at this point I have only access to the archive, not the underlying ofstream : template<class Archive> void serialize(Archive & ar, const MyClass& m, const unsigned int version) { ar & m.a_standard_string; } I do not have access there to the underlying ofstream of the implementation. Note that here, MyObject would contain a std::vector < MyClass * >. And I am serializing directly MyObject which causes the serialization of MyClass later. So, what do you mean? Have you located the bug ? _____ De : boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] De la part de Robert Ramey Envoyé : jeudi 9 février 2006 04:46 À : boost-users(a)lists.boost.org Objet : Re: [Boost-users] [Serialization]: bug while serializing astd:string Untl 1.34 comes out try the following on the output stream std::ifstream ofs("container.mob"); boost::archive::text_oarchive ia(ofs); oa << MyObject; ofs << std::endl; "RIVASSEAU Jean Noel" <JN.RIVASSEAU(a)oberthurcs.com> wrote in message news:87F60F7FA02FF749AFB02BD9FCAA6B04C87411@naserv31.nanterre.oberthurcs.com... I have been able to test with a string stream; the bug does not occur. That is, this code produces the bug: std::ifstream ifs("container.mob"); boost::archive::text_iarchive ia(ifs); ia & MyObject; while the following does not: std::ifstream ifs("container.mob"); std::ostringstream tmp; tmp << ifs.rdbuf(); std::istringstream is(tmp.str()); boost::archive::text_iarchive ia(is); ia & MyObject; a) try using a different stream buffer - strstream So do you have any clue, Robert ?? Clearly the problem has to do with buffering somewhere.... Jean-Noël _____ _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

pd darwin Symbol not found
by Locatelli Jérôme 09 Feb '06

09 Feb '06

I'am writing a application on MacOS 10.4.4 using darwin 8 with gcc 4, I compile and link my application without problem but when I run my application , I've got the error bellow. Has anybody got an idear?? Thanks. dyld: lazy symbol binding failed: Symbol not found: __ZN5boost6detail3tss4initEPNS_9function1IvPvSaINS_13function_baseEEEE Referenced from: /Users/locatelli/Projets/ics/IcsServer/obj.ics/../lib/ libIcsServerCommandSystem.0 Expected in: flat namespace dyld: Symbol not found: __ZN5boost6detail3tss4initEPNS_9function1IvPvSaINS_13function_baseEEEE Referenced from: /Users/locatelli/Projets/ics/IcsServer/obj.ics/../lib/ libIcsServerCommandSystem.0 Expected in: flat namespace Trace/BPT trap --------------------------------------------------- Locatelli Jérôme Institut Laue-Langevin Service Contrôle des instruments 6 rue Jules Horowitz, BP 156 38042 Grenoble cedex 9 http://www.ill.fr/ Tel: 33 (0)4 76 20 76 27 Fax: 33 (0)4 76 20 77 00 ---------------------------------------------------

1 0

[any] Converting between types
by Iulian M 09 Feb '06

09 Feb '06

Hello, I'm trying to use boost::any to store a group of types. For value types everything works as expected. The problem is that i need to store some polymorphic types also. The following code explains what i'm trying to accomplish. #include <boost/any.hpp> class A {}; //base calss class B: public A {}; // derived class int main() { B *b = new B(); //create derived class boost::any n(b); // place inside boost::any /* other part of code where i only know n contains a pointer to a class derived from A this will throw bad_any_cast: typeid(A*) != typeid (B*) */ A* a = boost::any_cast<A*>(n); } Is there any way of making boost::any_cast work work ok with "convertible" types? Or is there a way of storing different types of objects and having the possibility of retrieving them as compatible objects ? I've looked into any.hpp and found that it uses a typeid() equality test to see if the cast can be made. Can't there be a is_convertible test insted , at least for the case when is holding pointer to objects? Iulian -- Blessed is the man who, having nothing to say, abstains from giving wordy evidence of the fact. -- George Eliot

3 6

A problem with '^' and '$' in boost regex
by Alan Huang 09 Feb '06

09 Feb '06

Hi, In boost document I saw that the '^' matches all the blank chars in the begin of line and the '$' matches all the blank chars in the end of line. Just like regex( "^abc$" ) can't match the string " abc ". What's wrong? -- Yours Sincerely, Alan Huang

3 3

Re: [Boost-users] [GRAPH] Is the HP aCC compiler supported?
by Oliver.Kowalke＠infineon.com 09 Feb '06

09 Feb '06

Hello, aCC V6 is only available on HP/UX 11 Itanium systems. On HP/UX 11.11 PA-RISC only the old 3.x versions are available which are unable to compile boost libraries. I use gcc 3.4.4 on HP/UX together with boost - works fine. Oliver >The aCC V6 compiler is not listed as a supported compiler with boost >because nobody is running boost regression tests with this compiler, >but we are working on this... > >I ran boost regression tests with aCC A.06.06 for the latest boost >release: 1.33.1 and the results were on par with that for the supported >compilers. Specifically, for the graph library, 32 tests succeeded and >two tests failed to compile: graphviz_test and layout_test. > >graphviz_test failed to compile with: > >"/opt/aCC/include_std/utility", line 103: error #2144: a value of type > "const char *" cannot be used to initialize an entity of type > "std::pair<char [2], char [2]>::first_type" > >layout_test failed to compile because the compiler ran out of memory. > >I used extended acc jamfile with some compilation warnings disabled, >like "#2340-D: value copied to temporary, reference to temporary used", >"#2068-D: integer conversion resulted in a change of sign", "#2111-D: >statement is unreachable", ...etc. > >I suggest, you try BGL with aCC V6. > >Boris Gubenko >HP C++ team > >----- Original Message ----- >From: "rohit wadhwa" <rohit.wadhwa(a)teleatlas.com> >To: <boost-users(a)lists.boost.org> >Sent: Wednesday, February 08, 2006 12:01 PM >Subject: [Boost-users] [GRAPH] Is the HP aCC compiler supported? > > >> Can I use BGL with HP aCC compiler? If so, what version? >> >> Any help is really appreciated. >> >> Rohit. >> >> >> _______________________________________________ >> Boost-users mailing list >> Boost-users(a)lists.boost.org >> http://lists.boost.org/mailman/listinfo.cgi/boost-users >> >> > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0