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

Re: [Boost-users] [Thread] destructor called several times
by elconio＠commonworld.info 07 Oct '06

07 Oct '06

Hi, thanks for the explanations and the quick answer ! Conio.

1 0

[Thread] destructor called several times
by elconio＠commonworld.info 07 Oct '06

07 Oct '06

Hello, I am using boost::thread for a program, but it behaved in an unexpected way (to me at least). I extracted a small case that shows my problem. I searched in the archives and in the docs, but I couldn't figure this out. With the following code [1] I expect the constructor of my the class MyThread to be called only once and yet on my machine it is called 11 times! [2] The problem is that I was freeing some ressources from the destructor and that even before the operator() was called sometimes. I am using the CVS version from 3 days ago on a linux x86. Thanks for any help or any pointer ! [1]: #include <iostream> #include <boost/thread/thread.hpp> class MyThread { public: MyThread () { std::cerr << "Constructor called.\n"; } ~MyThread() { std::cerr << "Destructor called " << (++id) << ".\n"; } void operator()() { std::cerr << "Operator() called.\n"; } private: static int id; }; int MyThread::id = 0; int main() { MyThread * pt = new MyThread(); boost::thread th( *pt ); th.join(); std::cerr << "\nEnd\n"; delete pt; } [2] Constructor called. Destructor called 1. Destructor called 2. Destructor called 3. Destructor called 4. Destructor called 5. Destructor called 6. Destructor called 7. Destructor called 8. Destructor called 9. Operator() called. Destructor called 10. End Destructor called 11. -- Conio.

3 2

[review-results] pimpl_ptr
by Rene Rivera 07 Oct '06

07 Oct '06

Unfortunately I must say that the pimpl_ptr library will not be part of Boost at this time. The results where decidedly against the library in the proposed form. A few points about the library where clear: * Many want to see a pimpl_ptr in Boost. * The proposed design was considered to not do enough towards addressing the handle-body idiom within C++. * Some of the ideas, the policies, where both liked and disliked. Which raised the specter of the lack of a policy pointer. Hopefully we will see this idiom addressed in some near future. Last, I must apologize for the late results. But I must thank Ronald for the gentle prodding :-) -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

1 0

[Threads] Three Issues with Boost.Threads
by Daniel Winsor 07 Oct '06

07 Oct '06

Let's say I wanted to create my own thread pool class. I attempted this at first by making a thread manager class and in its constructor creating boost::thread objects and letting the objects expire naturally, when they went out of scope. The threads themselves and its function seemed to be running fine, but when I started encountering problems I wondered if the thread object needed to remain in scope. So I switched to creating boost::thread objects on the heap with new, but this didn't solve my problems, and by doing it this way I found out about the mem leaks. So my first question is what is the best way to make a thread pool, and are there any pratfalls to watch out for? I made what I thought was a simple input / output scheme for the thread functions. Please see the code attached. The thread functions sleep while waiting on the main thread to add data to an input queue. The thread then pops it, and locks a mutex while executing. When done, it puts the data on the output queue (order doesn't matter) and the mutex is unlocked. The main thread is checking for when the input queue is empty, and when it is, it tries to lock all the thread mutexes. Although the input queue may be empty, the thread may still be executing and the output would not be full, thus the thread mutex. The problem I was getting was that the thread manager seemed to be ignoring mutexes. You can see the problem in action with the attached code. What you are looking for is when the size of the output is not what you put in. You will need to run the code many times (~30) because the error is infrequent. A much more common error (1 in 3) I am getting is that I get an access violation in boost::mutex::scoped_lock. I don't know if they are related or what. I found if I have a memory intensive app in the background then there aren't as many access violations, so run one while looking for the other bug or you will go crazy. So basically I am new to threading and I don't know if these are bugs in my code, the boost::threads library, or what. Any suggestions on the abstraction of the thread pool or anything else is welcome. Boost 1.33.1 Windows XP x64 (win32 debug app) Intel Core 2 Duo

1 1

Re [Threads] Three Issues with Boost.Threads - Typo
by Daniel Winsor 07 Oct '06

07 Oct '06

There is a typo in the demo code. boost::mutex::scoped_lock lock3(Thread1Lock) should be Thread3Lock, and the same for 4. However, I still get the same errors as before.

1 0

[Serialization] un-archiving container of strings surfaces malloc/free bug
by Mark Melton 06 Oct '06

06 Oct '06

The following program archives a vector of one string to the file "/ tmp/foo" and then immediately un-archives "/tmp/foo" into an empty vector of strings. Deep in the bowels of the un-archiving process, malloc reports a bad call to free as noted in the comments at the end of the program. The bad call to free is made during a call to string.reserve(). Varying the program shows that the bad free occurs once for each element in the vector and is always the same address. The behavior is consistent across text, binary and xml archivers as well as vectors and maps of strings. Under different conditions the same error occurs during a call to string.assign(). I have tried gcc3.3/4.0 and 4.1 with no change in the results. My current work- around is to use vector<char> instead of string. I am hoping that someone will be willing to help me debug this problem. I am running on a PPC Mac using the darwin ports development environment: Darwin mars.notlem.com 8.7.0 Darwin Kernel Version 8.7.0: Fri May 26 15:20:53 PDT 2006; root:xnu-792.6.76.obj~1/RELEASE_PPC Power Macintosh powerpc boost @1.33.1_0 gcc41 @4.1.1_0+darwin_8 cheers, mark ======================================================================== = #include <iostream> #include <fstream> #include <vector> #include <string> #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/serialization/utility.hpp> #include <boost/serialization/vector.hpp> #include <boost/serialization/string.hpp> using namespace std; int main(int argc, char const **argv) { typedef vector<string> string_vector_t; string_vector_t output, input; output.push_back("a string"); { ofstream ofs("/tmp/foo"); assert(ofs.good()); boost::archive::xml_oarchive archiver(ofs); archiver << BOOST_SERIALIZATION_NVP(output); } { ifstream ifs("/tmp/foo"); assert(ifs.good()); boost::archive::xml_iarchive archiver(ifs); archiver >> BOOST_SERIALIZATION_NVP(input); } return 0; } // $ ./bug // bug(27658) malloc: *** Deallocation of a pointer not malloced: 0x209c33c; This could be a double free(), or free() called with the middle of an allocated block; Try setting environment variable MallocHelp to see tools to help debug ======================================================================== ======== #0 0x9000621c in free () #1 0x94afb8ec in std::string::_Rep::_M_destroy () #2 0x94afe6d0 in std::string::reserve () #3 0x0007617c in boost::archive::basic_xml_grammar<char>::my_parse () #4 0x000769b8 in boost::archive::basic_xml_grammar<char>::parse_start_tag () #5 0x00076c70 in boost::archive::basic_xml_iarchive<boost::archive::xml_iarchive>::load_s tart () #6 0x00004fc4 in boost::archive::detail::iserializer<boost::archive::xml_iarchive, std::vector<std::string, std::allocator<std::string> > >::load_object_data (this=0x300eb0, ar=@0xbfffe9d8, x=0xbfffed00, file_version=62) at /opt/local/include/boost/archive/ basic_xml_iarchive.hpp:79 #7 0x000638e8 in boost::archive::detail::basic_iarchive::load_object () #8 0x00002d3c in main (argc=3149488, argv=0xbfffe9d8) at /opt/local/ include/boost/archive/detail/iserializer.hpp:375

1 0

Serialization of char*
by Deane Yang 06 Oct '06

06 Oct '06

The code below won't compile for me using VC8. Can someone explain? I'm sure I'm doing something dumb, but I'm too dumb to know what. #include <boost/archive/xml_oarchive.hpp> #include <iostream> #include <fstream> struct Object { char* ptr; }; int main (int argc, char* argv[]) { try { std::cout << "Saving portfolio to archive...\n"; Object object; object.ptr = new char[10]; for (int i= 0; i < 10; ++i) { object.ptr[i] = 'i'; } std::ofstream ofs(filename); assert(ofs.good()); boost::archive::xml_oarchive oa(ofs); const Object& obj(object); oa << BOOST_SERIALIZATION_NVP(obj); } catch (const xml_schema::exception& e) { std::cerr << e << std::endl; return 1; } }

1 2

is there limitation on the file size generated by serialization
by Cheok Yan Cheng 06 Oct '06

06 Oct '06

hello all, we have been using boost serialization library for quite some time in small scale project. so far, we are happy with boost serialization, since it offers several features which are not found in MFC framework. before we migrate boost serialization library is large scale project, can we know 1. is there any limitation on the file size generated by serialization. i.e. what is the maximum file size which can be written by boost serialization? 2. is there any limitation on the object size, which can be loaded/ unloaded by boost serialization? 3. does boost serialization offer any way (checksum?) to ensure the file being written is not corrupted? thank you very much yccheok __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

problem using boost::archive::xml_oarchive to serliaze a map
by Thomas D\\\'Silva 06 Oct '06

06 Oct '06

Hello, I am trying to use boost::archive::xml_oarchive to serialize a vector<map<string, map<string,int> > >. When I examine the serialized xml file, I see that some of the map elements are duplicated which causes an assertion when I try to load the file because a map cannot have two elements with the same key. I would appreciate any insight into this error. Thank you Thomas D'Silva /include/boost/serialization/collections_load_imp.hpp:69: void boost::serialization::stl::archive_input_map<Archive, Container>::operator()(Archive&, Container&) [with Archive = boost::archive::xml_iarchive, Container = std::map<std::string, int, std::less<std::string>, std::allocator<std::pair<const std::string, int> > >]: Assertion `result.second' failed. #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> #include <boost/serialization/vector.hpp> #include <boost/serialization/map.hpp> #include <boost/serialization/string.hpp> #include <boost/serialization/nvp.hpp> using namespace std; using namespace boost::filesystem; class DistributionTable { private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(neTable); ar & BOOST_SERIALIZATION_NVP(mappedTable); } vector<map<string,map<string,int> > > neTable; vector<map<string,map<string,int> > > mappedTable; public: DistributionTable(){}; }; void saveTable(const DistributionTable &t, string filename) { ofstream ofs(filename.c_str()); boost::archive::xml_oarchive oa(ofs); oa <<BOOST_SERIALIZATION_NVP(t); ofs.flush(); } void loadTable(DistributionTable &t, string filename) { ifstream ifs(filename.c_str()); boost::archive::xml_iarchive ia(ifs); ia >>BOOST_SERIALIZATION_NVP(t); }

1 0

Ann: Statistical distributions / Mathematical Special Functions
by John Maddock 06 Oct '06

06 Oct '06

This package consists of 3 related components, and we're now seeking further feedback on these before the final push prior to a review: Statistical Distributions ~~~~~~~~~~~~~~~~~~~~~~~~~ Following feedback from the previous preview, we have re-organised so that distributions are now C++ classes. See the docs and especially the tutorial for illustrations on how this will be useful in real life. We would especially welcome feedback from those with experience of using statistics, and indeed from professional statisticians who are probably not Boosters. Mathematical Special Functions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ These are currently focused on those functions that are useful for statistical applications: Floating point classification. Gamma/lgamma/digamma/beta/erf/erfc Ratios of gamma functions Factorials The incomplete gamma and beta functions. The inverses of the gamma, beta, and error functions. Derivatives of the incomplete gamma and beta functions. Compared to the previous release, most of the changes involve accuracy and quality of implementation issues. Error handling has now been revised to provide a comprehensive package-wide error handling scheme complete with meaningful error messages. Toolkit ~~~~~~~ Provides tools to assist in the implementation of numerical methods: Infinite series evaluation. Continued fraction evaluation. Polynomial and rational function evaluation. Root finding with derivatives (Newton/Halley/Schroeder methods). Root finding without derivatives. Function minimisation. Misc Tools ~~~~~~~~~~ These are strictly experimental, but are used in the development of approximations: they are provided to hopefully encourage others to develop further special function implementations. Tools include: Polynomial arithmetic and manipulation. Minimax approximations (the Remez algorithm). Helper functions used in testing etc. Documentation and source downloads are available online at www.johnmaddock.co.uk/toolkit Regards, John Maddock.

2 2