March 2009 - Boost-users - lists.preview.boost.org

Re: [Boost-users] [Multi-Index] crash on iterator copy or assignment
by JOAQUIN M. LOPEZ MUÑOZ 13 Mar '09

13 Mar '09

Hi Dominique ________________________________________ De: boost-users-bounces(a)lists.boost.org [boost-users-bounces(a)lists.boost.org] En nombre de Dominique Devienne [ddevienne(a)gmail.com] Enviado el: jueves, 12 de marzo de 2009 21:46 Para: boost-users(a)lists.boost.org Asunto: Re: [Boost-users] [Multi-Index] crash on iterator copy or assignment > On Thu, Mar 12, 2009 at 3:36 PM, Steven Watanabe <watanabesj(a)gmail.com> wrote: > > After filling in enough to make your code compile, > > it works for me. I've attached what I did. If this > > works for you, can you post a minimal and complete > > example that fails? > > Hi Steven, > > Your example works for me as well, > since you've found the same fix my colleague found, which is to write > > std::auto_ptr<ByTypeFooIterator> iter_ptr(new ByTypeFooIterator); > > instead of > > std::auto_ptr<ByTypeFooIterator> iter_ptr = new ByTypeFooIterator; I was trying to reproduce your problem locally but as usual Steven outdid me :) FWIW, the last line should not compile, and the fact that it does in VS2005 is a compiler bug: http://tinyurl.com/abl9cj Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

5 6

asio stream_handle::async_read_some && pipe::PIPE_READMODE_MESSAGE = access violation
by Andrew Solodovnikov 13 Mar '09

13 Mar '09

Hi! I'm trying to write simple pipe redirector. Pipe have to be readed in message mode (it's awesome, i know, but i can't change this). So i should read all message content into buffer and then in one write send it to the service host. I'm using stream_handle::async_read_some to read message. When buffer bigger, then message, all ok - message readed and my handler called. Otherwise access violation (Unhandled exception at 0xfeeefeee). Call stack (buffer size 16 bytes): * feeefeee() * boost::asio::detail::win_iocp_io_service::operation::do_completion(unsigned long last_error=0x000000ea, unsigned int bytes_transferred=0x00000010) Line 77 + 0x16 bytes C++ * boost::asio::detail::win_iocp_io_service::do_one(bool block=true, boost::system::error_code & ec={...}) Line 509 C++ * boost::asio::detail::win_iocp_io_service::run(boost::system::error_code & ec={...}) Line 186 + 0xe bytes C++ * boost::asio::io_service::run() Line 58 + 0xf bytes C++ As you can see, transferred byte count exactly 16, and error code EA (MORE_DATA_AVAIL). This is correct response for message mode pipes, so i should repeat read operation in this case until success. So, after getting this error asio tries to call handler, but overlapped structure data is invalid. Sometimes several callbacks get called, but at the end - access violation. It seems that something wrong with syncronization, but i have only one thread with io_service::run. Thanks for answers! -- Andrew Solodovnikov ==================== Sample code to reproduce problem: #include <boost/bind.hpp> #include <boost/asio.hpp> #include <boost/thread.hpp> #include <memory> struct PipeServer { PipeServer():m_hPipe(INVALID_HANDLE_VALUE), m_sName(NULL) { } BOOL Create(LPCTSTR szName, DWORD dwTimeout = 2000, DWORD dwBufSize = 4096) { m_bStop = FALSE; m_sName = _tcsdup(szName); if (m_sName) { m_dwTimeout = dwTimeout; m_dwBufSize = dwBufSize; InternalCreatePipe(); if (m_hPipe != INVALID_HANDLE_VALUE) return TRUE; free(m_sName); m_sName = NULL; } return FALSE; } BOOL Accept(HANDLE &theStream) { if (m_hPipe != INVALID_HANDLE_VALUE) { if (ConnectNamedPipe(m_hPipe, NULL) || (GetLastError() == ERROR_PIPE_CONNECTED)) { if (!m_bStop) { theStream = m_hPipe; InternalCreatePipe(); return TRUE; } VERIFY(::DisconnectNamedPipe(m_hPipe)); } VERIFY(::CloseHandle(m_hPipe)); m_hPipe = INVALID_HANDLE_VALUE; } return FALSE; } void Stop() { ::InterlockedExchange((long *)&m_bStop, TRUE); // possible hang on ::CallNamedPipe - if another pipe connected between this calls... DWORD dwTemp; ::CallNamedPipe(m_sName, NULL, 0, NULL, 0, &dwTemp, NMPWAIT_NOWAIT); } void Close() { if (m_hPipe != INVALID_HANDLE_VALUE) { ::DisconnectNamedPipe(m_hPipe); ::CloseHandle(m_hPipe); m_hPipe = INVALID_HANDLE_VALUE; } free(m_sName); m_sName = NULL; } private: void InternalCreatePipe() { m_hPipe = ::CreateNamedPipe( m_sName, // pipe name PIPE_ACCESS_DUPLEX | FILE_FLAG_OVERLAPPED, // read/write access PIPE_TYPE_MESSAGE | // message type pipe PIPE_READMODE_MESSAGE | // message-read mode PIPE_WAIT, // blocking mode PIPE_UNLIMITED_INSTANCES, // max. instances m_dwBufSize, // output buffer size m_dwBufSize, // input buffer size m_dwTimeout, // client time-out NULL // no security attribute ); } private: BOOL m_bStop; LPTSTR m_sName; HANDLE m_hPipe; DWORD m_dwTimeout; DWORD m_dwBufSize; }; #define BLOCK_SIZE 64 struct pipe_session { pipe_session(boost::asio::io_service& io_service, HANDLE hClientHandle): m_hClientHandle(io_service, hClientHandle) { io_service.post(boost::bind(&pipe_session::client_read, this)); } private: void client_read() { //boost::asio::async_read(m_hClientHandle, m_hClientHandle.async_read_some( boost::asio::buffer(&szBuffer[0], BLOCK_SIZE), boost::bind(&pipe_session::handle_client_read, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred )); } void handle_client_read(const boost::system::error_code& error, size_t nReaded) { if (!error) { // end iteration int i = 0; } else { if (error.value() == ERROR_MORE_DATA) { //m_sbClientBuffer.commit(BLOCK_SIZE); client_read(); } } } private: boost::asio::windows::stream_handle m_hClientHandle; char szBuffer[4096]; }; void server_thread(PipeServer *server) { boost::asio::io_service io_service; boost::asio::add_service(io_service, new boost::asio::windows::stream_handle_service(io_service)); std::auto_ptr<boost::asio::io_service::work> work(new boost::asio::io_service::work(io_service)); boost::thread t(boost::bind(&boost::asio::io_service::run, &io_service)); HANDLE hStream; while (server->Accept(hStream)) { new pipe_session(io_service, hStream); } work.reset(); // Allow run() to exit. t.join(); } int _tmain(int argc, TCHAR* argv[], TCHAR* envp[]) { PipeServer server; if (server.Create(_T("\\\\.\\pipe\\test_channel"))) { try { { boost::thread t(boost::bind(&server_thread, &server)); getch(); HANDLE hService = CreateFile(_T("\\\\.\\pipe\\test_channel"), GENERIC_READ | GENERIC_WRITE, 0, NULL, OPEN_EXISTING, 0, NULL); if (hService != INVALID_HANDLE_VALUE) { DWORD mode = PIPE_READMODE_MESSAGE; ::SetNamedPipeHandleState(hService, &mode, NULL, NULL); std::vector<char> data(0xa5); DWORD dw; ::WriteFile(hService, &data.front(), data.size(), &dw, NULL); } getch(); server.Stop(); t.join(); } } catch (std::exception& e) { std::cerr << "Exception: " << e.what() << "\n"; } server.Close(); } return 0; }

1 0

boost::fusion zip_iterator and generic make_tuple call
by David Genest 13 Mar '09

13 Mar '09

Hi, I'm cross-posting here after a kind invitation from Joel. I hope I can tap into your expert wisom! :-) I have a fusion sequence of std::vectors which can be specified variably by users: typedef boost::fusion::vector<std::vector, std::vector, ...> Inputs; Using boost::zip_iterator, I would like to parallel iterate each std::vector of the sequence and call a user function, but I am not sure how to proceed. Of course I would like to do this in a generic manner so that the user can change the number of inputs (std::vectors) at will. My first non generic shot looks like this: typedef boost::fusion::vector<std::vector<int>, std::vector<int> > Inputs; Inputs my_input; struct UserFunction { static void apply(int i, int j) { } } template<class F> struct zip_func { void operator() (const boost::tuple<int, int>& t) const { F::apply(t.get<0>(), t.get<1>()); } }; int main (int argc, char* argv[]) { using namespace boost::fusion; std::foreach( boost::make_zip_iterator( boost::make_tuple( at_c<0>(my_input).begin(), at_c<1>(my_input).begin()), boost::make_zip_iterator( boost::make_tuple( at_c<0>(my_input).end(), at_c<1>(my_input).end()), zip_func<UserFunction>()) ); } To the experts on the list: I don't see how to generalize the zip_func operator() and make_tuple call based on the fusion::vector signature. Is it even feasible, or is my intuition leading me in a dead end? I saw that there is a zip_iterator for fusion sequence in a svn branch (http://svn.boost.org/svn/boost/branches/zip_iterator_fusion/), should I go more in that direction? Any pointers would be appreciated. Thank you for your time, David.

1 0

need some metaprogramming ideas
by Noah Roberts 13 Mar '09

13 Mar '09

I want to develop a way of checking if I can implement a trivial string converter through lexical_cast. If that converter doesn't exist then the developer is required to make one and a different converter is made that throws an exception if called (may not be the best way but for now that's what I'm working with). To be lexical castable T must be OutputStreamable, InputStreamable, CopyConstructable, and DefaultConstructable. I think I can implement checks for streamable through the sizeof trick...haven't thought of the exact way yet but I think it should work. I can't think of how to test for CopyConstructable or DefaultConstructable though. Any ideas?

2 1

[Multi-Index] crash on iterator copy or assignment
by Dominique Devienne 12 Mar '09

12 Mar '09

Hi, I've been using B.MI for a few days, and until now it's worked great for me. I've included below in Exhibit A a simplified (and a bit obfuscated, sorry) version of what I've been using (a proxy, indexed by name and type, and a method returning all instances of a given type). Filling in a list of all entries matching a given type (in exhibit A) works great. But now I was trying to return a custom (non-standard) iterator, where the client gets a pointer to a abstract base iterator (allows to hide the B.MI stuff in the concrete derived iterator), wrapped in an auto_ptr. The private concrete iterator internally stores the B.MI iterator from the index I'm interested in (exhibit B). This compiles (vs2005 sp1 on winXP 32), but it crashes at runtime on the call to boost::tie, during iterator assignment apparently. (call stack in exhibit C). So at this point I have no clue what's going on, and I would appreciate some help. The level of template complexity is more than stretching my abilities I'm afraid, just reading the type names makes my end spin ;) We're going to try it using GCC on Linux to see if it might be a compiler issue, and possibly try vs2008 as well, but if anyone knows if taking copies of the iterators is allowed (I know the index themselves shouldn't be copied, and we don't), and if yes why it's failing in this particular case, I'd appreciate. Thanks, --DD PS: Earlier I had a version of ByTypeFooIterator that took the begin/end iterator as param to its Ctor, taking a *copy*, but that also crashed w/o a stack trace this time. PPS: What I was trying to test in fact was whether I could get two equal ranges for two different MyTypeInfo, and iterate on those but interleaving the calls to ++iter, to verify nested or interleaved iteration was valid with B.MI. But I never reached that point. === Exhibit A: the code that works === struct FooProxy { Foo*const foo_; FooProxy(Foo* foo) : foo_(foo) {} const MyString& get_name() const { return foo_->name(); } const MyTypeInfo& get_type() const { return foo_->type_info(); } }; struct ByName {}; struct ByType {}; typedef boost::multi_index_container< FooProxy , indexed_by< hashed_non_unique< tag<ByName>, BOOST_MULTI_INDEX_CONST_MEM_FUN( FooProxy , const MyString&, get_name ) >, hashed_non_unique< tag<ByType>, BOOST_MULTI_INDEX_CONST_MEM_FUN( FooProxy , const MyTypeInfo&, get_type ) > > > FooIndex; FooIndex foo_index_; typedef FooIndex::index<ByName>::type FoosByName; typedef FooIndex::index<ByType>::type FoosByType; size_t select_foos(const MyTypeInfo& tinfo, MyList<Foo*>& list) { size_t count = 0; FoosByType::iterator range_begin, range_end; FoosByType& by_type = get<ByType>(foo_index_); boost::tie(range_begin, range_end) = by_type.equal_range(tinfo); for (FoosByType::iterator i = range_begin; i != range_end; ++i) { const FooProxy& proxy = *i; list.append(proxy.foo_); ++count; } return count; } === Exhibit B: the code that crashes === // public abstract base iterator class FooIterator { public: virtual ~FooIterator() {} virtual bool more() const = 0; virtual Foo* cur() const = 0; virtual void next() = 0; protected: FooIterator() {} }; // private concrete iterator struct ByTypeFooIterator : public FooIterator { FoosByType::iterator cur_, end_; virtual bool more() const { return cur_ != end_; } virtual Foo* cur() const { return cur_->foo_; } virtual void next() { ++cur_; } }; // public method std::auto_ptr<FooIterator> iterate(const MyTypeInfo& tinfo) { FoosByType& by_type = get<ByType>(foo_index_); std::auto_ptr<ByTypeFooIterator > iter_ptr = new ByTypeFooIterator; boost::tie(iter_ptr->cur_, iter_ptr->end_) = by_type.equal_range(tinfo); return iter_ptr; } // part of unit test Foo*const expected_bar = ...; // Bar derives from Foo for (std::auto_ptr<FooIterator> i = iterate(Bar::tinfo()); i->more(); i->next()) { Foo*const curr = i->cur(); CPPUNIT_ASSERT_EQUAL(expected_bar, curr); // only one Bar here } === Exhibit C: call stack of crash === (actual FooIndex contains more indices than shown in exhibit A above). > my_lib.dll!std::_Iterator_base::operator=(const std::_Iterator_base & _Right={...}) Line 374 + 0x8 bytes C++ my_lib.dll!std::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &>::operator=(const std::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::detail::iterator_base<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &>::operator=(const boost::detail::iterator_base<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &>::operator=(const boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> >::operator=(const boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::incrementable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > >::operator=(const boost::incrementable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > > & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::equality_comparable1<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::incrementable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > > >::operator=(const boost::equality_comparable1<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::incrementable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::dereferenceable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<& __that={...}) + 0x18 bytes C++ my_lib.dll!boost::input_iteratable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> >::operator=(const boost::input_iteratable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::forward_iteratable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> >::operator=(const boost::forward_iteratable<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy const *,boost::iterator<std::forward_iterator_tag,FooProxy,int,FooProxy const *,FooProxy const &> > & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::forward_iterator_helper<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy,int,FooProxy const *,FooProxy const &>::operator=(const boost::forward_iterator_helper<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,FooProxy,int,FooProxy const *,FooProxy const &> & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >::operator=(const boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > > & __that={...}) + 0x18 bytes C++ my_lib.dll!boost::tuples::tuple<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > > &,boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > > &,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>::operator=<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > > >(const std::pair<boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > >,boost::multi_index::detail::hashed_index_iterator<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::hashed_index_node<boost::multi_index::detail::index_node_base<FooProxy,std::allocator<FooProxy> > > >,boost::multi_index::detail::bucket_array<std::allocator<FooProxy> > > > & k=({node=0x04680080 buckets=0x04667bac },{node=0x04667c68 buckets=0x04667bac })) Line 638 C++ my_lib.dll!iterate(const MyTypeInfo & tinfo={...}) Line 510 + 0x51 bytes C++

2 3

Build Errors with Exceptions Support Disabled
by Gabi Davar 12 Mar '09

12 Mar '09

Hi All, While compiling the signals library (Boost v1.38) to a platform with exception support disabled (ARM RVDS v3.0Sp1 [EDG]) I've encountered build errors in the function libraries due to usage of try/catch keywords. See attached file for a fix where the keywords were replaced by BOOST_TRY/CATCH macros. -gabi

1 0

tokenizer problem
by Medora Schauer 12 Mar '09

12 Mar '09

Hi All! I'm having a problem passing strings into a boost tokenizer. I have been able to get the code to work but I don't think I should have had to do what I did. I don't know if it is tokenizer problem or a compiler problem or my problem. Basically, I start with a char[] and depending on how I convert it to an std::string, I may or may not get garbage tokens. I've attached a complete, but small, program to demonstrate the problem. I hope someone can tell what is going on. I'm using a Fedora 8 system with g++ v 4.1.2. Thanks in advance! Medora #include <iostream> #include <string> #include <boost/tokenizer.hpp> using namespace std; char *in_strs[] = { "S,1001,101,39.6511915386591,-108.199469861370,1240151.6,127361.5", "S,1001,111,39.6511191997175,-108.199521016802,1240136.4,127335.6", "S,1001,121,39.6510468691093,-108.199571817158,1240121.3,127309.7", "S,1001,131,39.6509745384787,-108.199622617407,1240106.2,127283.8", "S,1001,141,39.6509021994691,-108.199673772516,1240091.0,127257.9", "S,1001,151,39.6508298687935,-108.19972457255,1240075.9,127232.0", "S,1001,161,39.6507575380953,-108.199775372477,1240060.8,127206.1", "S,1001,171,39.6506851990178,-108.199826527263,1240045.6,127180.2", "S,1001,181,39.6506125938775,-108.199877316164,1240030.5,127154.2", NULL }; int main(int argc, char* argv[]) { typedef boost::tokenizer<boost::char_separator<char> > str_tokenizer; boost::char_separator<char> sep(","); cout << endl; #if 1 // Here I'm just tokenizing the strings and writing them out. // This version works. char **str = in_strs; while (*str){ string file1_str = *str; str_tokenizer file1_tokens(file1_str, sep); for(str_tokenizer::iterator iter=file1_tokens.begin(); iter!=file1_tokens.end();++iter){ cout << *iter << " "; } ++str; cout << endl; } #endif #if 0 // This version results in garbage in some tokens. char **str = in_strs; while (*str){ str_tokenizer file1_tokens(string(*str), sep); // <-- difference for(str_tokenizer::iterator iter=file1_tokens.begin(); iter!=file1_tokens.end();++iter){ cout << *iter << " "; } ++str; cout << endl; } #endif //==================================================================================================== #if 0 // Now I want to skip the first 2 characters of the string. // This version works. string file1_str; char **str = in_strs; while (*str){ string file1_str = *str; string sub1 = file1_str.substr(2); // <-- difference str_tokenizer file1_tokens(sub1, sep); // <-- difference for(str_tokenizer::iterator iter=file1_tokens.begin(); iter!=file1_tokens.end();++iter){ cout << *iter << " "; } ++str; cout << endl; } #endif #if 0 // This version results in garbage in some tokens. string file1_str; char **str = in_strs; while (*str){ string file1_str = *str; str_tokenizer file1_tokens(file1_str.substr(2), sep); // <--- difference for(str_tokenizer::iterator iter=file1_tokens.begin(); iter!=file1_tokens.end();++iter){ cout << *iter << " "; } ++str; cout << endl; } #endif return(0); }

2 2

A problem while building 1.38.0 for msvc
by Younie, Bradford 12 Mar '09

12 Mar '09

I downloaded the 1.38 build zip file and unzipped it. I then downloaded the latest bjam and unzipped that. I then used bjam to build boost, as follows: bjam --build-dir=c:\users\myuser\boost-build --toolset=msvc stage It all built successfully, but when I referenced libboost-thread-vc90-mt-gd-1_38_0.lib, I got a linker error where it is looking for libboost-regex-vc80-mt-gd-1_38_0.lib. The problem is that it is requiring the vc80 lib, but bjam only built vc90 libraries. When I renamed the 90 regex library to say 80, I got a host of unresolved externals when linking. (I didn't really expect that to work, but I tried anyway). Can someone advise what I should do to fix this problem? Here are the stats about my system: Windows Vista Business 32-bit Visual C++ 6.0 installed Visual Studio 2003 installed Visual Studio 2005 installed Visual Studio 2008 installed (we develop an SDK, so I need to build in the various environments.) Brad ---------- Learn more about Chase Paymentech Solutions,LLC payment processing services at www.chasepaymentech.com. THIS MESSAGE IS CONFIDENTIAL. This e-mail message and any attachments are proprietary and confidential information intended only for the use of the recipient(s) named above. If you are not the intended recipient, you may not print, distribute, or copy this message or any attachments. If you have received this communication in error, please notify the sender by return e-mail and delete this message and any attachments from your computer.

2 1

boost asio: Probelm with closing a blocking ip::udp::socket receive
by Lennyk 12 Mar '09

12 Mar '09

Hi, I'm using boost 1.35 on Linux. I have a thread which initiates a ip::udp::socket and has a blocking receive. When I try to terminate this thread from a different thread, in order to close the udp::socket I use the following: // m_socket is of type ip::udp::socket boost::system::error_code ec; m_socket::shutdown(boost::asio::ip::udp::socket::shutdown_receive, ec); std::cout << "error code = " << ec << std::endl; m_socket.close(); After shutdown I ALWAYS get system:107 error code - which means "Transport endpoint is not connected". Does anybody have any idea why I'm receiving such an error? Needless to say, the receive is working great - and I get the packets from the port I'm listening to. Thanks!

2 3

[asio] 1.37.0: trouble with multiple timers and serial line port
by Markus Werle 12 Mar '09

12 Mar '09

Hi! Using boost-1.37.0 Having some trouble with multiple timers in one class. Looks like the timers are going berserk when MemFun1() calls Process(): The expiration wait times then are ignored and the whole loop seesm to run as if timers are all set to 0. Any idea what went wrong? A little code sketch (does not compile) class Communicator { // the main IO service that runs this connection boost::asio::io_service io_service_; // the serial port this instance is connected to boost::asio::serial_port SerialPort_; // maximum amount of data to read in one operation static const int max_read_length = 256; // data read from the socket char read_msg_[max_read_length]; std::deque<char> received_data_; // buffered write data std::deque<char> write_msgs_; boost::thread * pThreadRead_; // etc. // timers boost::asio::deadline_timer timer1_; boost::asio::deadline_timer timer2_; // etc. // missing: // the whole async_write/async_read_some to serial line stuff void Process() { // MISSING CODE: // ... fill the send queue of the serial line cache and fire // async_write on serial port ... // Q: does this probaly interfere StartMemfun(0); } void StartMemFun1(int counter) { timer1_.cancel(); // Q: Is this necessary here? timer1_.expires_from_now(boost::posix_time::milliseconds(1000)); timer1_.async_wait(bind(MemFun1, this, counter)); } void MemFun1(int counter) { if (!(counter < 5)) Process(); bool success; // ... process data from read cache if (success) StartMemfun2(0); else StartMemFun1(counter + 1); // this here works fine! } void StartMemFun2(int counter) { timer2_.cancel(); // Q: Is this necessary here? timer2_.expires_from_now(boost::posix_time::milliseconds(1000)); timer2_.async_wait(bind(MemFun2, this, counter)); } void MemFun2(int counter) { if (!(counter < 5)) Process(); bool success; // ... process data from read cache if (success) StartMemfun2(counter + 1); else StartMemFun1(counter + 1); } };

2 2