- Boost-users - lists.boost.org

Seekg on a gzip'd file...Is this possible?
by Gerrick Bivins 14 Mar '09

14 Mar '09

Hello all, I¹m switching one of my projects over from java to c++. Basically it¹s a gzip file reader, where I¹d like to parse a bunch of numbers (floats, ints, doubles) and in some instances I¹d like to skip some of the data. I¹m a ³noob² to boost and after getting the file reader working using the boost like this: { .... std::ifstream ormodelFile(filenameForTimeStep.c_str(), std::ios_base::in|std::ios_base::binary); boost::iostreams::filtering_istreambuf compressedFileFilter; compressedFileFilter.push(boost::iostreams::gzip_decompressor()); compressedFileFilter.push(ormodelFile); this->inputFile.rdbuf(&compressedFileFilter); //where inputFile is a an istream. ... //do the file reading } I tried to add the skipping parts to the code by simply calling: { ... double value; this->inputFile->seekg(sizeof(double),ios::cur); ... } the fail bit is set on the inputFile istream. Is it possible to do this using boost? I¹m pretty ³green² with boost so I¹m sure it¹s just my misunderstanding of how to use the lib but I couldn¹t find my way around the docs to well. The best I could find was that I need to somehow make my stream ³seekable² but it wasn¹t too clear to me how to do this. Any help would be greatly appreciated. Gerrick

5 8

Re: [Boost-users] shared_ptr, auto_ptr and enable_shared_from_this confusion
by 张亚霏 14 Mar '09

14 Mar '09

shared_ptr's constructor in line 168 calls 'boost::detail::sp_enable_shared_from_this( pn, p, p );'. 'boost::shared_ptr<IDerived> c(new Derived());' is dispatched to 'sp_enable_shared_from_this' in line 86. pay attension to the second parameter, it can accept all types that inherit 'enable_shared_from_this<T>', so does 'Derived*'. While, 'boost::shared_ptr<IDerived> c((IDerived*)new Derived());' is dispatched to 'sp_enable_shared_from_this' in line 111, bacause the static type IDerived* has nothing to do with 'enable_shared_from_this<T>' and can not be implicitly converted to 'enable_shared_from_this<T>'. Then check out the difference between the two 'sp_enable_shared_from_this'. It is obvious that the second usage shall cause a BOOST_ASSERT failed in 'enable_shared_from_this::shared_from_this', so an exception is raised. However, I think the root reason is not the language nor boost, but the design of class IDerived. Its root and root1 seem a little bit irrational because there is not direct relationship between IDerived and Base(IBase). ------------------ Original ------------------ From: "Ryan McConnehey"<mccorywork(a)gmail.com>; Date: Sat, Mar 14, 2009 04:29 AM To: "Boost-users"<boost-users(a)lists.boost.org>; Subject: [Boost-users] shared_ptr,auto_ptr and enable_shared_from_this confusion I'm having some trouble with using shared_ptr, auto_ptr and enable_shared_from_this with my inheritance layout. The following is how my inheritance is laid out. // Inheritance Layout // // boost::enable_shared_from_this<T> // ^ // | // Base --------> IBase // ^ // | // Derived -----> IDerived In my test code (that I've included with this email) the following two ways to instantiate a shared pointer is equivalent. boost::shared_ptr<IDerived> r (new Derived()); std::auto_ptr<IDerived> k (new Derived); boost::shared_ptr<IDerived> l (k.release()); When creating the shared_ptr, both declaration take me to line 185 of shared_ptr.hpp. Instantiating the shared_ptr directly then takes me to line 105 of shared_ptr.hpp. Though when instantiating by releasing the auto_ptr into the shared_ptr I'm taken to line 129. It seems to me that I should be going to the same line of code. This wouldn't be such a problem except when trying to call shared_from_this() an exception is thrown if my shared_ptr was created from an auto_ptr. Is this behavior correct? Ryan

1 0

(no subject)
by isik bicer 14 Mar '09

14 Mar '09

Hello everybody, I am a newly boost user and have a problem in the following code. Since it is long, i put only the relevant parts. Could anyone help me please? When I try to get the vertices from a vector and put them to add_edge(), i get the following error massage : 'boost::add_edge' : cannot convert parameter 1 from 'char' to 'void *' The problematic part is highlighted. Code:typedefvector<Edge> simpleedgevector; vector<Edge>::iterator s_e_i;for simpleedge = Edge((*(*vec_it1)),(*(*vec_it1+1))); simpleedgevector.push_back(simpleedge);}edge_descriptor e; cout<< tie(e, inserted) = add_edge((*s_e_i).first, (*s_e_i).second, g); }for(s_e_i = simpleedgevector.begin();s_e_i != simpleedgevector.end();s_e_i++){boolinserted;"first part : "<<(*s_e_i).first<<" second part : "<<(*s_e_i).second<<endl;(vec_it1 = output1.begin();vec_it1 !=output1.end();vec_it1++){std::pair<char, char> Edge;

1 0

shared_ptr, auto_ptr and enable_shared_from_this confusion
by Ryan McConnehey 13 Mar '09

13 Mar '09

I'm having some trouble with using shared_ptr, auto_ptr and enable_shared_from_this with my inheritance layout. The following is how my inheritance is laid out. // Inheritance Layout // // boost::enable_shared_from_this<T> // ^ // | // Base --------> IBase // ^ // | // Derived -----> IDerived In my test code (that I've included with this email) the following two ways to instantiate a shared pointer is equivalent. boost::shared_ptr<IDerived> r (new Derived()); std::auto_ptr<IDerived> k (new Derived); boost::shared_ptr<IDerived> l (k.release()); When creating the shared_ptr, both declaration take me to line 185 of shared_ptr.hpp. Instantiating the shared_ptr directly then takes me to line 105 of shared_ptr.hpp. Though when instantiating by releasing the auto_ptr into the shared_ptr I'm taken to line 129. It seems to me that I should be going to the same line of code. This wouldn't be such a problem except when trying to call shared_from_this() an exception is thrown if my shared_ptr was created from an auto_ptr. Is this behavior correct? Ryan #include <iostream> #include <memory> #include <boost/shared_ptr.hpp> #include <boost/enable_shared_from_this.hpp> // Inheritance Layout // // boost::enable_shared_from_this<T> // ^ // | // Base --------> IBase // ^ // | // Derived -----> IDerived class IBase { public: virtual void hello(void) = 0; }; class Base : public IBase, public boost::enable_shared_from_this<Base> { public: Base(void){} ~Base(void){} void hello(void) { std::cout << "Hello" << std::endl; } }; class IDerived { public: virtual void world(void) = 0; static std::auto_ptr<IDerived> createAuto(void); static boost::shared_ptr<IDerived> createShared(void); virtual boost::shared_ptr<Base> root(void) = 0; virtual boost::shared_ptr<IBase> root2(void) = 0; }; class Derived : public IDerived, public Base { public: Derived(void){} ~Derived(void){} void world(void) { std::cout << "World" << std::endl; } virtual boost::shared_ptr<Base> root(void) { return shared_from_this(); } virtual boost::shared_ptr<IBase> root2(void) { return shared_from_this(); } }; std::auto_ptr<IDerived> IDerived::createAuto(void) { return std::auto_ptr<IDerived>(new Derived()); } boost::shared_ptr<IDerived> IDerived::createShared(void) { return boost::shared_ptr<IDerived>(new Derived()); } int main(int argc, char* argv[]) { boost::shared_ptr<IDerived> c = IDerived::createShared(); boost::shared_ptr<IBase> d = c->root2(); std::cout << c.use_count() << std::endl; boost::shared_ptr<IDerived> r (new Derived()); std::auto_ptr<IDerived> k (new Derived); boost::shared_ptr<IDerived> l (k.release()); boost::shared_ptr<IBase> m = l->root2(); std::cout << m.use_count() << std::endl; return 0; }

2 1

using boost :: thread in Linux
by ayansinhas＠gmail.com 13 Mar '09

13 Mar '09

HI, I am using boost::thread for my project. I have extracted only boost::thread through BCP and I have build my multithreading application. It is working properly in my local machine (Windows)now I have to port it over to linux box. I had couple of questions ...?? Is it ok to copy the boost::thread folder (That I have extracted) from my local machine to linux box and run my application?? or is it that I have to install the entire boost library in linux box and then by using bcp I have to extract only boost:thread part which I can use for my application?? Please let me know what can be the possible solution.If there is any other solution please feel free to advice. Thanks In Advance Ayan.

2 1

[signals] Chaining signals
by Remko Tronçon 13 Mar '09

13 Mar '09

Hi, Is it possible to 'chain' boost signals? I.e., connect signal a to signal b, such that whenever signal a fires, all slots of signal b are called. thanks, Remko

2 2

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