- Boost-users - lists.boost.org

Data type to handle C++ and Boost streams?
by Adam Nielsen 13 Sep '09

13 Sep '09

Hi all, Bit of a basic question - I'd like to have some functions that work with files (using std::fstream) but can also handle boost::iostreams. Is there a data type I can use that can handle both? For example: void writeSomeData(std::fstream out); I can't pass a boost::iostream::stream to this function. What type can I use to achieve this? Many thanks, Adam.

3 3

[boost.thread] boost thread on x64
by Pascal Kesseli 13 Sep '09

13 Sep '09

Hi friends Using the freshly created MinGW 64bit build, creating the boost 1.40 libraries with bjam worked flawlessly. However, when including boost/thread.hpp, I receive the following error: In file included from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/basic_timed_mutex.hpp:14, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/mutex.hpp:8, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/mutex.hpp:14, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/detail/thread.hpp:11, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/thread.hpp:22, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread.hpp:13, from ../src/main/com/emc/schedule/DailyRepetitiveTaskComposite.cpp:2: C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/interlocked_read.hpp: In function 'void boost::detail::interlocked_write_release(void* volatile*, void*)': C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/interlocked_read.hpp:71: error: cast from 'void*' to 'long int' loses precision In file included from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/mutex.hpp:8, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/mutex.hpp:14, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/detail/thread.hpp:11, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread/thread.hpp:22, from C:\Program Files\C++\boost\boost_1_40_0/boost/thread.hpp:13, from ../src/main/com/emc/schedule/DailyRepetitiveTaskComposite.cpp:2: C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/basic_timed_mutex.hpp: In member function 'void* boost::detail::basic_timed_mutex::get_event()': C:\Program Files\C++\boost\boost_1_40_0/boost/thread/win32/basic_timed_mutex.hpp:152: error: cast from 'void*' to 'long int' loses precision make: *** [src/main/com/emc/schedule/DailyRepetitiveTaskComposite.o] Error 1 However, on the stated line, I can't even make out a cast. Is boost_thread win64 compatible? g++ version: g++.exe (GCC) 4.4.1 20090505 (prerelease) Copyright (C) 2009 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Thanks for any help and best regards Pascal

2 1

What is the way to access the data contained in a stream buffer?
by Ramon F Herrera 13 Sep '09

13 Sep '09

This is the relevant code: boost::asio::streambuf request; std::ostream request_stream(&request); request_stream << "GET " << argv[2] << " HTTP/1.0\r\n"; request_stream << "Host: " << argv[1] << "\r\n"; request_stream << "Accept: */*\r\n"; request_stream << "Connection: close\r\n\r\n"; // Send the request. boost::asio::write(socket, request); How can I convert the 'request.data' into a string or something? TIA, -Ramon

3 3

Re: [Boost-users] What is the way to access the data contained in a stream buffer?
by Anders Dalvander 13 Sep '09

13 Sep '09

Ramon F Herrera wrote: > I have a follow-up question of a more general nature. I think the Boost > libraries are the greatest thing since sliced bread, but there are times > when a simpler approach will suffice. We can see in these 2 lines of > code that we are mixing STL with Boost: > > boost::asio::streambuf request; > std::ostream request_stream(&request); > > Is there any STL class that can be used above, instead of > boost::asio::streambuf? > Something like std::ostringstream? std::ostringstream request_stream; request_stream << "GET " << argv[2] << " HTTP/1.0\r\n"; request_stream << "Host: " << argv[1] << "\r\n"; request_stream << "Accept: */*\r\n"; request_stream << "Connection: close\r\n\r\n"; std::string request = request_stream.str(); Regards, Anders Dalvander

1 0

Is there anything wrong with the Gmane newsgroup? + Interested in parsing tools and networking code
by Ramon F Herrera 13 Sep '09

13 Sep '09

Hello, I just joined the mailing list - Greetings to all. I have been trying for several day to post messages to the Boost Users newsgroup on Gmane. I didn't get any error message, but my postings never appeared. Boost is one of the best resources I have found in a very long time - Thanks for the effort! One of my main current interests is parsing. Trying to decide among the choices: - Regex - Spirit - Xpressive I could use some help with ASIO, too. -Ramon

4 10

PBGL / Dynamic Property Maps complexity
by Mathieu Malaterre 12 Sep '09

12 Sep '09

Hi there, I am slowly discovering (Parallel) BGL. Could someone confirm the following: All the BGL examples I found handle dynamic properties (data attached to a vertex) using 'Dynamic Property Maps'. If I understand correctly dynamic_properties are basically a std::map. The complexity to get a value from a std::map is O(log(n)). Thus a graph traversal of all values is O(n * log(n)). If this is correct, my question is: is there another approach where I could have a O(1) complexity to access data associated with an edge ? Thanks, -- Mathieu

2 2

[program_options]If a parameter in the command line contains white space, how to parse it ?
by yaoyansi 12 Sep '09

12 Sep '09

hi, all i use boost::program_options to parse cmd line parameters. Attachment has my code. And here is my test: tst.exe --wav c:/Program Files/Microsoft/a.wav --words "Hi, how are you?" --d "so,\ may\ be." the output is : wav:[c:/Program ] words:["Hi, ] d:["so,\ ] -------------------------------------------------- But, I want the result to be : wav:[c:/Program Files/Microsoft/a.wav ] words:[Hi, how are you?] d:[so, may be. ] What should I do now? Sorry for my poor English, And ,thanks in advance. ___________________________________________________________ 好玩贺卡等你发，邮箱贺卡全新上线！ http://card.mail.cn.yahoo.com/

3 3

[interprocess] 1.40 compile error
by Moritz 12 Sep '09

12 Sep '09

Hi, I just started to port my application from boost 1.37 to 1.40 in order to get the remove() methods work in an appropriate way. Problem is now that the compiler is complaining about my use of named_condition.wait(lock). Here is what I get: ../boost_1_40_0/boost/interprocess/sync/named_condition.hpp|156|error: invalid application of 'sizeof' to incomplete type 'boost::STATIC_ASSERTION_FAILURE<false>' ../boost_1_40_0/boost/interprocess/sync/named_condition.hpp|254|instantiated from 'void boost::interprocess::named_condition::wait(L&) [with L = boost::interprocess::scoped_lock<boost::interprocess::named_upgradable_mutex>]'| Thank you in advance for any hint.

2 3

[streams] surplus output generated
by Sven Heithecker 12 Sep '09

12 Sep '09

Hi all, the code below using boost::iostream produces unwanted surplus output: Write: this ist a test output split into two parts Print: this flushes previous output Write: another test output t split into two parts The last line is a fragment of the first line which should not be there. I am using boost 1_39, gcc 4.3.2 [gcc-4_3-branch revision 141291] What am I doing wrong here ? BTW: the code is a stripped demo code. Originally, "true" devices are used instead of the dummy "0", and I use some home-brewn classes (instead of just "string") which require specialized output handling depending on the used device, therefore the direct call into the stream object. However, this makes no difference. Regards, Sven --------------------------------------------------------------------- #include <iostream> #include <boost/iostreams/stream.hpp> #include <boost/iostreams/concepts.hpp> using namespace std; namespace io = boost::iostreams; // "demo" strem sink class CTestSink: public io::sink { public: CTestSink(int dummy) {} // iostreams requires constructor argument public: streamsize write(const char *s, streamsize n) { cout << "Write: " << s; return(n); } void Print(const string &str) const { cout << "Print: " << str << endl; } }; // the real stream typedef io::stream<CTestSink> CTestOStream; // specialized output for "string" objects CTestOStream &operator<<(CTestOStream &stream, string &str) { // make sure previous output is flushed // (not required in this example) stream.flush(); stream -> Print(str); // specialized output handling return(stream); } int main() { // dummy "0" required by boost:iostreams CTestOStream testostream(0); testostream << "this ist a test output "; testostream << "split into two parts" << endl; string str("this flushes previous output"); testostream << str; testostream << "another test output " << endl; } -- Sven Heithecker GPG Fingerprint: 52F7 06C2 BF51 96CE DFF7 CC89 1A6B DFD5 BEC2 02A6

2 2

[thread] multiple producers, single consumer - bosst-specific and generic questions
by Julien Claassen 11 Sep '09

11 Sep '09

Hello everyone! Thanks for all the help I already got. And excuse my poor understanding of of multithreading. I would like to implement a "multiple producer single consumer" pattern. My current solution is working with signals, which, I think is not very efficient. And also it doesn't work properly with the code I have, (see end of this mail). But besides that problem I have a generic question. If I worked with buffers and mutex, there are two choices, it seems: 1. Use one buffer, at the cost of deletion from the middle, when one producer terminates (so consumer won't communicate with a dead thread). 2. Have a buffer per producer, but also have a synchronisation per buffer and the consumer must look in several buffers, plus I don't know, how many producers there will be. This all should lead to a user interface, with different input-methods. So what would you suggest? It must be speedy and very CPU friendly. Thanks in advance for consideration. I'd also be happy for some suggested reading. Only one thing: printed books or documents in non-plain-text form besides .doc, .pdf and .ps aren't of any use to me, for I'm blind and using Linux text based environment. Kindest regards (For code, see below) Julien *** CODE SNIPPET *** boost::mutex my_mtx; // the Mutex for passing on events/chars // assumptions: // the following variables are passed to the Input object as a reference // int& its_q_size; // condition& its_in_condition; // condition& its_out_condition; // boost::signals2::signal<void (char)>& its_sig; void Input::run() { char input; boost::unique_lock<boost::mutex> lk(my_mtx); while (input != 'q') { cin >> input; while (its_q_size != 0) { its_in_condition.wait(lk); } its_sig(input); its_q_size = 0; its_out_condition.notify_all(); } } void Controller::run() { sig_t sig; // passed as its_sig condition in_condition, out_condition; // passed respectively as // its_incondition and its_out_condition int q_size = 0; // passed as char input; boost::unique_lock<boost::mutex> lk(my_mtx); Input *ip = new Input(in_condition,out_condition,sig,q_size); ip->connect(bind(&Controller::slot1, this, _1)); ip->connect(boost::bind(&Controller::slot2, this, _1)); thread runner(boost::bind(&Input::run, ip)); // why no third param _1 // here? With one it won't compiler! while (input != 'q') { while (q_size != 1) { out_condition.wait(lk); } // Do something with data received from the slots q_size = 0; in_condition.notify_all(); } runner.join(); } *** END OF CODE *** Effects: Read once, send signal, read again print info from first read and send signal again. Only really sync, when q is pressed. I expect due to runner-thread ending. -------- Music was my first love and it will be my last (John Miles) ======== FIND MY WEB-PROJECT AT: ======== http://ltsb.sourceforge.net the Linux TextBased Studio guide ======= AND MY PERSONAL PAGES AT: ======= http://www.juliencoder.de

4 8