- Boost-users - lists.boost.org

compiler warnings in
by Chris Cleeland 17 Jan '12

17 Jan '12

Using Boost 1_47_0, in checking some automated builds on a CentOS 4 machine, I get the following errors from the compiler there (gcc v3.4.6). ../../../boost/exception/detail/error_info_impl.hpp:43: warning: `class boost::error_info<tag_answer, int>' has virtual functions but non-virtual destructor Indeed, if I go back and compile the boost tests for exceptions, I get the same compiler warning. GCC 3.4.6 is listed as a tested compiler for 1.47.0; is this warning then expected? Or not? The change that silences the compiler is declaring the destructor of boost::error_info_base to be virtual, but I'm not entirely sure if that's correct. I'm also hesitant to make this change only for CentOS 4 as that will change the object layout in memory and could result in linker confusion, though one could argue that linking against the wrong library is an error unto itself. If there's a more appropriate place to report/discuss these sorts of issues, please point me towards them. Thanks! -- Chris Cleeland

2 1

Jamfile.v2 for iostreams in Boost 1_48
by David Smith 17 Jan '12

17 Jan '12

Greetings, The Jamfile.v2 in the iostreams library of Boost 1_48 has the line (around 29): $(v) = [ modules.peek : $(v) ] ; Used to propogate path-constant settings to the local jam environment. I think it should be: $($(v)) = [ modules.peek : $(v) ] ; At least when I run with this the build of iostreams works correctly to compile bzip2 and zlib objects and fails otherwise. Has anyone else observed this? Regards David

3 2

boost::mpl, c++11 and clang 3.0: compiler bug?
by Allan Nielsen 17 Jan '12

17 Jan '12

Hi The following code compiles fine under clang 3.0 but I get unexpected output: #include <iostream> #include <boost/mpl/for_each.hpp> #include <boost/mpl/vector.hpp> template< typename _state > struct AccSize{ size_t s = 0; _state xxxx; template<typename T> void operator()(T) { std::cout << s << " " << sizeof(T) << std::endl; s += sizeof(T); }; }; struct Foo { Foo () {} Foo (const Foo & rhs) {} }; struct A{}; int main(int argc, const char *argv[]) { using namespace std; AccSize < Foo > p; typedef boost::mpl::vector<int, int, int> list; boost::mpl::for_each< list >( p ); return 0; } The output I get is: 0 4 0 4 0 4 The output I expected is: 0 4 4 4 8 4 If I remove the copy constructor in Foo, it works as I expect, and if I do a "normal" initialization of AccSize::s it also works as I expect. Is this a bug? If so, how can it be reproduced without mpl::foreach Best regards Allan W. Nielsen

3 3

[interprocess] hanging when open_or_create a mapped file
by yiling1 gong 17 Jan '12

17 Jan '12

hello, everyone: I’m a newbie to boost::interprocess, It’s really a cool library, so I try to write a simple program to test mapped file, it’s really simple: #include <boost/interprocess/managed_mapped_file.hpp> using namespace boost::interprocess; int main() { managed_mapped_file mfile(open_or_create, "/home/fractal/log.bin", 1024); //hanging here. return 0; } but it’s hanging on the creation of mapped file, I test it both on ubuntu 11.05(in virtualbox) and openSuse, the same problem remain. I’m sure the file exists, and I traced down the source code, it seems the program stuck in a while loop in function priv_open_or_create(...). except that, if I set the third argument of constructor from 1024 to 1, it’s even throw a exception( not good). any seggestion is appreciated!!!

1 0

boost::interprocess::detail::managed_open_or_create_impl::priv_open_or_create function
by Raghu Gyambavantha 16 Jan '12

16 Jan '12

Boost-ers, When I try to allocate more elements from a managed shared memory segment than allowed, I don't seem to be getting an error from the boost library. It looks like the reason for it is because the call to "construct_func" from "priv_open_or_create" in "boost::interprocess::detail::managed_open_or_create_impl" returns true/false, but that return value is never checked. I saw this error with boost versrion 1.43.0, but that specific portion of the code doesn't seem to have changed even in 1.48.0. >From the looks of it, it seems like the construct_func should throw instead of returning, but in 1.43.0, it is returning a value. Should the return value be check and appropriate action be taken here or am I missing something. Any help is much appreciated. Thanks, -Raghu

1 0

[BGL] Serialize boost::labeled_graph
by Alper Aydemir 15 Jan '12

15 Jan '12

Hi, Is there anyway to easily serialize labeled_graphs ? I can see that <boost/graph/adj_list_serialize.hpp> does it for adjacency lists, however no such header seem to exist for labeled graph type. -- Alper Aydemir

2 1

Multi Threaded Server with boost asio
by mark pashley 15 Jan '12

15 Jan '12

I am looking at writing a multithreaded tcp server using boost ASIO. I have read through the tutorials and had a look at some of the examples and just want to check that my understanding is correct. The server will accept connections then service requests from multiple clients. My understanding is as follows: 1 The server uses "a single io_service and a thread pool calling io_service::run()" 2 All threads call io_service::run(). 3 The calls to io_service::run() are not within a strand, ergo completion handlers can run simultaneously. 4 When a request arrives one of the threads is chosen, it's read handler will be called 5 Another request may arrive,starting the read handler on a second thread 6 When one of the threads has finished handling the request it calls async_write, from within a strand 7 Another thread also finishes processing its request, it also calls async_write, from within a strand 8 The writes to the io_service are serialised via the strand, ergo they are thread safe. 9 When the write operation completes the thread calls async_read() 10 This call is not protected by a strand and the thread will be used for handling requests Is my understanding correct? -- View this message in context: http://boost.2283326.n4.nabble.com/Multi-Threaded-Server-with-boost-asio-tp… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

What's the execution sequence of phoenix::lambda function?
by 饭桶 15 Jan '12

15 Jan '12

I'm a newbie for boost phoenix, I wrote a small piece of code but it totally confuses me about the execution sequence, you can check the code ------------------clip----------------- std::vector<int> v; v.push_back(1); ph::for_each(v, ph::lambda[ph::ref(cout)<<"a", ph::ref(cout)<<"b" ])(v); ph::for_each(arg1, ph::lambda[ph::ref(cout)<<"a", ph::for_each(v, ph::lambda[ph::ref(cout)<<"b", ph::ref(cout)<<"c" ]), ph::ref(cout)<<"d" ])(v); ------------------------clap-------------------------- The first output is "ab" but the second output is "dbca". It seams ph::for_each is lazy evaluate , but what is the correct one to make it execute from left to right? thanks -- 半亩薄田一头牛老婆孩子热炕头

1 0

[asio] usb howto
by László Veréb 15 Jan '12

15 Jan '12

Hi All! I would like to use the asio library to communicate through usb with some devices. I looked the documentation, and searched the internet. It was said that it can be done, but I could not find any specific information. I found a code for (old) serial ports, it mentioned the usb, but nothing more. I think the important parts were: boost::asio::serial_port port; port.open(devname); port.set_option(***); where one can set these options boost::asio::serial_port_base::baud_rate, parity, character_size, flow_control, stop_bits Are these option also available for usb ports? For example the data flows (found by be) are not the same. Thanks beforehand for your reply and help. Laci

1 0

Boost interprocess vector, creating multiple vector pointers
by Prasanth Madhavan 15 Jan '12

15 Jan '12

I want to insert a shared vector into a shared map using the following code: managed_shared_memory segment(create_only ,"MySharedMemory" ,65536); typedef allocator<int, managed_shared_memory::segment_manager> vecAllocator;typedef vector<int, vecAllocator> vec;typedef std::pair<int, vec* > ValueType;typedef allocator<ValueType, managed_shared_memory::segment_manager> ShmemAllocator;typedef multimap<int, vec*, std::less<int>, ShmemAllocator> MyMap; ShmemAllocator alloc_inst (segment.get_segment_manager());vecAllocator vectorallocator (segment.get_segment_manager()); MyMap *mymap = segment.construct<MyMap>("MyMap")(std::less<int>(),alloc_inst);vec *myvec = segment.construct<vec>("myvec")(vectorallocator);vec *myvec1 = segment.construct<vec>("myvec1")(vectorallocator); myvec->push_back(10);myvec->push_back(9);myvec->push_back(8);myvec1->push_back(987);myvec1->push_back(123);myvec1->push_back(456); for(int i = 0; i < 6; ++i){ for(int j = 0; j<6; j++) mymap->insert(std::pair<int, vec*>(i, myvec));}for(int i = 0; i < 6; ++i){ for(int j = 0; j<6; j++) mymap->insert(std::pair<int, vec*>(i, myvec1));}the code works.. but what i want to do is construct a vector without a name. that is do something like this vec *myvec = segment.construct<vec>(vectorallocator);vec *myvec1 = segment.construct<vec>(vectorallocator);so that if i put it inside a loop and there is no problem with naming them differently every loop. Is there a way to do so? Or atleast is there an automatic way to name them differently every loop? -- Prasanth Madhavan

3 2