August 2012 - Boost-users - lists.boost.org

Using boost::Asio for connected UDP socket
by Gheorghe Marinca 28 Aug '12

28 Aug '12

Hi, I want to implement an UDP server and have problems on defining on how the server handles coming concurrent udp "connections" from multiple clients. What I basically want to implement is an DNS proxy. Is there something equivalent to an tcp::acceptor for udp ? Looking over asio docs I see: template< typename Protocol <http://www.boost.org/doc/libs/1_35_0/doc/html/boost_asio/reference/Protocol…>, typename SocketAcceptorService <http://www.boost.org/doc/libs/1_35_0/doc/html/boost_asio/reference/SocketAc…> = socket_acceptor_service<Protocol>>class basic_socket_acceptor : public basic_io_object< SocketAcceptorService >, public socket_base For tcp we have following typedef: typedef basic_socket_acceptor< tcp > acceptor; Can I use basic_socket_acceptor with socket_acceptor_service<UDP> so that I can create additional udp sockets for every udp client that wants to send data to my server ? I've observed that if I use async_receive_from on the same socket I receive data from multiple clients making impossible exchanging data (read/write) with multiple clients at the same time. Any hints are appreciated. -Ghita

3 2

Re: [Boost-users] boost.Iostream OpenSSL Filter
by Neil Nelson 28 Aug '12

28 Aug '12

exocoder, Upon just reading the docs at http://www.boost.org/doc/libs/1_50_0/libs/iostreams/doc/index.html and particularly 2.2.8. Multi-Character Filters It looks like you could take the code from the following page http://stackoverflow.com/questions/1007337/encrypting-and-decrypting-a-smal… with the critical line being AES_cfb128_encrypt(indata, outdata, bytes_read, &key, ivec, &num, AES_ENCRYPT); that reads and writes to memory buffers. Two other aspects are: (1) software using encryption is export restricted and you can research that if needed for your application, (2) key management is likely the more complicated part, and (3) use of initialization vectors (added random bytes at the beginning of the input data). Neil Nelson > I am currently looking at using OpenSSL to write information to an > encrypted file on-the-fly so to speak. I have been researching my > options for doing this, and have come across the Boost.Iostream library, > which promises to be a very elegant solution (The documentation even > suggests that filters can be written to provide encryption/decryption > using OpenSSL). > It would seem that my best approach would be to use the BIO_f* interface > to encrypt a memory buffer, which could be written by overloading > write(). However the OpenSSL documentation is particularly sparse when > it comes to file/memory encryption and I would appreciate any advice on > how to do this. > Thank you in advance for any help proffered.

3 4

[Filesystem] path::generic_u16string() seems not to exist?
by Craig Dickson 28 Aug '12

28 Aug '12

Hello all, The online reference for Boost 1.50 show methods called u16string() and generic_u16string() to obtain the native and generic forms of the path as std::u16strings. However, if I try to compile code that uses them, the compiler can't find them, and the text "u16string" does not occur in boost/filesystem.hpp. (The same remarks apply to u32string and generic_u32string.) My workaround for now is to provide my own implementation of string<u16string>() and generic_string<u16string>(). This gets the job done, but the discrepancy between the docs and the code still bugs me. Will these methods be provided in a future release? If not, shouldn't they be removed from the docs? Thanks, Craig

1 0

[1.51] valgrind problem in boost::math::erfc_inv
by Jan Boehme 28 Aug '12

28 Aug '12

Hello, after switching to 1.51, I observe that none of the created apps or tests can be profiled using valgrind. When using 1.50 there's no problem in it. The issue that valgrind complains is the following: ==26716== Memcheck, a memory error detector ==26716== Copyright (C) 2002-2010, and GNU GPL'd, by Julian Seward et al. ==26716== Using Valgrind-3.6.1 and LibVEX; rerun with -h for copyright info ==26716== Command: optiSLang/optiSLang_4_0/kernel/tests/bin/serialize.test/782744749150d7ee13cdc314b59c8e95/serialize ==26716== terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::overflow_error> >' what(): Error in function boost::math::erfc_inv<e>(e, e): Overflow Error The system is a Fedora 16 powered by kernel 3.4.9-1.fc16.x86_64 Thanks for giving an advice, Jan.

2 2

Using boost::Asio for connected UDP socket
by Eugene Mamin 28 Aug '12

28 Aug '12

>>I want to implement an UDP server and have problems on defining on how the server handles coming concurrent udp "connections" from multiple clients. What I basically want to implement is an DNS proxy. You should see following implementation of transport layer from maidsafe-dht project: https://github.com/maidsafe/MaidSafe-Transport/tree/master/src/maidsafe/tra… Also, this code reviewed by boost.asio author Christopher M. Kohlhoff.

1 0

boost::graph, ref vs. copy during relaxation
by Tristram Gräbener 27 Aug '12

27 Aug '12

Hello, I use boost::graph and the Dijkstra algorithm. My edge property is rather big (I store public transport timetables on them). In the file boost/graph/relax.hpp, the property of an edge is copied : W w_e = get(w, e); As my properties are large, it yields some overhead. I changed that line to const W & w_e = get(w, e); My benchmarks (on 1000 runs) went from 101 secs down to 87 (using GCC 4.6 -O3) So my question is : – would it hurt if it used const& instead of copy? – am I not correctly using the lib? Maybe should I use an index on the edge property and use a functor to get the data in the different operations (edge_less, combine); but wouldn't it mean re-implementing the properties? Thank you for your help

2 6

simultaneous async_read_some and async_write using strand
by Chris Stankevitz 27 Aug '12

27 Aug '12

Hello, Q1: If I call async_write while there is a pending async_read_some in a single-threaded environment, can the write complete before the read completes? (i.e. will I need to receive some data before the write goes out?) Q2: Same question, but if I am using strand to serialize the operations. Thank you, Chris

2 2

Boost graph issues with Xcode 4.4
by Jon McCormack 27 Aug '12

27 Aug '12

Hello, I'm trying to use boost 1.51 graph library with Xcode 4.4 and running into problems. While boost will build fine, when I try some simple graph code as per the documentation: #include <iostream> // for std::cout #include <utility> // for std::pair #include <algorithm> // for std::for_each #include <boost/graph/graph_traits.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/dijkstra_shortest_paths.hpp> using namespace boost; int main(int,char*[]) { // create a typedef for the Graph type typedef adjacency_list<vecS, vecS, bidirectionalS> Graph; ...etc. I get the following errors: In file included from graph.cpp:4: In file included from /usr/local/include/boost/graph/graph_traits.hpp:26: In file included from /usr/local/include/boost/pending/property.hpp:13: In file included from /usr/local/include/boost/type_traits.hpp:35: In file included from /usr/local/include/boost/type_traits/has_operator.hpp:12: In file included from /usr/local/include/boost/type_traits/has_bit_and.hpp:43: /usr/local/include/boost/type_traits/detail/has_binary_operator.hpp:220:1: error: C++ requires a type specifier for all declarations BOOST_TT_AUX_BOOL_TRAIT_DEF3(BOOST_TT_TRAIT_NAME, Lhs, Rhs=Lhs, Ret=::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::dont_care, (::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::trait_impl < Lhs, ... ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ /usr/local/include/boost/type_traits/detail/has_binary_operator.hpp:220:30: error: use of undeclared identifier 'has_bit_and' BOOST_TT_AUX_BOOL_TRAIT_DEF3(BOOST_TT_TRAIT_NAME, Lhs, Rhs=Lhs, Ret=::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::dont_care, (::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::trait_impl < Lhs, ... ^ /usr/local/include/boost/type_traits/has_bit_and.hpp:12:29: note: expanded from macro 'BOOST_TT_TRAIT_NAME' #define BOOST_TT_TRAIT_NAME has_bit_and ^ In file included from graph.cpp:4: In file included from /usr/local/include/boost/graph/graph_traits.hpp:26: In file included from /usr/local/include/boost/pending/property.hpp:13: In file included from /usr/local/include/boost/type_traits.hpp:35: In file included from /usr/local/include/boost/type_traits/has_operator.hpp:12: In file included from /usr/local/include/boost/type_traits/has_bit_and.hpp:43: /usr/local/include/boost/type_traits/detail/has_binary_operator.hpp:220:230: error: expected ';' after top level declarator ...Lhs, Rhs=Lhs, Ret=::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::dont_care, (::boost::detail::BOOST_JOIN(BOOST_TT_TRAIT_NAME,_impl)::trait_impl < Lhs, Rhs, Ret >::value)) The errors continue for a number of type_traits/detail files. Is this an issue with clang? Thanks, Jon

3 3

how to find a point from linestring?
by rakesh patil 27 Aug '12

27 Aug '12

Hi, I am new to Boost. I am using boost::geometry to define my data structure as follows typedef boost::geometry::model::point<double, 3, bg::cs::cartesian> point_3d; typedef boost::geometry::model::linestring<point_3d> linestring_3d; and I declare a linestring variable as follows linestring m_linestring; Now after adding several points in m_linestring, is there any way to find out whether a particular point exists in the list m_linestring? Just like in case of std::vector, it can be found out using std::find function. Thanks in advance

1 0

Re: [Boost-users] seed for pseudo-random generator
by Neil Nelson 26 Aug '12

26 Aug '12

For a little more entropy you can add the seconds from time to the process PID (UUID). That will give you a difference each run at the seconds numeric scale that would tend to repeat at the PID cycle rate. You could add in microseconds from http://www.boost.org/doc/libs/1_51_0/doc/html/date_time/posix_time.html (Linux gettimeofday) for additional entropy. On Linux with some additional work you can use /dev/random and /dev/urandom (better). Neil Nelson > I create within a class the MT object with > > static boost::mt19937 m_random; > > so I need for a static variable for the initialization value, so I do this with > > boost::mt19937 tools::random::m_random( time(NULL) );= > > Now: My question is, can I use the time() value for initialization the MT object, so that the MT object returns values with a good entropie or > should I need another initialization value like a UUID of the OS? > > Because the time function (http://www.cplusplus.com/reference/clibrary/ctime/time/ ) returns a time_t value, so is this value good for initialization or should I need another value, if yes, which value if a good choice for > initialize the MT object?

1 0