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

[graph] adjacency_list constructor fails, wrong usage?
by Christoph 15 Nov '11

15 Nov '11

Hi, I originally wanted to construct an undirected graph from a directed graph and vice versa. For this purpose i tried to use template <class EdgeIterator> adjacency_list(EdgeIterator first, EdgeIterator last, vertices_size_type n, edges_size_type m = 0, const GraphProperty& p = GraphProperty()) as shown in http://www.boost.org/doc/libs/1_47_0/libs/graph/doc/adjacency_list.html but i do not even manage to construct a graph from the same type with this constructor. What am i missing? Please have a look at the example below. If you uncomment the last two lines, the code will not compile. // begin code #include <iostream> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/graph_utility.hpp> int main(int argc, char** argv) { typedef boost::adjacency_list < boost::listS, boost::vecS, boost::directedS> Dgraph; typedef boost::adjacency_list < boost::listS, boost::vecS, boost::undirectedS> Graph; Dgraph g(5); add_edge (0,1,g);add_edge (2,3,g);add_edge (4,5,g); print_graph (g, boost::get(boost::vertex_index, g)); boost::graph_traits<Dgraph>::edge_iterator i1 = (edges(g)).first; boost::graph_traits<Dgraph>::edge_iterator i2 = (edges(g)).second; // Dgraph g2 = Dgraph (i1, i2, num_vertices (g), num_edges (g)); // print_graph (g2, boost::get(boost::vertex_index, g2)); return 0; } // end code best regards Christoph

3 3

Using an mpl sepquence as a template parameter
by paul Fultz 14 Nov '11

14 Nov '11

Hi, I have class that takes an mpl sequence as its parameter like this: X<mpl::vector<A, B, C> > And I would like to be able to have the class be used without explicitly declaring the mpl sequence like this: X<A, B, C> I know this could be accomplished using variadic template parameters, but I am using an older compiler. I was wondering if there was a macro in boost to do this, such as: template<BOOST_MPL_VARIADIC_PARAM_PACK(T)> class X { typedef mpl::vector<BOOST_MPL_VARIADIC_PARAM_UNPACK(T)> type; }; If a macro is not available, does anyone know how I could write a macro like this? Or what is the default type used for mpl sequences when a type is unspecified? Thanks, Paul

3 3

Using an mpl sepquence as a template parameter
by paul Fultz 14 Nov '11

14 Nov '11

Hi, I have class that takes an mpl sequence as its parameter like this: X<mpl::vector<A, B, C> > And I would like to be able to have the class be used without explicitly declaring the mpl sequence like this: X<A, B, C> I know this could be accomplished using variadic template parameters, but I am using an older compiler. I was wondering if there was a macro in boost to do this, such as: template<BOOST_MPL_VARIADIC_PARAM_PACK(T)> class X { typedef mpl::vector<BOOST_MPL_VARIADIC_PARAM_UNPACK(T)> type; }; If a macro is not available, does anyone know how I could write a macro like this? Or what is the default type used for mpl sequences when a type is unspecified? Thanks,Paul

1 0

How to use boost::iostreams filter directly
by Allan Nielsen 14 Nov '11

14 Nov '11

Hi How can I use boost::iostreams::gzip_decompressor to decompress a sequence of boost::asio::streambuf's? Here is what I had in mind (non-working pseudo code): struct Foo { public: void sinkData(boost::asio::streambuf & buf) { z.write(boost::asio::buffer_cast<const char*>(buf.data()), buf.size()); while(write()); } void flush() { z.flush(); while(write()); } private: bool write() { char buf[1024]; size_t s = z.read(buf, 1024); std::cout << std::string(buf, s) << std::endl; return } gzip_decompressor z; } The sinkData and the flush functions can not be changed. sinkdata() is called a number of times, followed by a single call to flush(). Is boost::iostreams::gzip_decompressor the best library for this purpose, if so how to do it? Otherwise, what could else be suggested? Best regards Allan W. Nielsen

1 0

[smart_ptr] Reinterpret_pointer_cast()?
by Robert Jones 14 Nov '11

14 Nov '11

Even as I ask this I suspect the answer is obvious and I can't see it, but is there any reason why there is no reinterpret_pointer_cast() function for shared_ptr's? Thx, - Rob.

3 2

Using boost Barser to extract data from string
by Gaurav Shah 14 Nov '11

14 Nov '11

> > Hi, > I want to write parser for data I receive from a stock exchange server. > There is no delimeter that define different fields in buffer I receive. > However, it is provided in exchange spec's start and of each field. > How should I write parser for data I recveive? I am new to parser consept > and apprecite your suggestion. > I am putting following example data string just to insist my scenario > > Data my application recveive: > 1001TXPASSWORDSESSION1N > > I need to decode it(as per spec) as: > Messate type: 1001 > UserID: TX > password: PASSWORD > session: SESSION > data: N > > How should I be doing it? > > Thank You, > Gaurav >

3 3

[fusion] more trouble with transform
by Gennadiy Rozental 13 Nov '11

13 Nov '11

Hi, Now I am trying to perform transform which mutates: ------ #include <boost/fusion/container/list/list.hpp> #include <boost/fusion/container/list/convert.hpp> #include <boost/fusion/container/generation/make_list.hpp> #include <boost/fusion/algorithm/transformation/transform.hpp> namespace fusion=boost::fusion; struct C {}; struct B { C make_c() { return C(); } }; struct make_c { template<typename Sig> struct result; template <typename T> struct result<make_c(T&)> { typedef C type; }; template <typename T> struct result<make_c(T)> { typedef C type; }; template <typename T> C operator()(T& t) const { return t.make_c(); } }; int main() { B b1, b2; auto z1 = fusion::make_list( b1, b2 ); auto z2 = fusion::as_list( fusion::transform( z1, make_c() ) ); return 0; } ---- Both gcc and msvc complain like this: test.cpp:29:46: error: passing ‘const B’ as ‘this’ argument of ‘C B::make_c()’ discards qualifiers. What am I missing now? Gennadiy

4 9

Re: [Boost-users] [Spirit-general] Pattern matching with boost
by Alec Taylor 13 Nov '11

13 Nov '11

Thanks Seth, I would certainly like to have UNICODE support, so perhaps the AXE libraries would be a proper solution :) I just need to see if it will work with the MSVC10 compiler, then it will be the solution I am looking for :) The other option is to use something such as iswdigit with the unicode string library: ICU There's a lot more to consider, but it seems like the boost libraries aren't what I require for this project... so I will keep discussion off of these mailing-lists in future. Thanks all for your solutions, Alec Taylor On Sun, Nov 13, 2011 at 3:37 AM, Seth Heeren <bugs(a)sehe.nl> wrote: > On 11/12/2011 03:13 PM, Alec Taylor wrote: > > I am thinking to piece this together using just the stdlib > (std::isdigit &etc)... what do you think? > > By all means. My money is on the hybrid regex+stdlib approach. > > I haven't had the time to evaluate the Axe Library suggested on SO. On first > sight it looked like a regex+native hybrid but with nicer library support. > > But if you have a plan of attack, just go for it. Please, keep in mind that > this is the Boost Spirit general mailing list, and therefore not really > suited for continued discussion in other directions :) > > Cheers, > Seth >

1 0

boost::mpl
by Peter Foelsche 13 Nov '11

13 Nov '11

Again -- my thanks to the developers of boost::mpl -- something I'm really use and appreciate a lot! Some requests: 1) I talked already in the past about the value_init stuff used in for_each. I'm not certain which kind of object needs it's memory initialized to zero to work. Anyway -- even in case of pod types the contents of the value does not matter, AS ONLY THE TYPE IS NEEDED to select one or the other template routine. 2) The routine mpl::for_each must be inlined. I got an object file of a size which I've never seen before!!!! And memory usage by the compiler of up to 12GB! And after some examination it turned out, that some routines called by for_each are not inlined. And I did add again (like I did already some time ago and reported here) some __forceinline (the equivalent for g++ is some attribute) to some of the routines in boost/mpl/for_each.hpp Now everything is fine -- the compiler finishes after an appropriate time and things compile again on 32bit. Peter Foelsche

1 0

[ASIO] Flush output on socket
by Claude 12 Nov '11

12 Nov '11

Hi! I use this code for send a std::string with a socket: I have a doubt: this function automatically flush its output? -- View this message in context: http://boost.2283326.n4.nabble.com/ASIO-Flush-output-on-socket-tp3951921p39… Sent from the Boost - Users mailing list archive at Nabble.com.

2 6