January 2010 - Boost-users - lists.preview.boost.org

[function_types] About parameter_types
by Olivier Tournaire 30 Jan '10

30 Jan '10

Hi all, I am trying to extract the parameters types of a given function/functor in a mpl sequence. Digging boost libraries, I came across function_types, and in particular parameter_types, which might be the right tool. However, I cannot figure how to use it. Could you please help ? Suppose I have: void foo(int, double); And would like to generate the following sequence: mpl::vector<int, double> How can I do that ? Regards, Olivier

2 2

[doc] Add "Last Update" information
by Marco Guazzone 30 Jan '10

30 Jan '10

Dear Boosters, In my opinion it would be very helpful if the documentation include for each library the version of the last update. For instance, one could add this information in the summary page (http://www.boost.org/doc/libs/1_41_0): --- [snip] --- uBLAS uBLAS provides matrix and vector classes ... Author(s)...........Joerg Walter and Mathias Koch First Release...1.29.0 -> Last Update......1.40.0 <- Standard Build & Link......Header only Categories........Math and numerics --- [/snip] --- The reason for this is that it should simplify the discovery of what version of Boost one needs for his own software project. As a matter of fact, the use of "First Release" might not be sufficient because after the initial release of a library many other features could have been added. So the need for this additional piece of information. What do you think? Regards, -- Marco

1 0

asio UDP local port binding issue on Windows
by akos＠maroy.hu 30 Jan '10

30 Jan '10

Hi, I seem to encounter an issue when using boost asio. I'm trying to create a UDP client-server environment, where for the client I'm creating an asio socket like this: io_service _io_service; ip::udp::socket _socket(_io_service, ip::udp::v4()); my expectation at this point is that the _socket is created, and it can send and receive UDP packets. this includes receiving packets, which also means that the socket is bound to some random port on the local network interface. now this is all true on Linux and MacOS, but on Windows, the socket does not seem to bind to a port. interestingly, for _socket.is_open() I get true in return. but for _socket.local_endpoint(), I get an error excaption with the error code 10022, which is defined as WSAEINVAL. I also tried the following, and got the same results: io_service _io_service; ip::udp::socket _socket(_io_service); _socket.open(ip::udp::v4()); is my assumption that the above should result in a socket bound to a local random port incorrect? Akos

1 0

Property Tree - adding child nodes to a parent
by Jon Shuler 30 Jan '10

30 Jan '10

I am trying to build a property tree from hierarchy data in a database. I want to be able to add a child at a given node using something similar to: parentNode.add_child("child name", data_type). There is a add_child method but it expects the path to add the child at. I looked at the examples and did not see anything. Is this possible? Jon This e-mail message and any attachment(s) transmitted with it are intended only for the use of the recipient(s) named above.This message may be privileged and/or confidential. If you are not an intended recipient, you may not review, copy or distribute this message. If you have received this communication in error, please notify us immediately by e-mail and delete the original message.

2 1

Re: [Boost-users] Detect specialized template function?
by Andrew Chinkoff 30 Jan '10

30 Jan '10

Hi, Daniel. This is your example: template <class T> void S::f() { if(g is specialized) // (1) callbacks_ready=true; // else wait for a callback object to be registered } At (1) the runtime detection is used, but "S:g()" was specialized at compile time. Templates are specialized and compiled at compile time, but not runtime. We can't write "if(g is specialized)" - it is impossible to realize it at runtime (and it is senseless). So let me to recommend you some books of H. Sutter and A. Alexandrescu. There are a lot of recommendations about templates, their specializations, overloads, static/dynamic polymorphisms and more. 1st book: [http://www.gotw.ca/publications/xc++s.htm] Pay attention to this chapter of the book: "Item 7: Why Not Specialize Function Templates?" 2nd book; "C++ Coding Standards: 101 Rules, Guidelines, and Best Practices" (authors H. Sutter and A. Alexandrescu) Pay attention to 64, 65, 66 and 67 chapters. I hope you will find some useful ideas about templates in these books. There are also Boost libraries targeted to realize static polymorphism: 1) http://www.boost.org/doc/libs/1_40_0?view=category_Generic 2) http://www.boost.org/doc/libs/1_40_0?view=category_Metaprogramming Best, Andrew. > Slightly OT, but I was hoping someone here might know the magic incantation > (which may involve a Boost library). > Given a template class such as > template <class T> > struct S > { > void f(); > void g() { T t(h()); cb(t) } > T & h(); > boost::function1<void, T &> cb; > }; > Is there a way to have S.f() behave differently depending on whether S.g() uses > the default definition or whether it is specialized for this particular type? > This detection needs to occur before S.g() is ever called. > For example, the mere existance of the specialization > template <> void S<int>::g() { ... } > should trigger special logic in S<int>::f(). > Thanks, > Daniel > P.S. The motivation for this is an attempt to retrofit a callback API that > allows specialization of S.g() or the registration of a boost::function that > gets called by the default S.g(). We need to disable callbacks until an actual > function is registered via either method. It is easy to dispatch on > registration of the boost::function; I don't know how to detect the > specialization.

1 0

Boost 1.36.0 with Linux 64bits
by JongKwan Kim 29 Jan '10

29 Jan '10

Hi all, I would like to know if the boost 1.36.0 is working well with Linux 64 bits. If I want to compile with boost 1.36.0 and Linux 64 bits, how should I install the boost 1.36.0? Is it the same to 32 bits machine? Please, if someone knows that, let me know. Thanks,

5 8

fusion sequences, type erasure across translation units
by Hicham Mouline 29 Jan '10

29 Jan '10

Hello, I have a .cpp file which a sort of top level driver. I have say 100 different struct param_i {...}; where 1<=i<=100 Each param_i is defined in a param_i.hpp and param_i.cpp invokes metafunctions to obtain typedef fusion< variant< vector<double>, triplet<double>, pair<double>, double>, variant<...>, ..., variant<...> > seq_i_t; this sequence type is visible only in the param_i.cpp, not the header file. Generation of this sequence is The size of these sequences = number of members of param_i. The toplevel driver and other .cpp need to move instances of these types around to other .cpp. To avoid having the compiler call the metafunctions for these middle translation units, but only at the generating .cpp and at the consuming .cpp, I created a tag struct fusion_seq_tag {}; typedef <typename Arg> struct make_fusion_seq_i { typedef <undefined> type; }; typedef <typename Arg> struct make_fusion_seq_tag { typedef typename make_fusion_seq_i<Arg>::type base_type; struct type : fusion_seq_tag, base_type { type(...) : base_type(...) {} }; }; return smart_ptr<fusion_seq_tag>( new make_fusion_seq_tag<...>::type ); // this was the generating .cpp The function returns smart_ptr<fusion_seq_tag> which can move around in other .cpp without any metaprogramming until reaching the consuming .cpp Is this type erasure? I don't know how I can get the actual type back in the consuming .cpp if I don't know the originator but if I know only the shape of the fusion sequence. How to dynamic cast the type back to its unerased type? If I know the originator .cpp then I would know the exact fusion sequence and therefore I would just dynamic_cast the fusion_seq_tag. Regards,

2 1

[boost][task] task and fiber do not compile
by Kishalay Kundu 29 Jan '10

29 Jan '10

Hi Oliver, I have boost 1.41.0 installed. I have been trying to build boost-task and boost-fiber (I guess this is the pre-req for building boost-task). However when I do: "./bjam threading=multi --with-task", the compiler errors out with a long list of compiler errors. I have gcc 4.4.1 installed on my Fedore 11 system. Could you please let me know what I am doing wrong? Thank you. regards, Kishalay

1 0

[BGL] Color Maps
by pablo barrio lópez-cortijo 29 Jan '10

29 Jan '10

Hi there! I'm new to Boost (and also to this mailing list). I'm using depth_first_search and it's vital for me to be able to specify the root vertex (4th argument), and that means also specifying the Color Map, which I don't care too much about. It'd be great to have a default color map but I'm a newbie and I didn't found anything in the docs. My initial options were: a) Using sth like a "default_color_map" class. I ignore if there is such thing. b) Creating a property map out of a vector, something like "vector_property_map<default_color_type>". I tried and it didn't work. c) Including a property in my graph called "color" of type "default_color_type", and creating a property map out of it. Didn't work either. Actually, I get an error when trying to include the property with that type (compiles if the property is an int). d) Just copying the color map that the algorithm uses when called without the color map argument. Hoping to find out how the BGL code works was simply unrealistic so I abandoned the idea. Is there any way to easily instantiate a Color Map? I'd post my code, but I have 6 or 7 versions none of which work. Any help will be much appreciated. Thanks ahead!

2 1

Detect specialized template function?
by dherring＠ll.mit.edu 29 Jan '10

29 Jan '10

Slightly OT, but I was hoping someone here might know the magic incantation (which may involve a Boost library). Given a template class such as template <class T> struct S { void f(); void g() { T t(h()); cb(t) } T & h(); boost::function1<void, T &> cb; }; Is there a way to have S.f() behave differently depending on whether S.g() uses the default definition or whether it is specialized for this particular type? This detection needs to occur before S.g() is ever called. For example, the mere existance of the specialization template <> void S<int>::g() { ... } should trigger special logic in S<int>::f(). Thanks, Daniel P.S. The motivation for this is an attempt to retrofit a callback API that allows specialization of S.g() or the registration of a boost::function that gets called by the default S.g(). We need to disable callbacks until an actual function is registered via either method. It is easy to dispatch on registration of the boost::function; I don't know how to detect the specialization.

2 1