July 2012 - Boost-users - lists.preview.boost.org

Is there a straightforward Boost API for REST Web services calls
by Adam J Cooper 05 Jul '12

05 Jul '12

I am new to Boost and definitely still a novice with C++ in general. I would like to use Amazon Web Services (AWS) with C++. AWS supplies tookits (APIs) for .Net, Java, etc. However, they do not supply an API for C++. Therefore, I believe the best way for me to interact with AWS using C++ will be with REST. My rationale for wanting to use C++ is that I enjoy the flexibility and aesthetics of C++ and want to continue to find opportunities to learn and practice with it. (I do not have a specific application need and could use the AWS .Net API for my project if I needed to). Having said that, I understand that Boost is one of the most highly-regarded add-on libraries for C++. Therefore, is there a Boost library that will EASILY allow me to make REST (get and post) calls? If not, can someone recommend another well-tested and widely used C++ library that would allow me to make such REST calls? Thank you, Adam

5 4

Question about Program Options Vector processing for config file
by Charles Mills 05 Jul '12

05 Jul '12

I have the following in my code cnfOptions->add_options()("pacing", po::value< std::vector<float> >(&foo)->composing(), "Pacing experiment") If my config file contains pacing= 3.14 pacing= 2.718 then it works as expected (foo contains the two float numbers). But "pacing" is really a pair of values, not some collection of independent values. I've tried to find a config file syntax more like pacing= 3.14 2.718 that Boost PO will accept, but I can't find one. (I get syntax error complaints about pacing.) Is there either (a.) a syntax for a single pacing= that PO will "compose" into the vector, or (b.) a change I could make to add_options that would make this work. For example, I'm not in love with vector if there is some other container that would work better. Thanks, Charles

1 0

[Multi Index Container]: Retrieving a set by ordered non-unique criteria?
by Steve Hibbert 05 Jul '12

05 Jul '12

I'm a newbie use of Multi Index Containers, apologies if this question is facile. I have read through the Boost online documentation but am still a bit stumped. I have a container that contains shared pointers to instants of a custom object type (in this case, financial market orders). The container has an ordered-non-unique index defined by a member function of the object. Specifically it is the object's state code, an integer. At any point in time, I would expect several objects with a state value of 1, several with 2, etc. I need to pull back the set of objects in a particular state, to a set. I intend to use the set in testing to 'do something' with one random object in that set, but to all intents and purposes I just need to get the shared pointer to an item in State=1. How would I pull back the set (or analogous container) of objects that are in State=1? I am thinking that I first get a reference to the index in question, which is fine. Would I then do a find(), and would that find() operation return a set-like object? Or, would I need to do a lower_bound() and upper_bound() operation to find the two enclosing iterators, and then copy all shared-pointers-to-objects to a std::set, and then operate on the set as normal? Thanks in advance for assistance with this, help is greatly appreciated. Steve [I'm aware that if I change state values, I have to do so via the modify function, via the index, been bitten on that point before.]

2 1

Compiling boost asio 1.50.0 under Xcode 4.3.3 with Clang or GCC and c++11 support
by Graham Reitz 05 Jul '12

05 Jul '12

(Below) Is this is known issue, am I missing something, or does the libboost_system.a library need to be compiled differently than the default directions to work with c++11 support? The following code produces linker errors when the compiler options -std=c++11and libc++ are used: #include <iostream> #include <boost/asio.hpp> int main( int argc, char* argv[] ) { return 0; } Undefined symbols for architecture x86_64: "std::allocator<char>::allocator()", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string(char const*, std::allocator<char> const&)", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::allocator<char>::~allocator()", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string()", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::string::operator=(char const*)", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string(std::string const&)", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) "std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string()", referenced from: (anonymous namespace)::generic_error_category::message(int) const in libboost_system.a(error_code.o) ___tcf_0 in libboost_system.a(error_code.o) ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation)

2 2

[test] defining a macro to ignore tests
by Marco Craveiro 05 Jul '12

05 Jul '12

boost-uers, I vaguely recall a discussion a while ago regarding support for ignoring tests in boost test, and the conclusion being that its not possible. However, is it as simple as defining a macro such as this? #define BOOST_IGNORE_AUTO_TEST_CASE(test_name) void test_name() If that's the case, it would really be useful to have this in boost test. Cheers Marco -- So young, and already so unknown -- Pauli blog: http://mcraveiro.blogspot.com

3 3

[1.50.0][scope exit] dead link documentation
by Akira Takahashi 05 Jul '12

05 Jul '12

Hi, I found dead link documentation in follow page's "Tutorial" link. http://www.boost.org/doc/libs/1_50_0/libs/scope_exit/doc/html/index.html "404 Not Found File "/home/grafik/www.boost.org/archives/live/boost_1_50_0/libs/scope_exit/doc/html/scope_exit/Tutorial.html" not found. Unable to find file." Correct page is "tutorial.html". Please fix link. Thanks, Akira >>======================== Akira Takahashi mailto:faithandbrave@gmail.com site : https://sites.google.com/site/faithandbrave/about/en

2 2

[boost.python] exec_file crashes when using python 3.2
by yjaeger＠free.fr 04 Jul '12

04 Jul '12

Hi, I'm very new to boost, and I first want to thank you all for the hard and impressive work that has been done. I'm currently trying to use Boost.python and I'm facing an annoying issue. The following source code compiles fine (with msvc10 and mingw gcc 4.7): #include <boost/python.hpp> #include <iostream> using namespace boost::python; void greet() { // Retrieve the main module. object main = import("__main__"); // Retrieve the main module's namespace object global(main.attr("__dict__")); // Define greet function in Python. object result = exec("def greet():\n" " return 'Hello from Python!' \n", global, global); // Create a reference to it. object greet = global["greet"]; // Call it. std::string message = extract<std::string>(greet()); std::cout << message << std::endl; } void greet_file() { // Retrieve the main module. object main = import("__main__"); // Retrieve the main module's namespace object global(main.attr("__dict__")); // Define greet function in Python. object result = exec_file("hello.py", global, global); // Create a reference to it. object greet = global["greet"]; // Call it. std::string message = extract<std::string>(greet()); std::cout << message << std::endl; } void main(int argc, char** argv) { Py_Initialize(); greet(); greet_file(); } It even runs perfectly well when I'm using boost.python linked against python 2.7.3 library. I have following output: Hello from Python! Hello from Python! The issue appears when I link boost.python against python 3.2 (tried with 3.2.2 and 3.2.3) the first 'Hello from Python!' shows up, but then it crashes and all I can see is a windows message box saying 'prog.exe has stopped working'. It's not that I absolutely want to use python 3, but as I'm working on a new project, I'm concerned with durability, so I would prefer 3.2. Any idea would be welcome. Thanks for reading. Yannick.

1 0

[bgl] depth_first_search does not compile. color_map howto?
by Joachim Faulhaber 04 Jul '12

04 Jul '12

Hi, When using visitors with the depth_first_search algorithm I can successfully compile and run this code: //========================================================= template<class Graph> class Visitor: public default_dfs_visitor { public: typedef typename graph_traits<Graph>::vertex_descriptor Vertex; void discover_vertex(Vertex v, const Graph& g)const {cout << v << " "; return;} }; void GLV_visit() { typedef adjacency_list<listS, vecS, directedS> GraphLV; GraphLV g; add_edge(0, 1, g); add_edge(0, 2, g); add_edge(1, 2, g); add_edge(1, 3, g); Visitor<GraphLV> vis; boost::depth_first_search(g, visitor(vis)); cout << endl; } BOOST_AUTO_TEST_CASE(graph_test) { GLV_visit(); } //========================================================= With a slightly different graph (listS instead of vecS for vertices) the same call of depth_first_search fails to compile with msvc-10. //========================================================= struct Int{ Int(): _value(0){} Int(int val): _value(val){} int _value; }; void GLL_visit() { typedef adjacency_list<listS, listS, directedS, Int> GraphLL; typedef graph_traits<GraphLL>::vertex_descriptor VertexLL; GraphLL g; Int val(42); VertexLL v0 = boost::add_vertex(g); VertexLL v1 = boost::add_vertex(g); g[v0] = Int(0); g[v1] = Int(1); add_edge(v0, v1, g); Visitor<GraphLL> vis; boost::depth_first_search(g, visitor(vis)); //compile errors } //========================================================= >From the tiny error message of 108 long verbose lines of internal graph template instantiations (see attachment) it seems that the compiler is unable to create a default color_map (but I might be wrong with this interpretation of the template jungle). 1. Do I need to provide a color_map parameter here? 2. If so, what is the simplest way to do that and provide an object of the appropriate type? I'd greatly appreciate a working code example. Examples and best practices for defining and handling external propery maps in general would be helpful. TIA, Joachim -- Interval Container Library [Boost.Icl] http://www.joachim-faulhaber.de

2 5

Re: [Boost-users] [boost.test] Missing entry point when statically linking
by Tan, Tom (Shanghai) 04 Jul '12

04 Jul '12

> Date: Tue, 3 Jul 2012 19:25:50 +0000 (UTC) > From: Gennadiy Rozental <rogeeff(a)gmail.com> > >Tan, Tom (Shanghai <TTan <at> husky.ca> writes: > > #include <boost/test/unit_test.hpp> > using namespace boost::unit_test::framework; using namespace > boost::unit_test; ? > test_suite* init_unit_test_suite(int argc, char *argv[]) { ??? > BOOST_TEST_MESSAGE("hello test."); ? > ??? return 0; > } > With the same build options: > cl.exe /EHsc /MT main.cpp > ? > VC2010 compiles fine. VC2012 outputs: > LINK : fatal error LNK1561: entry point must be defined. >Don't you need to link with static library? Can you try using single header variants? >Gennadiy I am trying to use the static library, because I was under the impression that the feature in this variant would be more complete than others. With VC++ I do not have to explicitly specify the static file name in the command line if it's what you mean. And if this is the problem, the error would be like LINK : fatal error LNK1104: cannot open file 'libboost_unit_test_framework-vc110-mt-s-1_50.lib' I had a re-read of the single header variant documentation. Seems I can just switch to it. But it still seems strange that the static variant works with VC10 but not VC11.

2 1

Re: [Boost-users] [boost.test] Missing entry point when statically linking
by Tan, Tom (Shanghai) 04 Jul '12

04 Jul '12

>Date: Tue, 3 Jul 2012 19:52:31 +0100 >rom: Marco Craveiro <marco.craveiro(a)gmail.com> > >Hi Tom, > >shouldn't it be > >#include <boost/test/included/unit_test.hpp> > >on the include? Thanks, Marco. I was intending to use the library variant because I recall reading somewhere the header only variant missing some test tools. Judging from this link: (http://www.boost.org/doc/libs/1_50_0/libs/test/doc/html/utf/user-guide/ usage-variants/single-header-variant.html) I just read, I seem to be wrong. The code in question was in place since around 2005. I never thought of changing it until it's rejected by vc11. Thanks for the suggestion.

1 0