- Boost-users - lists.boost.org

[PtrContainer] Inserting "this" Pointers
by Hossein Haeri 22 Sep '10

22 Sep '10

Dear all, Is inserting "this" pointers into pointer containers valid? If so, what are the consequences? Would that need special care for example? TIA, --Hossein

2 2

circular_buffer iterators
by Bosko Markovic 22 Sep '10

22 Sep '10

I have a simple use case for circular_buffer but I need a few advices. There is one producer and one consumer which only inspects elements and circular buffer will only be keeping history of last CB_SIZE elements. As far as I have understood there are two ways to access buffer elements, using indexes and operator[] or using iterators. I have chosen iterators. Since adding elements to buffer can invalidate iterators what is the proper way to check if they are still valid? Is it comparing it to .end() and .begin()? typedef boost::circular_buffer<data_type> CircularBuffer; CircularBuffer cb(CB_SIZE); CircularBuffer::iterator it = cb.begin(); ... if( it >= cb.begin() && it < cb.end() ){ cout << *it; } Regards, BoskoM.

2 2

Compiling boost::regex v4 with VC8 on Windows
by Wolfgang Goldbach 22 Sep '10

22 Sep '10

Hi, In my current project boost::regex was used previously from version v3 (boost 1.30.2) and there was no problem. Because I use a third party SDK that contains boost 1.33.1 I want to use boost and regex (v4) from this SDK. In the project there is a source file that includes "boost/regex.hpp". In the release version the file compiles, but in the debug version I get a lot of errors like: char_regex_traits.hpp(34) : error C2976: 'boost::regex_traits' : too few template arguments /basic_regex.hpp(186) : error C2825: 'traits': must be a class or namespace when followed by '::' basic_regex.hpp(186) : error C2039: 'size_type' : is not a member of '`global namespace'' I have now removed the preprocessor definition '_DEBUG' from the VC8 project that contains the source file but this makes no difference. Is there any idea what's going on here ? Wolfgang

2 2

boost thread, shared_mutex deadlock
by George George 22 Sep '10

22 Sep '10

threadA owns a shared_lock on the mutex, threadB tries to get a unique lock on the mutex and blocks, threadC tries to get a shared_lock on the mutex and blocks also!! I guess there is a LIFO kind of stack serving the mutex to the requesting locks. But what happens if threadA depends on threadC in order to complete and release the lock?I n this case I have got a deadlock. Is this a design flow of my app or boost?

2 1

[Serialization] compilation error on boost/serialization/item_version_type.hpp
by Paul Graphov 22 Sep '10

22 Sep '10

Hello Boost Users, File mentioned in subj (included via boost/serialization/vector.hpp -> /boost/serialization/collections_save_imp.hpp) fails to compile at my Ubuntu box because of blatant use of assert function. Problem was solved by including <boost/assert.hpp> and changing assert to BOOST_ASSERT Is it a bug or just my misconfiguration? Thanks.

1 0

Unresolved overloaded function type with local_time and function
by Nevin Liber 21 Sep '10

21 Sep '10

I'm getting the dreaded "unresolved overloaded function type" error on the following line of code: boost::function<boost::posix_time::ptime()> fn(boost::posix_time::microsec_clock::local_time); If I wrap the call to local_time() in a function or a function object, it works without any issues. I don't see any other overloads for local_time(). I'm sure it's something simple I'm just not seeing... -- Nevin ":-)" Liber <mailto:nevin@eviloverlord.com> (847) 691-1404

2 2

[Proto]: How to use complex grammars in a domain/extension
by Roland Bock 21 Sep '10

21 Sep '10

Hi, I am trying to create a DSEL with Boost.Proto, which (among other things) should allow expressions like i == j + 3 && t != "foo" and want to prevent disallowed operators to be generated by Proto. Please consider the code below. // ----------------------------------------------------------------------- #include<boost/proto/proto.hpp> using namespace boost; typedef proto::terminal<int>::type terminal; struct addition: proto::or_ < terminal, proto::plus<addition, addition> > {}; struct equation: proto::equal_to<addition, addition> {}; struct condition: proto::or_ < equation, proto::logical_and<equation, equation> > {}; template<class Expr> struct extension; struct my_domain: proto::domain < proto::pod_generator< extension>, condition > {}; template<class Expr> struct extension { BOOST_PROTO_EXTENDS( Expr , extension<Expr> , my_domain ) }; int main() { terminal c; BOOST_PROTO_ASSERT_MATCHES(c + c, addition); // OK BOOST_PROTO_ASSERT_MATCHES(c == c, equation); // OK BOOST_PROTO_ASSERT_MATCHES(c == c, condition); // OK BOOST_PROTO_ASSERT_MATCHES(c == c&& c == c, condition); // OK extension<terminal> i; // I was hoping for the following to be legal i + i; // ERROR: operator+ not found i == i; // ERROR: operator== not found i == i&& i == i; // ERROR: operator== not found } // ----------------------------------------------------------------------- In my_domain, I can replace condition by addition, which allows i + i, but then the == operator is still missing (and the &&, too, probably). How can I solve this? Thanks and regards, Roland

6 18

How to write a predicate for mpl::find_if that is able to work with placeholders?
by Michael Schulze 21 Sep '10

21 Sep '10

Hi list, I have a problem that bothers me since days. It seems, that I am always doing something wrong. I try to write a meta function, that is able to work with placeholders and is applicable to mpl::find_if. I understand the example from the documentation where is_same is used, which is in my understanding also a meta function. is_same gets always the current element of the sequence and compares it with the type given as second template argument to itself. Correct me if I am wrong at this point. So far so good, what I need is not the comparison to the type but instead a comparison with an encapsulated type, a kind of tag type if you want. Here is my class template template<int i> struct Slot { typedef Slot type; typedef int_<i> slot_id; }; and I would like to test whether a certain slot_id is contained within a sequence, because the Slot class template may have different further template arguments, here omitted for the sake of clarity. My created meta function is: template<typename A, typename B> struct contained : is_same< typename A::slot_id ,B> {}; A possible sequence may be defined as: typedef vector<Slot<0>, Slot<10>, Slot<15> > slots; and the find_if's invocation should look like this typedef find_if<slots, contained< _1, int_<10> > >::type t10; and a comparison with the dereferenced returned iterator should give true BOOST_MPL_ASSERT(( is_same< deref<t10>::type, Slot<10> > )); However, until now I have not much luck with my different tries because I get always compile errors that state that the mpl::arg<1> did not has slot_id type. This is totally right, however, how I reach it that not the placeholder is given to the meta function but instead the current element of the sequence to which the iterator points and for which the placeholder is a substitute. If I reflect it seems I have not fully understand how the placeholder mechanism works in detail but I would like to get it. Okey, I have two solution that work and give me the correct result for my problem. I created two different "contained" definitions and find_if invocations 1. version: template<typename Nr> struct contained1 { template <typename Slot> struct apply : ::boost::is_same< typename Slot::slot_id, Nr > {}; }; typedef find_if<slots, contained1<int_<10> > >::type t11; 2. version: struct contained2 { template<typename A, typename B> struct apply : is_same< typename A::slot_id,B>::type {}; }; typedef find_if<slots, bind< contained2, _1, int_<10> > >::type t12; Both versions work, because they are meta function classes, but the invocation of find_if and how the parameter is delivered to the predicate is not clear at least in the first version. In the second version this is clear but we need an additional bind. Is there a way, to get my initial version working, and to use a meta function instead of a meta function class. In my eyes, my initial find_if invocation is clearer and more readable e.g. for maintainers. Thanks and regards Michael

1 2

graph: PropertyMap needs lvalue for vertex_index
by Jens Weller 21 Sep '10

21 Sep '10

Hi, I've been playing arround with boost::graph (and the layout algorithms in special). Now I encountered a little problem, I'm trying to get Kamada Kawai Spring to work, but I do get an compile error when calling put for vertex_index. So, this problem has nothing to do with KKS. My Code: typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS, // Vertex properties boost::property<boost::vertex_index_t, int, boost::property<vertex_position_t, point> > // Edge properties ,boost::property<boost::edge_weight_t, double> > Graph; Graph g(scene->getNodeCount()); boost::graph_traits<Graph>::vertex_iterator vi,vi_end; int i = 0; for (boost::tie(vi, vi_end) = boost::vertices(g); vi != vi_end; ++vi) boost::put(boost::vertex_index, g, *vi, i++); put throws now a compile Error: ..\..\..\cpp\boost_1_43/boost/property_map/property_map.hpp:361: error: lvalue required as left operand of assignment The last argument of put does throw this error. I expected the code to work, as I have copied it out of the layout_tests file: http://www.boost.org/doc/libs/1_43_0/libs/graph/test/layout_test.cpp So, where is my error here? I tried some different versions and some casts, but none made this code work. And I don't think that something that trivial is broken in the testcases. regards, Jens Weller -- GRATIS: Spider-Man 1-3 sowie 300 weitere Videos! Jetzt freischalten! http://portal.gmx.net/de/go/maxdome

2 2

this_thread::get_id() never returns occasionally
by Samuel Williams 21 Sep '10

21 Sep '10

Hi, I'm running boost on an iPhone 3G compiled with the official 3.2 SDK on a regular iPhone (i.e. not jailbroken). Sometimes I get a strange situation where calling this_thread::get_id() never returns. #0 scoped_lock [inlined] () at spinlock_gcc_arm.hpp:45 #1 boost::detail::spinlock::lock () at boost_1_44_0/boost/smart_ptr/detail/spinlock_pool.hpp:41 #2 boost::detail::sp_counted_base::weak_add_ref () at spinlock_gcc_arm.hpp:61 #3 atomic_increment [inlined] () at boost_1_44_0/boost/smart_ptr/detail/sp_counted_base_spin.hpp:364 #4 0x000b6c38 in boost::(anonymous namespace)::make_external_thread_data () at spinlock_gcc_arm.hpp:189 #5 0x000b849c in boost::this_thread::get_id () at libs/thread/src/pthread/thread.cpp:189 Is this a bug? Kind regards, Samuel

2 1