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

[function_types] comparing arity
by Jeff Flinn 03 Mar '12

03 Mar '12

I'm trying to come up with the condition for an enable_if condition based on the function_arity of a signature. The following fails to compile: bool z = boost::mpl::equal_to < boost::function_types::function_arity<SIG>::value , boost::mpl::int_<3> >; Any help is appreciated. Thanks, Jeff

1 2

enable_if and templated operator() of member of class template
by MM 02 Mar '12

02 Mar '12

Hello, I am using the enable_if construct to disable an operator() template member function of a templated variant visitor: template <typename iter_t> class iterator_compare_visitor_ : public boost::static_visitor<bool> { public: iterator_compare_visitor_(const iter_t& to_compare_to) : to_compare_to_(to_compare_to) {} template <typename variant_iter_t> typename std::enable_if< boost::mpl::contains<const_iterators_t, iter_t>::value, bool >::type operator()(const variant_iter_t& it) const { return std::is_same<variant_iter_t, iter_t>::value && it==to_compare_to_; } template <typename variant_iter_t> typename std::enable_if< boost::mpl::contains<const_reverse_iterators_t, iter_t>::value, bool >::type operator()(const variant_iter_t& it) const { return std::is_same<variant_iter_t, iter_t>::value && it==to_compare_to_.base(); } private: const iter_t& to_compare_to_; }; The compilation fails in msvc2010 with: error C2039: 'type' : is not a member of 'std::tr1::enable_if<_Test,_Type>' 1> with 1> [ 1> _Test=false, 1> _Type=bool 1> ] in this case, iter_t is not contained in const_reverse_iterators_t and so the 2nd operator() should not be enabled. Rds, MM

3 4

Using boost::container::map on OpenVMS
by De Rudder, Stephen L. 02 Mar '12

02 Mar '12

I am trying to get the boost::container::map to work on OpenVMS (Alpha). I am getting an %CXX-E-BADBASECLS, (1) invalid base class error but have not been able to track it down. If I use std::map instead of boost::container::map in the code then it will work btw. Thanks for any help, SLDR (Stephen L. De Rudder) rdls22 AT tditx.com More Info: OpenVMS V8.2 on Alpha $ cxx /ver HP C++ V7.3-009 for OpenVMS Alpha V8.2 The compile command I am using is: $ cxx/stand=strict_ansi/list/show=all/inc=userdisk4:[boost_1_49_0] t.cpp I have enclosed the example code and the section of the listing output from the cxx compile. #include <tree> #include <stdio.h> #include <stdlib.h> #include <boost/container/string.hpp> #include <boost/container/map.hpp> #define cwstring std::string class SecPrivOrderRec { public: boost::container::map<long, bool> orderMap; //map order -> grant }; class SecPrivOrderMap { public: boost::container::map<long, SecPrivOrderRec> privMap; //map priv -> order/grant }; class SecUserProfCompPrivMap { public: SecUserProfCompPrivMap(); bool valid; SecPrivOrderMap keyPrivMap; // map priv -> order }; SecUserProfCompPrivMap::SecUserProfCompPrivMap() { this->valid = false; } class SecUserProfPrivMap { public: SecUserProfPrivMap(); bool valid; boost::container::map<long, long> compPrivMap; // map priv -> order boost::container::map<cwstring, SecUserProfCompPrivMap> keyMap; // map comp -> key }; SecUserProfPrivMap::SecUserProfPrivMap() { this->valid = false; } class SecUserProfMap { public: boost::container::map<cwstring, SecUserProfPrivMap> compMap; // map comp -> priv }; int main(int argc, char **argv) { boost::container::map<std::string, int> j1; j1["foobar"]=22; boost::container::map<cwstring, SecUserProfMap> userProfileMap; //map profile -> comp -> key -> priv long maxndx = 10000; long maxndx2 = 1; int testtype = 0; char ndxstr[16]; cwstring s1 = "my_large_test_user"; cwstring s2 = "this_is_a_key"; cwstring s3; //fprintf(stderr,"Running test type 0 -- expected to consume 10x memory\n"); for (long ndx = 0; ndx < maxndx; ndx++) { //sprintf(ndxstr, "%d", ndx); for (long ndx2 = 0; ndx2 < maxndx2; ndx2++) { s3 = ndxstr; userProfileMap[s1].compMap[s2].keyMap[s3].keyPrivMap.privMap[ndx2].orderMap[1] = true; } } } ========================================================================================================= Source Listing 2-MAR-2012 10:49:24 HP C++ V7.3-009 Page 1910 1-JAN-2012 13:14:12 has_member_function_callable_with.hpp;1 I6 92339 ␍ I6 92340 BOOST_INTRUSIVE_HAS_MEMBER_FUNCTION_CALLABLE_WITH_NS_BEGIN␍ E namespace boost { namespace container { namespace container_detail { I6 92341 ␍ I6 92342 template <typename Type>␍ I6 92343 class BOOST_PP_CAT(has_member_function_named_, BOOST_INTRUSIVE_HAS_MEMBER_FUNCTION_CALLABLE_WITH_FUNCNAME)␍ E class has_member_function_named_select_on_container_copy_construction I6 92344 {␍ I6 92345 struct BaseMixin␍ I6 92346 {␍ I6 92347 void BOOST_INTRUSIVE_HAS_MEMBER_FUNCTION_CALLABLE_WITH_FUNCNAME();␍ E void select_on_container_copy_construction(); I6 92348 };␍ I6 92349 ␍ I6 92350 struct Base : public Type, public BaseMixin { Base(); };␍ ..............................1 %CXX-E-BADBASECLS, (1) invalid base class detected during: instantiation of class "boost::container::container_detail::has_member_function_ named_select_on_container_copy_construction<Type>::Base [with Type=const boost::container::container_detail::node_alloc_holder<std ::allocator<std::pair<const std::string, SecUserProfPrivMap>>, boost::container::container_detail::intrusive_rbtree_type <std::allocator<std::pair<const std::string, SecUserProfPrivMap>>, boost::container::container_detail::value_compare_impl<st d::string, std::pair<const std::string, SecUserProfPrivMap>, std::less<std::string>, boost::container::container_detail::select1st<std::pair<c onst std::string, SecUserProfPrivMap>>>>::type, std::less<std::string>>::NodeAlloc]" at line 98 of "USERDISK4:[DERUDDER.BOOST_1_49_0.BOOST.INTRUSIVE.DETAIL] has_member_function_callable_with.hpp;1"

1 0

How to assign a preallocated buffer boost::format?
by niXman 02 Mar '12

02 Mar '12

Hello, I need to assign a preallocated buffer for the boost::format. Any ideas? Thanks. -- Regards, niXman

2 1

Re: [Boost-users] Interprocess container segmentation fault (example provided)
by Davidson, Josh 02 Mar '12

02 Mar '12

Ion, Terribly sorry, but I goofed when I cobbled that code together. My actual application constructs "TestStore" from the segment, and not on the stack like what I pasted in before. I'm including the full test code again, although only the main function changed. Since the map itself a member of TestStore and TestStore is being constructed from the segment, I would think that should be roughly equivalent to just creating the map by itself from the segment. However, as you'll see if you run it, it still seg faults when accessing the list iterator. ===================================== BEGIN CODE============================================== #include <stdint.h> #include <cstddef> #include <string> #include <functional> #include <iostream> #include <utility> using std::cerr; using std::endl; using std::pair; #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <boost/interprocess/offset_ptr.hpp> #include <boost/interprocess/containers/string.hpp> #include <boost/interprocess/containers/list.hpp> #include <boost/interprocess/containers/map.hpp> #include <boost/ref.hpp> using namespace boost::interprocess; class SharedMemoryDeleter { public: SharedMemoryDeleter(const std::string& name) : mName(name) { shared_memory_object::remove(mName.c_str()); } ~SharedMemoryDeleter() { shared_memory_object::remove(mName.c_str()); } protected: std::string mName; }; class TestStore { public: struct TestData { uint16_t commandStatus; size_t wordCount; uint16_t data[32]; }; ///Allocates characters in shared memory typedef allocator<char, managed_shared_memory::segment_manager> CharAllocator; ///String in shared memory typedef basic_string<char, std::char_traits<char>, CharAllocator> SharedMemoryString; ///Allocates list in shared memory typedef allocator<TestData, managed_shared_memory::segment_manager> ListAllocator; ///List of TestData items in shared memory typedef list<TestData, ListAllocator> SharedMemoryList; ///Key/Value pair for the map (SA, Data List) typedef pair<const unsigned, SharedMemoryList> data_map_value_t; ///Allocates maps in shared memory typedef allocator<data_map_value_t, managed_shared_memory::segment_manager> MapAllocator; ///Maps lists to SAs in shared memory typedef map<unsigned, SharedMemoryList, std::less<unsigned>, MapAllocator> SharedMemoryMap; TestStore(managed_shared_memory& segment, const std::string& name) : mSegment(&segment), mSegmentName(name.c_str(), CharAllocator(mSegment->get_segment_manager())), mData(std::less<unsigned>(), MapAllocator(mSegment->get_segment_manager())){ } /////////////////////////////////////////////////////////////////////////// /// Destructor /////////////////////////////////////////////////////////////////////////// virtual ~TestStore() { } void printData() { TestStore::SharedMemoryMap::iterator it = mData.begin(); while(it != mData.end()) { TestStore::SharedMemoryList::iterator lIt = it->second.begin(); cerr << "SA: " << it->first << endl; cerr << it->second.size() << endl; while (lIt != it->second.end()) { cerr << "\titerating..." << endl; cerr << "wc: " << lIt->wordCount << endl; cerr << "cs: " << lIt->commandStatus << endl; cerr << "\t" << lIt->commandStatus << ", WC: " << lIt->wordCount << endl; ++lIt; } ++it; } cerr << endl; } bool addMessage(unsigned key, const TestData& data) { TestStore::SharedMemoryMap::iterator it = mData.find(key); cerr << "Adding word count " << data.wordCount << endl; if(it != mData.end()) { it->second.push_back(data); } else { cerr << "Looking for " << mSegmentName.c_str() << endl; managed_shared_memory m(open_only, mSegmentName.c_str()); cerr << "found it" << endl; cerr << m.check_sanity() << " - " << m.get_free_memory() << "/" << m.get_size() << endl; cerr << mSegment->check_sanity() << " * " << mSegment->get_free_memory() << "/" << mSegment->get_size() << endl; ///!!!!!!!!!!!!!!!BROKE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! TestStore::SharedMemoryList sList(TestStore::ListAllocator( m.get_segment_manager())); //==========This Works (obviously only for current process)=========== // TestStore::SharedMemoryList sList(TestStore::ListAllocator( // mSegment->get_segment_manager())); sList.push_back(data); cerr << "pushed" << endl; cerr << sList.front().wordCount << endl; mData.insert(TestStore::SharedMemoryMap::value_type(key, sList)); cerr << "added list" << endl; } cerr << "====================Data post add===================" << endl; printData(); return true; } private: ///Segment housing this class. Only valid in the constructing process managed_shared_memory* mSegment; ///Name of the segment SharedMemoryString mSegmentName; SharedMemoryMap mData; }; int main() { std::string segName = "MemSegmentHere"; int num_bytes = 645360; SharedMemoryDeleter deleter(segName); managed_shared_memory shm(open_or_create, segName.c_str(), num_bytes); std::string storeName = "TestStoreHere"; shm.destroy<TestStore>(storeName.c_str()); TestStore* store = shm.construct<TestStore>(storeName.c_str())(boost::ref(shm), boost::ref(segName)); TestStore::TestData data; data.commandStatus = 10; data.wordCount = 14; store->addMessage(3, data); return 0; } -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Ion Gaztañaga Sent: Thursday, March 01, 2012 3:24 PM To: Boost User List Subject: EXTERNAL: Re: [Boost-users] Interprocess container segmentation fault (example provided) El 01/03/2012 19:01, Davidson, Josh escribió: > I'm running into issues with using interprocess containers have other containers added to them during runtime. A very high level description is I have a set of data structures in shared memory that can be accessed and modified by more than one process. I'm creating the data structures in the first process and storing the name of the segment within the data structure. Methods that modify the data by adding another container first lookup the segment using its name, build the new container, and then insert it into the existing container. In the example I'm providing below, I have a map of lists. For some reason, when I add a new list, if I open the segment by name, and allocate the list using the segment manager returned via the lookup, the list isn't initialized correctly and accessing > the list (via printData here) results in a segfault. However, if I bypass the segment lookup and just cache the raw pointer to the segment, it works. It's probably easiest just to look at > the code below. In addMessage, you can see how I'm trying to initialize the list in a manner that will work across multiple processes. This breaks. Commented out directly below it is a hack that will work. I can't see any reason why one works and the other doesn't. The "managed_shared_memory" object returned from the lookup appears to be valid and exactly the same as the pointer I'm caching with the data structure in this example. > > I am currently using 1.49 final, but I've observed similar behavior in other 1.4x releases. You are storing a *local* map that allocates memory in shared memory (that is, the internal pointers of the map are in the stack, whereas the allocated nodes are in shared memory), which is really a bad idea as once you unmap the segment, as the local map's internal pointers point to segmentation-fault friendly memory. I know that "managed_shared_memory shm" outlives "TestStore store" but the problem is that you open the shm again in "managed_shared_memory m" and that maps the shared memory *in another address, different to the address mapped by "managed_shared_memory shm"* (two views of the same shared memory, as you map a file twice). So you end up with a local map pointing to *two different address ranges* (although the share the same underlying shared memory object, just think it's like a file). Then you unmap "managed_shared_memory m" (the second address range disappears) and then when you try to iterate, when the local map tries to read an element through the second address range, you get a segmentation fault. If you put the map itself in shared memory (e.g. use a named construct) you can avoid the problem. And if you have already mapped the shared memory once, avoid wasting OS resources (the OS must synchronize the two address ranges in the same process with the same underlying device/shm/mapped file) and application address space mapping the same shared memory/file several times. Sometimes is useful to open the shared memory segment more than once from the same process (say, when you communicate a DLL with an application, or when trying to implement a intermodule singleton type), but in general, try to avoid this. Best, Ion _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Boost.Interprocess containers question
by Davidson, Josh 02 Mar '12

02 Mar '12

I have an object that I'm storing in shared memory that contains a map data member whose values are lists. The processes that interact with this object don't have direct access to the map data structure. Instead, the object is accessed and modified through its public interface. The issue I am running into is how to handle new key, value insertions. I would prefer to leave the fact that the object is in shared memory an implementation detail, but in order to construct a new list to insert in the map, the list needs to be initialized with an allocator using the segment manger. Unless I pass that segment manager into every method that might add a new list to the map, I'm struggling how to get it working. I can't simply store a pointer to the managed shared memory or segment manager within the class since the address would only be valid in the process that constructed it. The only possible implementations I can think of would be to require passing in the segment manager into all of the methods and then create a simple wrapper class that contains a pointer to this object along with a pointer to the segment manager. The wrapper class would essentially "duplicate" each of the methods in the wrapped class but drop the segement manager parameter since it has it as a member. Processes that "find" this object in shared memory would then be responsible for wrapping it in an object of the wrapper class. The other possible implementation would be to always open the shared memory segment by name before modifying the map. Josh

3 3

Re: [Boost-users] Interprocess container segmentation fault (example provided)
by Davidson, Josh 02 Mar '12

02 Mar '12

> Ion, > Terribly sorry, but I goofed when I cobbled that code together. My actual application constructs "TestStore" from the segment, and not on the stack like what I pasted in before. I'm including the full test code again, although only the main function changed. Since the map itself a member of TestStore and TestStore is being constructed from the segment, I would think that should be roughly equivalent to just creating the map by itself from the segment. However, as you'll see if you run it, it still seg faults when accessing the list iterator. Ok, I had to re-read your message again before it sunk in. While allocating the TestStore on the stack was an obvious issue, I understand now why it still blew up when iterating the list. Also, it looks like I was further complicating things. It appears from my testing that I can store allocator instances in the shared memory segment (e.g. a ListAllocator in TestStore from earlier example) and just use that in addMessage rather than creating a new allocator each time I insert into the map. Thanks, Josh

1 0

[interprocess::offset_ptr] operator>> error in boost_1_46_1
by 乔志强 02 Mar '12

02 Mar '12

#include <boost/interprocess/offset_ptr.hpp> I use boost_1_46_1, boost::interprocess::offset_ptr<int> p; std::cin >> p; // error!! offset_ptr.hpp std::ptrdiff_t get_offset() const { return internal.m_offset; } //!operator>> //!for offset ptr template<class E, class T, class Y> inline std::basic_istream<E, T> & operator>> (std::basic_istream<E, T> & is, offset_ptr<Y> & p) { return is >> p.get_offset(); }

1 0

Is there a self contained vector-like container?
by Frank Birbacher 01 Mar '12

01 Mar '12

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi! Is there in Boost a container that is statically "max sized" and contains its own storage like a std::array, but manages a size and can construct and destruct elements. I guess it would look something like: template<typedef T, size_t N> struct self_contained_vector { size_t max_size() const { return N; } size_t size() const { return size_; } private: size_t size_; aligned_storage<T> data_[N]; //...implement random access container... }; Frank -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.17 (Darwin) Comment: GPGTools - http://gpgtools.org Comment: keyserver x-hkp://pool.sks-keyservers.net iEYEARECAAYFAk9P2GcACgkQhAOUmAZhnmqtdACfa6D7j8THF2UtGvTSSHERbffe i6UAn2d+bW1jREGupp1wPMVafXiKQpOm =mwRW -----END PGP SIGNATURE-----

4 7

Iterating over the output of a member function in std::for_each
by Andrew Hundt 01 Mar '12

01 Mar '12

I have a class with an accessor member function that I want to call and apply the result to a functor using std::for_each. I have a working version below that uses a for loop and for_each, but the for_each version is cryptic and cumbersome. Is there a way I can make the for_each version more concise? #if 0 // for loop version: for(value_vector_type::iterator it = values.begin(); it!=values.end(); it++){ avg(it->getValue()); // I want to put this in a for_each loop } #else // bind version: std::for_each(values.begin(), values.end(), // iterate over all values boost::bind( boost::mem_fn(&average_type::operator()), // attach the averaging functor to the output of the getvalue call &avg, boost::bind( boost::mem_fn(&value_wrapper_type::getValue), // bind the getValue call to each element in values _1 ) ) ); #endif I also have the question posted in stack overflow, with a full implementation: http://stackoverflow.com/questions/9507830/iterating-over-the-output-of-a-m… Cheers! Andrew Hundt

2 3