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

[Smart Pointer] Reference counting when you don't want to manage memory [Honeywell Internal]
by Hickman, Steve (AdvTech) 26 Aug '14

26 Aug '14

Classification: Honeywell Internal Smart pointers (as in http://www.boost.org/doc/libs/1_56_0/libs/smart_ptr/smart_ptr.htm ) assume that the action you want to perform when the count goes to zero is to free the memory pointed to. That isn't the only use for reference counting. Why doesn't the Smart Pointer library decouple the reference counting from the action to be taken and provide a generic reference counting mechanism that will call an arbitrary function when the count goes to zero? The default can be "delete" to provide the expected memory management capability but it would provide a robust mechanism for those who need other capabilities. Reviewing the <boost/shared_ptr.hpp> code, it appears this decoupling already exists - through the use of boost::checked_delete and boost::checked_deleter (<boost/checked_delete.hpp>). It just doesn't appear to be documented. Why not document this so others know how to use it? Steve Hickman System Architect, Flight Deck of the Future 480-236-8367 ============================================================================================================================================================================================================ This message classified as Honeywell Internal by Hickman, Steve (AdvTech) on Tuesday, August 26, 2014 at 12:35:32 PM. The above classification labels are in accordance with the Honeywell Corporate Classification Policy. The information contained in this electronic message is confidential and intended only for the use of the individual/entity named above, and the information may be privileged. If you, the reader of this message, are not the intended recipient or an employee or agent responsible to deliver this message to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please immediately notify the sender and delete the original message. ============================================================================================================================================================================================================

3 3

Multi-Index: custom iterator
by Sensei 26 Aug '14

26 Aug '14

Dear all, I'm not yet satisfied with my MultiIndex games, so now I am trying to do something that is purely cosmetic. However, I have *no idea* how I can do this. How can I build a custom iterator on a MultiIndex? The objective is simple. As you may recall, I am building a MultiIndex with 128 bits, indexing them by MSB, LSB, and HSB (in the middle). It works like a charm. Now I'd like to write something like this (or similar): auto it = storage.myfind<byMSB>(MSB(value128bits)); for (; i != it.end(); i++) In essence, I'd like to hide the get<> and everything I am using right now (as shown below). So I need an iterator for this data structure: typedef __uint128_t Storage; // Indexing placeholders struct byMSB { }; struct byHSB { }; struct byLSB { }; // Storage type with indexing typedef boost::multi_index::multi_index_container< Storage, boost::multi_index::indexed_by< // Hashed by subject/predicate/object boost::multi_index::hashed_non_unique<boost::multi_index::tag<byMSB>, boost::multi_index::global_fun<Storage, soSub, &Data::msb>>, // 0 boost::multi_index::hashed_non_unique<boost::multi_index::tag<byHSB>, boost::multi_index::global_fun<Storage, soPre, &Data::hsb>>, // 1 boost::multi_index::hashed_non_unique<boost::multi_index::tag<byLSB>, boost::multi_index::global_fun<Storage, soObj, &Data::lsb>> // 2 > > DataStorage; DataStorage storage_; With your suggestions I've achieved to find all items by MSB, but sincerely, now I don't know how to write a custom iterator for this type. The code for finding elements is this: //typedef typename DataStorage::template nth_index<0>::type byHashedIndex; //typedef typename DataStorage::template index<byMSB>::type byHashedIndex; //byHashedIndex &p = storage_.template get<byMSB>(); auto &p = storage_.template get<byMSB>(); auto q = p.equal_range(subject(v)); while (q.first != q.second) { std::cout << " EQID " << *(q.first) << " | " << msb(*(q.first)) << " | " << hsb(*(q.first)) << " | " << lsb(*(q.first)) << std::endl; l.push_back(*(q.first)); q.first++; } Note: I *have* written iterators, but they were much easier (for instance char*). This is far too complex for making my way on my own! From the code above, I guess I have begin() and end() point to an equal_range output .first and .second. But what will the "pointer" I need in the iterator will be? Of course, my first attempt is a complete failure: // Iterator template <class T> class iterator : public std::iterator<std::input_iterator_tag, Storage> { typedef typename DataStorage::template index<T>::type byHashedIndex; byHashedIndex &p; // Default constructor iterator() : p(storage_.get<T>()) { }; // Copy constructor iterator(const iterator& i) : p(i.p) { }; // Pre-increment ++it iterator& operator++() { ++p; return *this; }; // Post-increment it++ iterator operator++(int) {iterator tmp(*this); operator++(); return tmp; }; // Equality bool operator==(const iterator& rhs) { return p == rhs.p; }; // Non-equality bool operator!=(const iterator& rhs) { return p != rhs.p; }; // Access Storage& operator*() { return *p; }; // Difference iterator operator-(int i) { return iterator(p - i); }; // Addition iterator operator+(int i) { return iterator(p + i); }; }; Can you point me in the right direction? Thanks!

2 3

"newbie" Build Boost Matrix
by Ismael Fadiga 25 Aug '14

25 Aug '14

hi all After executing the below sql query in c++, the printed results are: WITH rnum AS (SELECT s.ISN, s.Date, ClosedPrice, row_number() OVER (partition BY s.ISN ORDER BY s.ISN, s.Date DESC) AS rownumber FROM SecurityData s INNER JOIN PortfolioHoldings p ON s.ISN = p.ISN WHERE p.Pfid = 1876) SELECT DISTINCT ISN, Date, ClosedPrice FROM rnum WHERE rownumber <= 250)) //ID Date Price US345 2014-08-17 342 US345 2014-08-16 310 ... FR123 2014-08-17 18.20 FR123 2014-08-16 15 Does anyone know how to build a price matrix for the IDs like:? |342 18.20| MAT=|310 15 | |... ... | Your feedback would be very appreciated Cheers

2 5

[ConceptCheck] Same type concept
by Robert Jones 25 Aug '14

25 Aug '14

I'd like a concept check to check for two types being the same type. I can see how to write such a concept, but this is such an obvious requirement that either there should be one already, or there's some blindingly obvious reason why you never need one, a reason that escapes me. A little guidance would be appreciated. Thanks, Rob.

3 3

Re: [Boost-users] [Chrono] The default BOOST_CHRONO_VERSION is still 1
by DonM 24 Aug '14

24 Aug '14

-------- Original Message -------- Subject: Re: [Boost-users] [Chrono] The default BOOST_CHRONO_VERSION is still 1 From: "Vicente J. Botet Escriba" <vicente.botet(a)wanadoo.fr> Date: 8/23/2014 6:28 AM To: boost-users(a)lists.boost.org > > Le 23/08/14 01:41, DonM a écrit : >> The docs (boost_1_56_0/doc/html/chrono/reference.html) say: >> >> "The default value for BOOST_CHRONO_VERSION will be changed to 2 since Boost 1.55. " >> >> However, it still has the value 1 in both 1.55 and 1.56. >> >> Not a big deal, I just noticed and figured I'd mention it. >> > Hi, > > I must remove this comment as as there is a incompatibility this will break user code. I planned to do a smooth trarnsition, but there were some complains. > > Thanks for pointing this issue, > Vicente Are you saying that you never intend to change the default value? That would be a mistake, in my opinion. Version 2 is much better -- the I/O stuff works. If users need backwards compatibility, they can define version 1, but encourage people to use version 2. Just my $0.02. Don

2 1

[move] Movable but not copyable bug?
by Mostafa 24 Aug '14

24 Aug '14

Given the following code, if (1) is uncommented then it fails to compile under C++03 mode yet works in C++11 mode. If commented out, then it compiles successfully under both modes. (g++ 4.8.2 is being used) Is this behavior expected or a bug? #include <boost/move/core.hpp> #include <boost/move/utility.hpp> struct Foo { Foo(int x, int y) {} Foo(BOOST_RV_REF(Foo) rhs) {} private: //Purposefully declared and not defined. Foo & operator=(Foo); private: BOOST_MOVABLE_BUT_NOT_COPYABLE(Foo) }; template <typename T> struct Bar { Bar() : f(0, 1) {} // (1) // Bar(Bar const & rhs) // : f(rhs.f) // {} Bar(BOOST_RV_REF(Bar) rhs) : f(::boost::move(rhs.f)) {} private: //Purposefully declared and not defined. Bar & operator=(Bar); private: BOOST_COPYABLE_AND_MOVABLE(Bar) private: T f; }; Foo f1() { // (2) // Foo f(0, 1); // return f; // (3) return Foo(0, 1); } Bar<Foo> f2() { return Bar<Foo>(); } int main() { // Foo f( f1() ); Bar<Foo> b(Bar<Foo>()); return 0; }

2 2

BGL: boost::astar_search_no_init
by Georg Gast 24 Aug '14

24 Aug '14

Hi there! Currently i try to implement an astar search with BGL and boost. My graph is infinite, so i searched for an example of astar_search_no_init and found this: http://lists.boost.org/boost-users/2012/03/73707.php I am on gcc-4.9.0 and boost 1.52.0 on linux. I try to compile the above example and get the following error which i cant decipher ... make gcc -std=c++11 example.cpp -o example example.cpp: In function 'int main(int, char**)': example.cpp:274:140: error: no matching function for call to 'boost::associative_property_map<default_map<std::pair<XY, XY>, unsigned int> >::associative_property_map(default_map<std::pair<XY, XY>, unsigned int>)' . weight_map(boost::associative_property_map<default_map<std::pair<XY,XY>,unsigned> >(default_map<std::pair<XY,XY>,unsigned>(1))) I added the code as an attachement. For me it seems that this line is not ok .... but i currently have no clue how to fix it ... try { astar_search_no_init(g, start, distance_heuristic<XYGraph>(goal) , visitor(astar_goal_visitor(goal)) . distance_map(d) . predecessor_map(boost::ref(p)) . weight_map(boost::associative_property_map<default_map<std::pair<XY,XY>,unsigned> >(default_map<std::pair<XY,XY>,unsigned>(1)) // < --- line 274 . vertex_index_map(boost::associative_property_map<std::map<XY,unsigned> >(std::map<XY,unsigned>())) . rank_map(boost::associative_property_map<std::map<XY,unsigned> >(std::map<XY,unsigned>())) . color_map(boost::associative_property_map<std::map<XY,boost::default_color_type> >(std::map<XY,boost::default_color_type>())) . distance_compare(std::less<unsigned>()) . distance_combine(std::plus<unsigned>()) ); with clang 3.4 i get this error line. clang --std=c++11 example.cpp -o ex example.cpp:274:26: error: no matching conversion for functional-style cast from 'default_map<std::pair<XY, XY>, unsigned int>' to 'boost::associative_property_map<default_map<std::pair<XY, XY>, unsigned int> >' . weight_map(boost::associative_property_map<default_map<std::pair<XY,XY>,unsigned> >(default_map<std::pair<XY,XY>,unsigned>(1)) So what ant the author of this code do? Allocate a property map of size 1? Currently i dont see the forrest, there are so many trees in my view .... Any help would be appreciated! Thank you! Georg

1 0

[Chrono] The default BOOST_CHRONO_VERSION is still 1
by DonM 23 Aug '14

23 Aug '14

The docs (boost_1_56_0/doc/html/chrono/reference.html) say: "The default value for BOOST_CHRONO_VERSION will be changed to 2 since Boost 1.55. " However, it still has the value 1 in both 1.55 and 1.56. Not a big deal, I just noticed and figured I'd mention it. Don

2 1

Boost.ASIO calling handler guarantees
by Фёдор Трушкин 23 Aug '14

23 Aug '14

Hello, Dear All, I'm playing with Boost.ASIO framework and failed to find a complete answer on the following question. Under which circumstances is it guaranteed that asynchronous operation completion handler is called? Plainly speaking, is this example code correct? void my_handler(const error_code &ec) { if(!ec) { /* process successful operation */ } else { /* process failed operation */ } try { async_write(port, my_handler2) } catch(...) { // processing failure to execute asynchronous operation } } Obviously here I rely heavily upon the above mentioned guarantee. Looking at least at async_write() implementation, it cannot be relied upon: - async_write executes a chain of boost::asio::detail::write_op() operations: - if asio_handler_allocate() may throw then - write_op::stream_.async_write_some() may throw then - write_op.handler_() might never be called So, either we should never rely on this assumption, or there's a bug in async_write(). Not relying on this assumption what is the reliable way of building applications with Boost.Asio? Sincerely yours, Fedor Trushkin

3 3

Installing 1.56 on Fedora 20
by Robert Jones 22 Aug '14

22 Aug '14

It's been a while since I've installed Boost. Running this command on a completely vanilla Fedora 20 b2 --prefix=~/local/boost-1.56.0 --layout=tagged link=shared threading=multi,single install Finally says... ...failed updating 58 targets... ...skipped 12 targets... ...updated 12214 targets... Is this normal? Thx, Rob.

3 5