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

shared_ptr and circular dependency
by Christopher Pisz 30 Mar '12

30 Mar '12

What are one's options when the code one is working on contains a circular dependency and one is trying to update it to use a shared_ptr vs a raw pointer? I don't think I can easily get rid of the preexisting circular dependency. The original author has created an inheritance where the base needs the derived. We all know this is wrong, but it would take quite a bit to analyze how it is used and be removed. Am I out of luck? Fictitious Example (best minimal example including the problems the preexisting code has): // File MasterClient.h #include "BaseClient.h" #include <set> class MasterClient : BaseClient { public: // Constructor adds this to the collection of clients // It is later included in the processing of all clients MasterClient(); // SNIP private: typedef std::set<Base *> Clients; Clients clients_; // Thread function that iterates over all clients // including this instance and makes calls on them void Process(); }; // file BaseClient.h class MasterClient; class BaseClient { public: // Constructor BaseClient(MasterClient * master); // SNIP interface private: MasterClient * master_; // I really need this to be a shared_ptr // Thread function that makes calls to the MasterClient virtual void Process(); }; ---------------------------- I gave it my best shot and have 2 errors that I can't figure out: C2664: BaseClient::BaseClient(boost::shared_ptr<T>) cannot convert parameter 1 from boost::shared_ptr<T> to boost::shared_ptr<T> C2385: ambiguous access of 'shared_from_this' I have to use enable shared from this to be passing around the shared_ptr from myself, no? I am also worried about using this in the intializer list, is it a problem? I believe both classes cannot inherit from it. I am not sure how to change it up. any suggestions? We are using VC2008, so I cannot use the standard shared_ptr yet, but I believe the boost version is the same... My attempt: /* * BaseClient.h */ #ifndef BASECLIENT_H #define BASECLIENT_H #include <boost/shared_ptr.hpp> #include <boost/enable_shared_from_this.hpp> //------------------------------- class MasterClient; //------------------------------- class BaseClient : public boost::enable_shared_from_this<BaseClient> { public: typedef boost::shared_ptr<BaseClient> SharedPtr; // Enforces that every instance must be a shared_ptr for shared_from_this requirement static BaseClient::SharedPtr Create(boost::shared_ptr<MasterClient> master); protected: BaseClient(boost::shared_ptr<MasterClient> master); virtual ~BaseClient(); boost::shared_ptr<MasterClient> master_; }; #endif /* * BaseClient.cpp */ #include "BaseClient.h" #include "MasterClient.h" //------------------------------- BaseClient::SharedPtr BaseClient::Create(boost::shared_ptr<MasterClient> master) { return BaseClient::SharedPtr(new BaseClient(master)); } //------------------------------- BaseClient::BaseClient(boost::shared_ptr<MasterClient> master) : master_(master) { master->InsertClient(shared_from_this()); } //------------------------------- BaseClient::~BaseClient() { } /* * MasterClient.h */ #ifndef MASTERCLIENT_H #define MASTERCLIENT_H #include "BaseClient.h" #include <boost/shared_ptr.hpp> #include <boost/enable_shared_from_this.hpp> #include <set> //------------------------------- class MasterClient : public BaseClient , public boost::enable_shared_from_this<MasterClient> { public: typedef boost::shared_ptr<MasterClient> SharedPtr; // Enforces that every instance must be a shared_ptr for shared_from_this requirement static MasterClient::SharedPtr Create(); protected: typedef std::set<BaseClient::SharedPtr> Clients; Clients clients_; MasterClient(); ~MasterClient(); }; #endif /* * MasterClient.cpp */ #include "MasterClient.h" //------------------------------- MasterClient::SharedPtr MasterClient::Create() { return MasterClient::SharedPtr(new MasterClient()); } //------------------------------- MasterClient::MasterClient() : BaseClient(MasterClient::shared_from_this()) { } //------------------------------- MasterClient::~MasterClient() { }

3 3

Crash in boost::thread::sleep(xt) (boost 1.42 on Windows 32 bit)
by Arno 30 Mar '12

30 Mar '12

Hi all, we recognize in our tests, that one of our threads seem to crash in boost:thread::sleep(xt). Is there a known issue with the sleep() function under MS Visual Studio 2008, with boost 1.42? It was tested on Windows 7 and Windows XP 32 Bit OS. Our thread crashes also in the debugger, when all exceptions are on, without any notice. But only this one thread goes away, the debugger stop at a breakpoint, which was set at a place, where it was recognized that this thread no longer runs. The error occure, in a multithreaded environment, with at least around 30 active threads. The sleep is outsite any critical section (no mutex and/or no semaphore is active). Any hints are welcome. regards Arno

4 16

[boost.locale] Question about boundary rules
by alex_perry 29 Mar '12

29 Mar '12

Have just been exploring boost.locale which I hadn't used before. However I'm not quite understanding some of the behaviour for boundary rules when segmenting text. Fortunately I can see this behaviour happening in the example code so its probably my misunderstanding and easily corrected. If I compile http://www.boost.org/doc/libs/1_49_0/libs/locale/doc/html/boundary_8cpp-exa… and run it I get :- [...skipped to avoid long quote ] Part [Linux2.6] has number(s) Part [ ] has no word characters Part [and] has letter(s) Part [ ] has no word characters Part [Windows7] has number(s) letter(s) Part [ ] has no word characters [...] However I don't understand why "Linux2.6" is detected as having number(s) but no letters whilst "Windows7" is detected as having both. It doesn't appear to be the decimal point "Linux26" has the same behaviour (whilst "Linux2" is detected as having both). I haven't debugged this just glanced at the code (which seems to be setting these flags based on all the icu ruleBasedBreakIterator getRuleStatusVec()). Thought I would just ask whether I'm misunderstanding something fundemental here before trying to understand what is going on here and where (if there is one) the problem is TIA Alex Perry ps Just in case this is a known platform / version issue I was running this on :- Windows7 MSVC 10 boost 1.49 icu 4.9.1 -- View this message in context: http://boost.2283326.n4.nabble.com/boost-locale-Question-about-boundary-rul… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

boost 1.46.1 vs. boost 1.49.0 ?
by JIA Pei 28 Mar '12

28 Mar '12

Hi, all: My code is as simple as: > path tempp(dir_path, native ); It used to work for boost 1.42. And now, it's working fine with boost 1.49.0 ... However, current Ubuntu 11.10 affords boost 1.46.1. Whenever I tried to compile my code with boost 1.46.1, I got the following errors: invalid conversion from ‘bool (*)(const string&) {aka bool (*)(const > std::basic_string<char>&)}’ to ‘boost::enable_if_c<true, void>::type* {aka > void*}’ [-fpermissive] In addition, I notice there are some other errors at the same time, which seems to have something to do with filesystem3 ? Invalid arguments ' > Candidates are: > bool is_directory(boost::filesystem3::file_status) > bool is_directory(const boost::filesystem3::path &) > bool is_directory(const boost::filesystem3::path &, > boost::system::error_code &) > ' However, why my code is able to compile with boost 1.49.0 then? Cheers Pei -- Pei JIA Email: jp4work(a)gmail.com cell: +1 604-362-5816 Welcome to Vision Open http://www.visionopen.com

1 2

Boost interprocess message_queue thread-safe?
by Rajalakshmi Iyer 28 Mar '12

28 Mar '12

Hello, I am trying to use the Boost Inter-process message queue and wanted to know the application needs to take care of synchronization. The documentation lists 'Message queues' under the 'Synchronization' section and mentions: 'A message queue is similar to a list of messages. Threads can put messages in the queue and they can also remove messages from the queue' So is the message queue thread-safe? My use case is several multi-threaded processes on a single host writing to the message queue and another multi-threaded process on the same host reading from the message queue. Thanks! -- Twitter: @Blismobile <http://twitter.com/#!/blismobile> BlisMobile Media 32 Percy Street, London W1T 2DE www.blismobile.com [image: BlisMobile] <http://www.blismobile.com/>[image: Follow on Twitter]<http://twitter.com/#!/blismobile>[image: Blis Website] <http://www.blismobile.com/> This communication is from BlisMobile Media, which is a trading name of Breeze Tech (UK) Ltd, a company registered in England and Wales with registered number 06455773. Its registered office is 32 Percy Street, London W1T 2DE, United Kingdom. This communication contains information that is confidential and may also be privileged. It is for the exclusive use of the intended recipient(s). If you are not the intended recipient(s), please (1) notify info(a)blismobile.com by forwarding this email and delete all copies from your system and (2) note that disclosure, distribution, copying or use of this communication is strictly prohibited. Email communications cannot be guaranteed to be secure or free from error or viruses. All emails sent to or from a Blismobile email account are securely archived and stored by an external supplier. This email does not constitute a contractual agreement; such agreements are in specified contractual or Insertion Order (IO) form only and exclusively contain all the terms to which Breeze Tech )UK) Ltd will be bound. To the extent permitted by law, Breeze Tech (UK) Ltd does not accept any liability for use of or reliance on the contents of this email by any person save by the intended recipient(s) to the extent agreed in a contract or Insertion Order. Opinions, conclusions and other information in this email which have not been delivered by way of the business of Breeze Tech (UK) Ltd are neither given nor endorsed by it.

2 1

[code] multi_array serialization
by Andriy Gapon 28 Mar '12

28 Mar '12

I needed to serialize some multi_array objects in my project and came up with the following code. How does it look? It's probably missing serialization of the "quirks" like memory layout, dimension bases and similar. template<typename Archive, typename T, std::size_t N> inline void save(Archive & ar, const boost::multi_array<T, N> & t, const unsigned int file_version) { ar << boost::serialization::make_nvp("dimensions", boost::serialization::make_array(t.shape(), N)); ar << boost::serialization::make_nvp("data", boost::serialization::make_array(t.data(), t.num_elements())); } template<typename Archive, typename T, std::size_t N> inline void load(Archive & ar, boost::multi_array<T, N> & t, const unsigned int file_version) { typedef typename boost::multi_array<T, N>::size_type size_type; boost::array<size_type, N> dimensions; ar >> boost::serialization::make_nvp("dimensions", boost::serialization::make_array(dimensions.c_array(), N)); t.resize(dimensions); ar >> boost::serialization::make_nvp("data", boost::serialization::make_array(t.data(), t.num_elements())); } template<typename Archive, typename T, std::size_t N> inline void serialize(Archive & ar, boost::multi_array<T, N>& t, const unsigned int file_version) { split_free(ar, t, file_version); } -- Andriy Gapon

1 0

[Bind] Understanding protect
by Robert Jones 28 Mar '12

28 Mar '12

Hi All I'd appreciate some helpful words about Boost.Bind protect, 'cos I'm not quite understanding it! I thought its purpose was allow binding, but not evaluating, of a function object. So that this should work. #include <iostream> #include <boost/bind.hpp> #include <boost/bind/protect.hpp> void f( int i ) { std::cout << i << std::endl; } template <typename NullaryCallable> void g( const NullaryCallable & callable ) { callable( ); } int main( ) { g( protect( bind( f, _1 ) )( 1 ) ); } But it doesn't, and instead thinks I'm calling g() with the return type of f(), just as if I'd not used protect at all. Is it possible to have bind generate a fully bound, nullary callable object? TIA, - Rob.

4 12

[shared_ptr] Multiple enable_shared_from_this in inheritance tree
by Edward Diener 28 Mar '12

28 Mar '12

There was a Code Project article (http://www.codeproject.com/Articles/286304/Solution-for-multiple-enable_sha…) about a problem using multiple enable_shared_from_this derivations in an inheritance tree. Does this problem actually exist in the current Boost implementation ? If so, should there not be some documentation regarding it ?

2 5

[random] number of seed values required to seed engine
by Thomas Mang 27 Mar '12

27 Mar '12

Hi, Consider the mersenne_twister_engine<...> member function seeding a engine with an iterator range of seed values: template<typename It> void mersenne_twister_engine<...>::seed(It &, It); Is there some nice way to figure out how many seed values this function will need? I have used mt19937 and had to inspect the code to figure out it needs 624. I would find it very useful if this information is somewhere readily provided through some organized interface, e.g. a member function or a static class member stating that value. Same for other engine types. BTW, I find it surprising that the first parameter is a reference to an iterator, and hence the input argument itself will be altered. Is that by design? If so what's the rationale? I think it's fairly odd behaviour, giving rise to unnecessary dangerousness and should become documented more clearly (big warning in bold flashing red letters etc. :) ). Thanks Thomas

2 1

[ptr_container] Disable indirection interface
by Philipp Moeller 27 Mar '12

27 Mar '12

I have an API that expects iterators to pointers and using it with ptr_container would simplify some code. But the indirection interface of ptr_container prevents this. I can of course add an iterator adapter on top of the ptr_vector but that seems like overkill to me. -- Philipp Moeller GeometryFactory

2 1