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

[chrono] Can not build with intel-win
by Bo Jensen 13 Jun '11

13 Jun '11

Newest intel C++ release and vs2010, can not build with chrono from trunk. I get : pnet.cpp ..\..\..\external\boost\boost/chrono/duration.hpp(94): error: expected an expression && (boost::is_convertible<Rep2, ^ ..\..\..\external\boost\boost/chrono/duration.hpp(124): error: expected an expression && (boost::is_convertible<Rep, ^ compilation aborted for ..\network\pnet.cpp (code 2) call "C:\Program Files (x86)\Intel\ComposerXE-2011\bin\intel64//iclvars.bat" > nul icl @"bin\intel-win\release\threading-multi\pnet.obj.rsp" ...failed compile-c-c++ bin\intel-win\release\threading-multi\pnet.obj... ...failed updating 1 target... Any suggestions welcome. Thanks /Bo

2 1

[Smart Pointers] enable_shared_from_this::shared_from_this() behaviour query
by eddyaod＠gmail.com 13 Jun '11

13 Jun '11

shared_from_this() throws bad_weak_ptr if it's called in a derived destructor even if the object is heap allocated, does anyone know if this is intended? And if so, why? Example: class DestructorCheck : public std::tr1::enable_shared_from_this <DestructorCheck> { public: DestructorCheck() { } ~DestructorCheck() { shared_from_this(); } }; BOOST_AUTO_TEST_CASE(enable_weakTests) { BOOST_CHECK_THROW( { DestructorCheck d; } , std::tr1::bad_weak_ptr ) BOOST_CHECK_NO_THROW( { std::tr1::shared_ptr<DestructorCheck> d = boost::make_shared<DestructorCheck>(); } ) }

2 2

intrusive_ptr and multi-threading
by Andrej van der Zee 13 Jun '11

13 Jun '11

Hi, I am looking for an efficient way to use boost::intrusive_ptr in a multi-threaded application. In the source-tree I encountered the macro BOOST_INTERLOCKED_INCREMENT/DECREMENT. Should I use these too and how are these implemented in Linux? Where can I find documentation on these macros? Thank you, Andrej

2 4

Boost.Phoenix Exception Handling
by Stuart Dootson 13 Jun '11

13 Jun '11

I've been using Phoenix (the one that came with 1.44 - that's Phoenix v2?) for a bit, as a Boost.Lambda replacement, mostly happily. But I've come across a scenario I don't think Phoenix caters for, namely accessing an exception from a catch_ body, something like this: std::vector<std::string> errors; using namespace boost::phoenix; using namespace boost::phoenix::arg_names; std::for_each(changes.begin(), changes.end(), try_ [ bind(&SomeClass::AddChange, this, arg1, std::string()) ] .catch_<std::runtime_error>() [ push_back(ref(errors), bind(&std::runtime_error::what, /* Can I access the exceptin that's been thrown? */)) ]); I want to aggregate any error information thrown by SomeClass::AddChange in the 'errors' vector, to be rethrown after the for_each. In Boost.Lambda, I'd be able to use the '_e' placeholder to access the exception object, but there seems to be no way of doing so in Phoenix? So, I'm just wondering if a) I've missed something, and b) is there any likelihood of this being added into Phoenix v3 at some point? Stuart Dootson

2 3

Boost configuration - particularly no wchar
by Pace, Antony [Harman Pro Group UK] 13 Jun '11

13 Jun '11

Hi, Our compiler does not support wchar. I have found that #defining BOOST_NO_STD_WSTRING and BOOST_NO_INTRINSIC_WCHAR_T helps the compilation but I suspect there is a correct place or method of defining this. Could someone point me at the documentation for this please? Many thanks Ant

2 2

exception, filesystem, system_error and consistency
by Michael Schuerig 13 Jun '11

13 Jun '11

I haven't used C++ for several years and it's only a few weeks since I've looked at boost for the first time. Thus it's likely that my understanding of things isn't quite up to scratch. In my first C++ project for years, I'm already using boost::filesystem and I'm using boost::system::system_error to wrap error codes returned by Linux API functions. For the latter purpose, I intend to use boost::exception for its ability to add diagnostic information to exceptions. At the top-level of my own code, I implement plain C interfaces and therefore need to convert all exceptions back to their original error codes. Poring over docs and headers, I noticed that neither boost::system::system_error nor boost::filesystem::filesystem_error are derived from boost::exception. Instead they implement an independent exception hierarchy. I found this disconcerting. For one thing, I have to catch and handle the respective exceptions separately. Far more importantly, though, I was expecting consistency among boost libraries and now see that this is not the case apparently. When there is more than one way to do something, this automatically raises the question which one is the right way. Regarding boost, I'm wondering whether all libraries are created equal or whether there is an undocumented gradation from "stable, recommended" to "experimental, beware". Michael -- Michael Schuerig mailto:michael@schuerig.de http://www.schuerig.de/michael/

4 10

[Serialization] Boost serialization how to speed it up?
by tang qingxiong 13 Jun '11

13 Jun '11

I am currently trying to use boost serialization in my code and is encountering some performance issues mainly with speed of serialization and de-serialization. I am coding such that it is able to choose to serialize in both binary and XML format. My binary size is currently about 180+MB which takes a while to de-serialize Below is a sample of my code for the serialization of a struct. I have quite a few of them but they are all similarly implemented namespace boost { namespace serialization { template<class Archive> void serialize( Archive& theArchive, sim::SpeedStruct& theValue, const unsigned int theVersion ) { theArchive & BOOST_SERIALIZATION_NVP(theValue.XValue); theArchive & BOOST_SERIALIZATION_NVP(theValue.YValue); theArchive & BOOST_SERIALIZATION_NVP(theValue.ZValue); } } } BOOST_CLASS_VERSION( SpeedStruct, 1 ); BOOST_CLASS_EXPORT_GUID( SpeedStruct, "SpeedStruct" ); and the speed struct is just a simple struct with 3 double values of X,Y and Z. MY question is I read the FAQs and there is indication of various ways to speed up the process, but I am unclear of how to implement it in my code , specifically how to use BOOST_IS_BITWISE_SERIALIZABLE and also removing tracking and versioning. Thanks! Lester

1 0

how to change boost thread priority ?
by chyujian 12 Jun '11

12 Jun '11

I am new using boost thread, in my project I need to change the priority of the threads, but I cannot find help in the documentation.

2 1

Last of the first batch of Boostcon videos
by Marshall Clow 12 Jun '11

12 Jun '11

Gordon (Thursday): http://blip.tv/boostcon/introducing-mpl-graph-5262812 Bryce (Friday): http://blip.tv/boostcon/ast-construction-with-the-universal-tree-5266608 I am looking into why Bartoz' Haskell preso is not showing up. You can download the mp4 directly here: http://blip.tv/file/get/Boostcon-HaskellThePseudocodeLanguageForCTemplateMe… but the conversion to H.264 seems to time out (I've tried several times) I'll be contacting blip.tv about this in the morning. -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

1 0

Re: [Boost-users] [boost] Even more Boostcon videos
by Marshall Clow 12 Jun '11

12 Jun '11

On Jun 11, 2011, at 9:29 AM, John Bytheway wrote: > On 10/06/11 16:56, Marshall Clow wrote: >> Joel (Thursday): >> http://blip.tv/boostcon/practical-simd-acceleration-with-boost-simd-5261117 > > This one doesn't entirely work for me; the audio cuts off at about > 1:10:43. Fails at the same point in mplayer and VLC. I downloaded it > twice with the same result. > > $ md5sum < Boostcon-PracticalSIMDAccelerationWithBoostSIMD525.mp4 > 56c08b18dc5141a383736e1456d9c48b - I went back and checked the tapes - the audio is missing there, too It appears that the battery on the wireless mic that we were using ran out. :-( Thanks for the heads-up, but I don't think there's anything that I can do to fix this. -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

1 0