December 2011 - Boost-users - lists.preview.boost.org

[Units] How do you declare a plane angle to be dimensionless?
by Stephen Torri 06 Dec '11

06 Dec '11

Boost ver: 1.45 Compiler: Visual Studio 2008 I am trying to figure out how to declare a plane angle to be dimensionless. In the code I have so far I am using: typedef boost::units::make_system<radii_base_unit, minute_base_unit, boost::units::angle::radian_base_unit >::type myproject_system_t; I have one velocity in radii^1.5/min and a second velocity in radian/min. When I divide the radii^1.5/min by radian/min I am expecting to see radii to be the final dimension. What I am seeing is the denominator has radians in it. Here is my equation ke = radii^1.5/min n = radian/min a1 = ( ke / n ) ^ (2/3) The 'min' dimension cancel out so I am left with radii^1.5/radian before applying the power of 2/3. Applying the power of 2 I see the units are radii^3/radian^2 and then the 1/3 to get a final of radii/radian^(2/3). According to the almighty Wikipedia the radian angle is a dimensionless term. I am wondering what is the recommended course of action I should be taking here? Stephen

2 1

[function] How to use function with templated member functions
by Christian Henning 06 Dec '11

06 Dec '11

Hi there, I would like to boost::function with a templated member function. Please consider the following code: struct my_reader { my_reader(bool full) { if( full ) { _read_function = &my_reader::read_full<int>; } else { _read_function = &my_reader::read_partial<int>; } } void read() { _read_function(this); } private: template< typename T > void read_full() { std::cout << "reading full image" << std::endl; } template< typename T > void read_partial() { std::cout << "reading partial image" << std::endl; } boost::function<void (my_reader*)> _read_function; }; int main(int argc, char* argv[]) { my_reader r(false); r.read(); return 0; } This doesn't compile with Visual Studio 2010 since the compiler cannot: "cannot convert from 'overloaded-function' to 'boost::function<Signature>'" Is there is way to make this happen? Regards, Christian

3 3

Using boost::asio::io_service with boost::packaged_task
by Kelvin Chung 06 Dec '11

06 Dec '11

I'm trying to create a thread pool / task queue like system via boost::asio::io_service and boost::packaged_task rather than attempt to make 800+ threads for 800+ tasks that I have to make. (I'm not sure of the various concurrency issues that I might have, so reducing thread creation overhead is probably one of my big refactorings in my project.) A cursory glance from various sources reveals that I could do something like the following: template <class Ret> class TaskWrapper { // Wrapper around a boost::shared_ptr<boost::packaged_task<Ret>>, as boost::asio::io_service::post() // requires copyable input boost::shared_ptr<boost::packaged_task<Ret>> task; boost::unique_future<Ret> future; public: TaskWrapper(boost::shared_ptr<boost::packaged_task<Ret>>& task_) : task_(task) {} void operator()() { (*task)(); } boost::unique_future<Ret> get_future() { return future; } }; class TaskQueue { boost::asio::io_service io_service; boost::asio::io_service work; boost::thread_group threads; void run() { io_service.run(); } public: TaskQueue(std::size_t size = boost::thread::hardware_concurrency()) : work(io_service) { for (std::size_t i = 0; i < size; ++i) { // Create threads in pool threads.create_thread(boost::bind(&TaskQueue::run, this)); } } ~TaskQueue() { io_service.stop(); threads.join_all(); } template<class Ret> boost::unique_future<Ret> schedule(TaskWrapper<Ret>& task) { io_service.post(boost::bind(&TaskWrapper<Ret>::operator(), boost::ref(task))); return task.get_future(); } }; There are a few questions that I'd like to ask with this: * Would you recommend this? * To retrieve the results, I'd still have to extract all the futures (from TaskQueue::schedule() returns) into a list of boost::shared_futures, and then call boost::wait_for_all() on the list, right? * If a TaskWrapper itself posts other tasks, would a priority inversion or resource starvation issue occur (I keep thinking that it will, for some reason). How should I address this?

1 0

[BGL] clear_vertex in subgraph
by Avishay Ben Shabtai 05 Dec '11

05 Dec '11

> > > According to Ticket #4753 <https://svn.boost.org/trac/boost/ticket/4753> clear > vertex should work on subgraph BUT according to documentation clear_vertex > should remove all edges of specified vertex but in the boost 1.47 > implementation of clear_vertex of subgarph it removes only the out_edges() > is there a prbolem with documentation or the issue was not fixed? > > The code of subgraph is: > > template <typename G>void > clear_vertex(typename subgraph<G>::vertex_descriptor v, subgraph<G>& g) { > while (true) { > typedef typename subgraph<G>::out_edge_iterator oei_type; > std::pair<oei_type, oei_type> p = out_edges(v, g); > if (p.first == p.second) break; > remove_edge(*p.first, g); > } > } > > Thanks > > Avishay > >

2 1

lexical_cast<uint_value_t<10>::least>("7") = 55
by Nathan Ridge 05 Dec '11

05 Dec '11

Whoops! uint_value_t<10>::least is unsigned char... Can something be done about this, like having uint_value_t<10>::least be something that looks and acts like an unsigned char but is recognized by lexical_cast to be an integer type, not a character type?

2 2

Container with insertion order
by Claude 05 Dec '11

05 Dec '11

Is available in BOOST some container (set or map) that can keep the insertion order of elements? -- View this message in context: http://boost.2283326.n4.nabble.com/Container-with-insertion-order-tp4097717… Sent from the Boost - Users mailing list archive at Nabble.com.

8 35

[Bimap][Function] Create a Boost.bimap of SDLKeys and Boost.functions
by John M 05 Dec '11

05 Dec '11

What is the correct syntax for creating a boost::bimap<SDLKey, boost::function<void(boost::shared_ptr<Player>, bool)> >? (SDLKey is part of the SDL library. It is an enumeration. See here http://www.libsdl.org/cgi/docwiki.cgi/SDLKey) The code above does not compile, generating long cryptic template-related errors. Using an std::map<SDLKey, boost::function<void(boost::shared_ptr<Player>, bool)> > works, however. I could use some help overcoming this obstacle.

4 4

[Units] Unable to raise constant to power of two
by Stephen Torri 05 Dec '11

05 Dec '11

In the attached code I am trying to take a constant physical unit and raise it to a power of two. The compiler is complaining that it cannot find a template to satisfy multiply_typeof_helper. It says that there are more than one partial specialization that matches the template argument list. I read something on the web about there was no pow operation for constants. The solution to that problem was to upgrade to Boost 1.48. Unforunately I am constrained to using Boost 1.45. What is an work around for this problem? Here is the compiler output with the file attached: 1>------ Build started: Project: boost_units, Configuration: Debug Win32 ------ 1>Compiling... 1>boost_units.cpp 1>C:\boost\boost/units/detail/static_rational_power.hpp(126) : error C2752: 'boost::units::multiply_typeof_helper<X,Y>' : more than one partial specialization matches the template argument list 1> with 1> [ 1> X=boost::units::quantity<myproject::second_types::squared_radii_unit>, 1> Y=boost::units::one 1> ] 1> C:\boost\boost/units/quantity.hpp(689): could be 'boost::units::multiply_typeof_helper<boost::units::quantity<Unit1,T1>,Y>' 1> C:\boost\boost/units/detail/one.hpp(36): or 'boost::units::multiply_typeof_helper<T,boost::units::one>' 1> C:\boost\boost/units/detail/static_rational_power.hpp(93) : see reference to class template instantiation 'boost::units::detail::static_int_power_impl<1,false>::apply<Y,R>' being compiled 1> with 1> [ 1> Y=boost::units::quantity<myproject::second_types::squared_radii_unit>, 1> R=boost::units::one 1> ] 1> C:\boost\boost/units/detail/static_rational_power.hpp(158) : see reference to class template instantiation 'boost::units::detail::static_int_power_impl<N>::apply<Y,R>' being compiled 1> with 1> [ 1> N=2, 1> Y=boost::units::constant<boost::units::physical_constant<myproject::second_types::equatorial_radius_of_earth_t>>, 1> R=boost::units::one 1> ] 1> C:\boost\boost/units/detail/static_rational_power.hpp(187) : see reference to class template instantiation 'boost::units::detail::static_int_power_sign_impl<N>::apply<Y>' being compiled 1> with 1> [ 1> N=2, 1> Y=boost::units::constant<boost::units::physical_constant<myproject::second_types::equatorial_radius_of_earth_t>> 1> ] 1> C:\boost\boost/units/pow.hpp(52) : see reference to class template instantiation 'boost::units::detail::static_rational_power_impl<R,Y>' being compiled 1> with 1> [ 1> R=boost::units::static_rational<2>, 1> Y=boost::units::constant<boost::units::physical_constant<myproject::second_types::equatorial_radius_of_earth_t>> 1> ] 1> ..\..\..\apps\boost_units\boost_units.cpp(104) : see reference to class template instantiation 'boost::units::power_typeof_helper<X,Y>' being compiled 1> with 1> [ 1> X=boost::units::constant<boost::units::physical_constant<myproject::second_types::equatorial_radius_of_earth_t>>, 1> Y=boost::units::static_rational<2> 1> ] 1>Build log was saved at "file://c:\documents and settings\stephen.torri\my documents\visual studio 2008\projects\satsim_boost_units\vs2008\apps\boost_units\boost_units.dir\Debug\BuildLog.htm" 1>boost_units - 1 error(s), 0 warning(s) ========== Build: 0 succeeded, 1 failed, 1 up-to-date, 0 skipped ========== Regards, Stephen

3 3

[python] converting C++ object to PyObject
by Johan Råde 05 Dec '11

05 Dec '11

I have three C++ classes X, Y and Z that I have exported from C++ to Python using python::class_. Now I want to write a factory function like PyObject* f(char c) { if(c == 'x') return new X; else if(c == 'y') return new Y; else if(c == 'z') return new Z; else PyErr_SetString(PyExc_ValueError, "Bad function argument"); } and export that function to Python. What is missing here is code that converts X*, Y* and Z* to PyObject*. How do I write that? Thank you, Johan

4 7

[units] EE Acceleration, Velocity, Displacement in terms of decibels
by Michael Powell 04 Dec '11

04 Dec '11

Boost Units, Has anyone done any work, conversions, etc, with Acceleration, Velocity, Displacement (Length) in terms of decibels? dB, dBA, dBV, etc, and conversions to/from more "conventional" units like meters, meters per second, meters per second per second? I am learning, for example, that it is possibly to integrate across the Time dimension in order to yield Velocity from Acceleration, or similarly, Displacement (Length) from Velocity. I am doing a job with these type conversions going on across a spectrum analysis and would be interested in hooking up the calculations using boost::units to help enforce dimensional / unit integrity. Thanks... Best regards, Michael

2 2

BOOST_THROW_EXCEPTION causing Abort Trap
by Kieran Simpson 04 Dec '11

04 Dec '11

I'm trying to use `boost::exception` and have condensed my prototype code down to the example in the Boost exception tutorial (http://www.boost.org/doc/libs/1_47_0/libs/exception/doc/tutorial_transporti…) however when running the code with the `BOOST_THROW_EXCEPTION` macro I get an abort from the program. #include <iostream> #include <boost/exception/all.hpp> typedef boost::error_info<struct tag_my_info,int> my_info; struct my_error: virtual boost::exception, virtual std::exception { }; void f() { BOOST_THROW_EXCEPTION(my_error() << my_info(42)); // Uncomment the below (and comment the above) for the program to work //throw my_error() << my_info(42); } int main(int argc, char** argv) { try { f(); } catch(my_error& x) { if(int const* mi = boost::get_error_info<my_info>(x)) { std::cout << "My info: " << *mi << std::endl; } } return 0; } Running the code with the BOOST_THROW_EXCEPTION macro: $ ./a.out Abort trap If as the comment says, I swap the code, all is well $ ./a.out My info: 42 Below is the output from the g++ preprocessor for f() void f() { ::boost::exception_detail::throw_exception_(my_error() << my_info(42),__PRETTY_FUNCTION__,"main.cpp",14); } Software versions are: $ g++ -v Using built-in specs. Target: x86_64-apple-darwin10 Thread model: posix gcc version 4.4.6 (GCC) $ port list boost boost @1.47.0 devel/boost I'm on OSX SL using the tools provided by MacPorts. I've double checked the g++ search paths and there's only one copy of the boost hpp files and that's the ones that belong to the aforementioned boost package. I have no idea why the abort trap is being called. I admit I'm newish to C++. I've also asked this question on StackOverflow (http://stackoverflow.com/questions/8256874/boost-throw-exception-causing-ab…) Cheers,

1 1

Embedded python using boost
by Shane Kruk 03 Dec '11

03 Dec '11

I'm in the process of learning python, and embedding python into C++. I know very little about python so I've been playing around with many different types of interfaces and API's, with varying degrees of successes and failures. What I'm currently working on is the boost interface with C++. I am, however, stuck, for lack of a better word. Since the system I'm playing on is at work and on an isolated network, I do not have the ability to post the code directly to this forum. As such, I was able to print a copy of some experimental code, brought it home, and scanned it into a PDF file. I've attached the file(s) containing the code. If someone could please take a look at it and provide some advice as to what my problems are, it would be greatly appreciated. You'll notice some lines commented out as I've been experimenting with different things in an attempt to get something to work. My main problem at the moment is that the functions import(), eval(), and exec() cannot be found by the compiler. I get compiler errors indicating they are not declared in scope. I'm running on a redhat linux system (5.4) with a 1.33.1 version of boost. Python is version 2.4. I forgot to check the GNU C++ compiler version, but I'm guessing that probably doesn't matter.

1 1

boost::posix_time::time_period - how to round seconds?
by qplace 03 Dec '11

03 Dec '11

I am building intervals using 0.5h increments (1800sec) in a cycle. The interval length is 1800sec also 8:30 am is used as start date. last of my interval looks like this: [2009-Apr-07 14:30:00/2009-Apr-07 14:59:59.999999] because of this weired endpoint my data with the timestamp of Apr-07 15:00:0000 does not fit into the interval How can I fix the interval to be [2009-Apr-07 14:30:00/2009-Apr-07 15:00:000000] ? Thanks

2 1

[thread] Reducing the number of open tickets.
by Vicente J. Botet Escriba 03 Dec '11

03 Dec '11

Hi, I have committed the following tickets in trunk. Fixed Bugs: * [@http://svn.boost.org/trac/boost/ticket/2639 #2639] documentation should be extended(defer_lock, try_to_lock, ...). * [@http://svn.boost.org/trac/boost/ticket/3762 #3762] Thread can't be compiled with winscw (Codewarrior by Nokia). * [@http://svn.boost.org/trac/boost/ticket/3885 #3885] document about mix usage of boost.thread and native thread api. * [@http://svn.boost.org/trac/boost/ticket/3975 #3975] Incorrect precondition for promise::set_wait_callback(). * [@http://svn.boost.org/trac/boost/ticket/4480 #4480] OpenVMS patches for compiler issues workarounds. * [@http://svn.boost.org/trac/boost/ticket/4819 #4819] boost.thread's documentation misprints. * [@http://svn.boost.org/trac/boost/ticket/5423 #5423] thread issues with C++0x. * [@http://svn.boost.org/trac/boost/ticket/5739 #5739] set-but-not-used warnings with gcc-4.6 * [@http://svn.boost.org/trac/boost/ticket/6168 #6168] recursive_mutex is using wrong config symbol (possible typo). * [@http://svn.boost.org/trac/boost/ticket/6200 #6200] patch to have condition_variable and mutex error better handle EINTR. These are tickets with a simple solution or that had associated a complete patch or a clear solution. Anthony I would prefer that you do the merge so that you can check what goes on the release version. I have not closed the tickets. These tickets have just release 1.49 as Milestone. Please, close them when you merge to release. There were 113 tickets two days ago. There are now 114 tickets associated to Boost.Thread. I have closed some duplicates, but some news tickets have been created since. I would like to reduce by 10% the number of tickets every month. I encourage the Boost users that have created tickets for Boost to provide a clear an complete patch when possible so that we can reduce together the number of tickets associated to this Boost library and make it more portable/usable. Don't forget to provide tests that make evident the bug, or to add doc when you are proposing a anew feature. There are a lot of tickets associated to move semantics emulation. I have created a ticket to adapt Boost.Thread to Boost.Move hoping most of these bugs will be solved. As posted in another thread there is already a patch using Boost.Move that could be a good starting point. Of course the help offered by Ion will be too much appreciated. I have created another to adapt the interface to Boost.Chrono, but I will prefer to take care of this once the one for Boost.Move adaptation is committed, as the chance of collisions is big. On the mean time I will continue with some tickets specific to Windows platform. Best, Vicente

1 0

[GIL] Saving DPI in PNG files
by PB 02 Dec '11

02 Dec '11

Hello, Is it possible to state a DPI setting when saving a PNG file using Boost' Generic Image Library? Thanks, Pete

3 4

[serialization] Warning "type qualifier is meaningless on cast type" in extended_type_info_typeid.hpp
by Patrik Jonsson 02 Dec '11

02 Dec '11

Hi, When I compile code that uses the 1.48.0 serialization library with icpc 12.3.174, I get this warning, along with screenfuls of instantiation context: /n/home00/pjonsson/include/boost/serialization/extended_type_info_typeid.hpp(134): warning #191: type qualifier is meaningless on cast type static_cast<T const * const>(p) ^ This is the offending code: virtual void destroy(void const * const p) const { boost::serialization::access::destroy( static_cast<T const * const>(p) ); The compiler is complaining (correctly, I think) that you can't dictate whether something returned by value (here, the T const *) is const or not. Changing the cast to: static_cast<T const *>(p) makes the warning go away. Perhaps someone can make this change? Regards, /Patrik J.

1 0

C0x11 threading capabilities
by Erotavlas_turbo＠libero.it 02 Dec '11

02 Dec '11

Hi, I have some question about the threading programming capabilities of boost library. First question: The new C++ standard C++11 has a support for threading programming. I have read some times ago, that with high probability the threading library of the C++11 should be the same of boost library? If it is true, can I use the boost thread library and in the future, when the C++ standard will be completed, could I transparently switching to it? What are the differencies between the boost thread library and the future C++11 thread library? Second question: this question is strictly related to the first one. At the moment GCC supports only a small subset of threading programming features http: //wiki.apache.org/stdcxx/C++0xCompilerSupport. So using thread boost library allows me to use the full features of future C++ standard library? Or I'm limited by the compiler? Last question: what about the OpenMP directives? How they are different in terms power and flexibility for the parallel programming task respect to the boost library? Thank you Best Regards

3 2

Changelog for 1.47.0 does not properly warn about critical bug
by Paul Harris 02 Dec '11

02 Dec '11

Hello, I discovered this bug by accident https://svn.boost.org/trac/boost/changeset/68866 which was only linked via the ticket mentioned below, The code below demonstrates that ANY wide-character string of the same length resolves to the same UUID ! Thats pretty darn serious, I expected to see a much more serious mention in the Changelog in boost 1.47.0... for UUID, there was only a mention of a few tickets including this one: https://svn.boost.org/trac/boost/ticket/5145 which only mentioned a WARNING in the description ! "Fixed warning: use of logical && with constant operand; switch to bitwise & or remove constant [-Wconstant-logical-operand]" Surely it should say something like "UUIDs for wide character strings are totally wrong" Note also that the Changelog link for UUID goes to Utility, not UUID. The code: #include <boost/uuid/name_generator.hpp> #include <boost/uuid/nil_generator.hpp> #include <boost/uuid/uuid_io.hpp> #include <iostream> using namespace boost::uuids; using namespace std; int main() { { cout << "WIDE" << endl; const wchar_t* a = L"one long string that is something like this and that blahh"; const wchar_t* b = L"hello there you long long string that looks nothing like t"; nil_generator nil; name_generator na(nil()); name_generator nb(nil()); uuid ua = na(a); uuid ub = nb(b); cout << to_string(ua) << endl; cout << to_string(ub) << endl; } { cout << "NARROW" << endl; const char* a = "one long string that is something like this and that blahh"; const char* b = "hello there you long long string that looks nothing like t"; nil_generator nil; name_generator na(nil()); name_generator nb(nil()); uuid ua = na(a); uuid ub = nb(b); cout << to_string(ua) << endl; cout << to_string(ub) << endl; } return 0; } the result: $ g++ main.cpp && ./a.out WIDE 4ffa48af-9685-5089-97ac-fe6627ead94c 4ffa48af-9685-5089-97ac-fe6627ead94c NARROW 73030b72-a864-5c7b-9963-c0bedff1283e da8943a6-26c9-5641-bd37-427b1b70d619 $ The two different wchar_t strings resolve to the same uuid !!! What am I doing wrong? thanks Paul

3 3

Problem with user defined Boost Units system
by Stephen Torri 01 Dec '11

01 Dec '11

I am using Boost 1.45 and trying to use Boost Units to map the types of astrodynamic equation variables. I am getting an error with the following equation. The error is saying that the lvalue type for k2_divded_by_semi_major is wrong. The questions I have are: Q1: Did I defined my system correctly? (System file is attached - types.hpp) Q2: How do I define ratio in the system? For example I have some constants that define a conversion (1 earth radii = 6378 km). Q3: How do I define a variable type if its a fraction (e.g. radii / radian )? ---- EQUATION ---- sgp4::types::radii_t k2_divided_by_semi_major = m_k2 / pow<2> ( semi_major ); The types of variables are: m_k2 sgp4::types::radii_t semi_major sgp4::types::radian_t ---- ERROR ---- 1> [ 1> Unit=boost::units::unit<boost::units::list<boost::units::dim<boost::units::length_base_dimension,boost::units::static_rational<-1>>,boost::units::detail::merge_dimensions_impl<0,0>::apply<boost::units::dimensionless_type,boost::units::dimensionless_type>::type>,boost::units::homogeneous_system<boost::units::list<sgp4::types::radii_base_unit,boost::units::list<sgp4::types::radian_base_unit,boost::units::dimensionless_type>>>>, 1> Y=double 1> ] 1> and 1> [ 1> Unit=sgp4::types::sgp4_length_t, 1> Y=double 1> ] 1> Constructor for class 'boost::units::quantity<Unit,Y>' is declared 'explicit' 1> with 1> [ 1> Unit=sgp4::types::sgp4_length_t, 1> Y=double 1> ] Regards, Stephen

3 8

Multicast using Asio
by Vince Deters 01 Dec '11

01 Dec '11

Hi, I'm working on a rudimentary pub/sub mechanism using Asio, with the requirement that the publisher be notified when a new subscriber appears. Using multicast, this would correspond to a subscriber joining the multicast group. But it appears there's no way for the publisher to know this, correct? If so, is the only alternative to dedicate a socket to each pub/sub pair? Thanks, Vince

2 1

[Units] Temperature conversion problem
by Belcourt, Kenneth 01 Dec '11

01 Dec '11

Hi, There appears to be a temperature conversion problem between ºC and ºF. Here's a simple test that shows the error under both gcc and Intel on both Darwin and Linux (with Boost trunk). brisc:~ kbelco$ cat test3.cpp #include <iostream> #include <boost/units/io.hpp> #include <boost/units/quantity.hpp> #include <boost/units/systems/temperature/celsius.hpp> #include <boost/units/systems/temperature/fahrenheit.hpp> int main(int, char **) { using namespace boost::units; quantity<fahrenheit::temperature> t(100 * celsius::temperature()); // expect output of 212 F, produces 180 F // seems like not adding 32º offset when computing F = (C * 9/5 + 32) std::cout << t << std::endl; return 0; } brisc:~ kbelco$ icpc -g -Iboost test3.cpp brisc:~ kbelco$ ./a.out 180 F -- Noel

2 5

Re: [Boost-users] [Units] Temperature conversion problem
by Alex Perry 01 Dec '11

01 Dec '11

On Wed, 30 Nov 2011 15:00:28 -0800 Matthias Schabel <boost(a)schabel-family.org> wrote: ...snip... > This will have the added benefit of allowing conversions but disallowing > operations that don't make sense... Some code: > > #include <iostream> > > #include <boost/units/absolute.hpp> > #include <boost/units/systems/temperature/celsius.hpp> > #include <boost/units/systems/temperature/fahrenheit.hpp> > #include <boost/units/systems/si/temperature.hpp> > > #include <boost/units/base_units/temperature/conversions.hpp> > > using namespace boost::units; > > int main() > { > quantity<celsius::temperature> > dTC(10.*celsius::temperature()); > quantity<fahrenheit::temperature> > dTF(10.*fahrenheit::temperature()); > quantity<si::temperature> dTK(10.*si::kelvin); > > std::cout << dTC << std::endl; > std::cout << dTF << std::endl; > std::cout << dTK << std::endl; > > quantity<absolute<celsius::temperature> > > TC(0*absolute<celsius::temperature>()); > quantity<absolute<fahrenheit::temperature> > > TF(32.*absolute<fahrenheit::temperature>()); > quantity<absolute<si::temperature> > > TK(0.*absolute<si::temperature>()); > > std::cout << TC << std::endl; > std::cout << TF << std::endl; > std::cout << TK << std::endl; > > std::cout << quantity<absolute<celsius::temperature> >(TF) << > std::endl; > std::cout << quantity<absolute<celsius::temperature> >(TK) << > std::endl; > > std::cout << quantity<absolute<fahrenheit::temperature> >(TC) << > std::endl; > std::cout << quantity<absolute<fahrenheit::temperature> >(TK) << > std::endl; > > // std::cout << TC+quantity<absolute<celsius::temperature> >(TF) << > std::endl; // fails because you cannot add two absolute temperatures > std::cout << TC+quantity<celsius::temperature>(dTF) << std::endl; > std::cout << TF+quantity<fahrenheit::temperature>(dTC) << > std::endl; > > return 0; > } Not being a current user of boost units I was curious about above code and just thought I'd try it (it looked totally straightforward and I was thinking hmm maybe I should be using boost units at times) - However it didn't compile (msvc 2010 and boost 1.48) - It wasn't exactly hard to change it so it would compile - just played around a little and ended up adding :- #include <boost/units/quantity.hpp> #include <boost/units/io.hpp> However given your email had outputs pasted in so obviously code was running at some point I wondered : Was this just a cut & paste typo? Or do you need to be careful with headers in code using boost unit to make sure that it compiles x-platform or some such? Alex

2 1

[phoenix] problem when trying to access member via arg1
by Christian Hoffmann 01 Dec '11

01 Dec '11

Hi, I have below test program in which I try to call a function on each map object. It compiles and runs fine with gcc 4.5 but not on vc10. Is arg1 and ->* operator supposed to work when arg1 is an object and not pointer? Or is there another way of accessing members with arg1 (i.e. without phoenix::bind)? Regards, Chris ----------------------- #include <boost/phoenix/core.hpp> #include <boost/phoenix/operator.hpp> #include <iostream> #include <vector> #include <boost/range/algorithm.hpp> #include <boost/shared_ptr.hpp> #include <boost/shared_ptr.hpp> struct A { void f (int s) const { std::cout << "Calling f (" << s << ")...\n"; } }; int main (int argc, const char **argv) { typedef std::map<int, boost::shared_ptr<A> > map_type; typedef map_type::value_type value_type; map_type a_map; a_map[3] = boost::shared_ptr<A> (new A ()); namespace phoenix = boost::phoenix; using phoenix::arg_names::arg1; boost::for_each (a_map, ((arg1->*&value_type::second)->*&A::f) (10)); } ---------------------------

1 0

Subversion repository not accessible as anonymous
by Marko Lindqvist 01 Dec '11

01 Dec '11

http://www.boost.org/users/download/#repository says that "Anonymous, read-only access to the Boost Subversion repository is available at http://svn.boost.org/svn/boost." My attempts to make checkout results in svn prompting for credentials. Seems like misconfiguration to me. - ML

2 2

Using hash_range with polymorphism
by Kelvin Chung 01 Dec '11

01 Dec '11

I'm having a hard time reconciling two distinct notions. Suppose I have a class hierarchy with root class Foo, and subclasses Foo1, Foo2, etc. (all of which are leaf classes) Now, suppose that hash_value() is defined for all the subclasses but not Foo itself (say, if Foo was abstract). Now suppose I have this: struct FooList { std::vector<boost::shared_ptr<Foo>> foos; } std::size_t hash_value(const FooList& list) { std::size_t seed = 0; typedef boost::indirect_iterator<std::vector<boost::shared_ptr<Foo>>::const_iterator> Iterator; boost::hash_range(seed, Iterator(list.foos.begin()), Iterator(list.foos.end()); return seed; } This seems a bit incorrect (and it would fail to compile due to "ambiguous overload" of hash_combine or something). What I want to do is something along the lines of std::size_t hash_value(const FooList& list) { std::size_t seed = 0; for (std::vector<boost::shared_ptr<Foo>>::const_iterator it = list.foos.begin(); it != list.foos.end(); ++it) { boost::hash_combine(seed, **it); } return seed; } That is, hash the Foos pointed to by the foos (using the correct derived class hash_value() functions) and not the shared_ptrs themselves. Is there a "shorter" version of the latter (which I presume is correct), or will the latter have to do?

3 4