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

[range] unique
by Bill Buklis 23 Feb '12

23 Feb '12

This may simply be my lack of understanding of the range library as I'm still just working my way through the algorithms, but is boost::unique returning the correct result range? Given that std::unique returns an iterator to the first non-unique value you can do something like: vec.erase( std::unique(vec.begin(), vec.end(), vec.end() ); However, boost::unique returns the range of the "unique" values. I would expect instead that it would return the range of non-unique values so that it matched the standard algorithm version, which effectively returns the range of non-unique values. Correspondingly trying: boost::erase( vec, boost::unique(vec) ); eliminates the complete opposite of what the standard version does. I can get around this by calling: vec.erase( boost::unique(vec).end(), vec.end() ); but this seems rather counter-intuitive to me. Is this correct? Is this by intent? What is the benefit of having it this way? The documentation doesn't mention anything at all about the return range in the description. Since I'm still just beginning to use the range library, I don't know all the tricks and benefits yet, so I ask. -- Bill

2 2

Boost.Units: Issue with user-defined and unitless quantities
by Pieter 22 Feb '12

22 Feb '12

LS, This is a repost of an earlier question that unfortunately remained unanswered - I still hope sincerely someone out there can help me out as I feel really stuck here. I have some problems working with user defined quantities/units in combination with unitless quantities (such as the results of a sin() or cos() operations).Taking the example of the 'nautical mile', which is presented in the documentation: ---START CODE --- namespace nautical { struct length_base_unit : boost::units::base_unit<length_base_unit, length_dimension, 1> { static std::string name() { return "nautical mile"; } static std::string symbol() { return "nmi"; } }; typedef boost::units::make_system<length_base_unit>::type system; typedef unit<length_dimension,system> length; static const length mile,miles; } BOOST_UNITS_DEFINE_CONVERSION_FACTOR(nautical::length_base_unit, boost::units::si::meter_base_unit, double, 1.852e3); ---END CODE --- The following does not compile: ---START CODE --- typedef quantity<boost::units::si::plane_angle> TAngleInRad; typedef boost::units::quantity<nautical::length> TNauticalDistanceInNMI; TNauticalDistanceInNMI NauticalDistanceInNMI; TNauticalDistanceInNMI SumNauticalDistanceInNMI = NauticalDistanceInNMI + NauticalDistanceInNMI; // OK TNauticalDistanceInNMI ScaledSumNauticalDistanceInNMI = NauticalDistanceInNMI + NauticalDistanceInNMI*sin(AngleInRad); // Last line does not compile due to lack of a suitable operator+ ---END CODE --- It does work correctly when using the 'nautical mile' as it has been defined in the library itself in "nautical_mile.hpp" (as a scaled base unit related to the si::meter): ---START CODE --- namespace metric { typedef scaled_base_unit<boost::units::si::meter_base_unit, scale<1852, static_rational<1> > > nautical_mile_base_unit; } // definition of the nautical mile from "nautical_mile.hpp" typedef boost::units::metric::nautical_mile_base_unit::unit_type nautical_mile_unit; typedef quantity<nautical_mile_unit> TNauticalDistanceInNMI; ---END CODE --- However, I don't think I can specify my user-defined units/quantities in this manner as they use a scale factor represented by an arbitrary floating point value. My perception is that the example of the documentation does not work due to the mixing of quantities from the si system (the sin() returns a dimensionless quantity in that system) with user-defined quantities in another ('nautical') system, while the 'nautical mile' from 'nautical_mile.hpp' is a scaled version of the si::meter and thus defined within the realms of the si::system - but please correct me if I am wrong here. My problem can thus be summarized as follows: - the solution from the documentation does not compile - the solution from 'nautical_mile.hpp' uses a scaled base unit which is limited to scaling factors of the form scale<Base,Exponent> and my question is therefore: how can I define my user-defined units/quantities in such a way that they allow the required operations while being related to base units by an arbitrary scale factor? Help would be highly appreciated, Pieter ------------------- Nautical mile version #1 (from the included nautical_mile.hpp): namespace metric { typedef scaled_base_unit<boost::units::si::meter_base_unit, scale<1852, static_rational<1> > > nautical_mile_base_unit; } template<> struct base_unit_info<metric::nautical_mile_base_unit> { static const char* name() { return("nautical mile"); } static const char* symbol() { return("nmi"); } }; typedef quantity<boost::units::si::plane_angle> TAngleInRad; typedef quantity<boost::units::si::length> TDistanceInM; typedef boost::units::metric::nautical_mile_base_unit::unit_type nautical_mile_unit; typedef quantity<nautical_mile_unit> TDistanceInNMI; typedef boost::units::quantity<nautical::length> TNauticalDistanceInNMI; # this is OK: TAngleInRad Angle(0.5*radians); TDistanceInNMI DistanceInNMI; TDistanceInNMI SumDistanceInNMI = DistanceInNMI + DistanceInNMI; TDistanceInNMI ScaledSumDistanceInNMI = DistanceInNMI + DistanceInNMI*sin(AngleInRad); TNauticalDistanceInNMI NauticalDistanceInNMI; TNauticalDistanceInNMI SumNauticalDistanceInNMI = NauticalDistanceInNMI + NauticalDistanceInNMI; TNauticalDistanceInNMI ScaledSum NauticalDistanceInNMI = NauticalDistanceInNMI + NauticalDistanceInNMI*sin(AngleInRad); // DOES NOT COMPILE for lack of a suitable operator+

1 0

[Serialization] base64 encoding/decoding an archive as it's written/read
by Sean Farrow 22 Feb '12

22 Feb '12

Hi: Is there any way of base64 encoding/decoding an archive as it is wread/written? I have found some code that encodes, but wondered whether there was already any ioStreams written. Cheers Sean.

2 2

Acknowledgement messages
by norulez＠me.com 22 Feb '12

22 Feb '12

Hi, can anybody explain me on how to stop the "Boost-users post acknowledgement" messages after I submit a message? Thanks in advance Best Regards NoRulez

2 2

Re: [Boost-users] undefined reference to '__sync_add_and_fetch_4'whencross-compiling boost::asio fors mips
by Jayne 22 Feb '12

22 Feb '12

Hello Andreas, Thanks for the help. I have eventually managed! In short I... - defined BOOST_ASIO_DISABLE_THREADS both while building Boost, and in the application which used it - implemented the atomic primitives using code from the gcc headers All working now! thanks, Jayne. On Fri 17/02/12 12:37 , "Andreas Wehrmann" a.wehrmann(a)centersystems.com sent: Alright, I looked into how I did when I faced the problem (my previous reply was a fast shot). I put the following #defines in the user config header (boost/config/user.hpp): #define BOOST_ASIO_DISABLE_FENCED_BLOCK #define BOOST_SP_USE_SPINLOCK This did the trick for me. From what I recall, the problem lies with shared_ptr since the implementation tries to use gcc builtins for the locking mechanism it needs for refcounting but which are not implemented on every platform (in my case it was Coldfire/m68k). Regards -- Dipl.-Ing. (FH) Andreas Wehrmann Software Development -------------------------------------------------------------- Center Communication Systems GmbH A-1210 Wien, Ignaz-Köck-Straße 19 Sitz in Wien FN 796 88p, Firmenbuchgericht Wien http://www.centersystems.com [1]">www.centersystems.com Tel.: +43 (0) 190 199 - 3616 Mobile: +43 (0) 664 884 75916 Fax: +43 (0) 190 199 - 2110 E-Mail: a.wehrmann(a)centersystems.com Links: ------ [1] http://www.centersystems.com

1 0

Re: [Boost-users] mpi/serialization: broadcasting derived class through base pointer
by MM 22 Feb '12

22 Feb '12

> -----Original Message----- > From: Matthias Troyer [mailto:troyer@phys.ethz.ch] > Sent: 16 December 2010 15:20 > To: > Subject: Re: [Boost-users] mpi/serialization: broadcast function and > the value argument > > > On Dec 16, 2010, at 10:54 PM, Hicham Mouline wrote: > > > 1.44 > > As I wrote, I am able to debug deep into boost mpi and archive. > > Please let me know if I should send any more stacks or other data, > > > > Please try with 1.45. Pointer serialization in Boost.MPI has been > broken in 1.44 due to changes in Boost.Serialization. > > Matthias Hello Matthias, Robert Ramey, I am revisiting the topic of mpi broadcast-receiving a polymorphic object through a base class null pointer: void recv_histodata(boost::mpi::communicator& world) { historical_data_base* hd = nullptr; boost::mpi::broadcast(world, hd, master_process); } historical_data_base has 3 derived classes. I get the unregistered class exception from the serialization library, and so I am trying to register the derived type as shown in the serialization doc at /libs/serialization/doc/index.html : reference ar.template register_type<derived_one>(); but where do I put that? Regards, MM

6 19

return std::vector<> as boost::python::list
by Andy Falanga (afalanga) 21 Feb '12

21 Feb '12

Hi everyone, I'd originally worked out this issue by following the guidelines at the python wiki for exposing C++ containers into python. Basically, I made my vector act like a python list. This, however, created a problem for our teams build environment, specifically, that we didn't have python_d.lib to link with. This irritation cannot be tolerated in our environment so, I'm tasked with moving that vector to a boost::python::list object. However, all of my searches haven't returned much in terms of how to use this list object. The vector is of a simple structure that doesn't have anything but primitive types. Because of this, returning the list by value isn't a problem. Could I simply make a wrapper around the function that returns the vector with function that returns the list? For example: std::vector<myStruct>& GetVector() { // important stuff return vector; } boost::python::list GetPythonList(std::vector<myStruct>& structure) { boost::python::list boostPyList; for(int i(0); i < structure.size(); i++) { // copy each element into the python list } return boostPyList; } Would this work? Is it that simple? Thanks, Andy

1 0

[Multi-Index] Design Rationale behind erase in ordered_index container
by Ovanes Markarian 21 Feb '12

21 Feb '12

Hello Joaquín, the doc states in erase that an iterator or end-iterator is returned when erasing an element. What was the rationale behind returning the iterator? Where other members return the number of elements being removed. My point is that end-iterator is also returned when erasing an element where iterator points one before the end iterator. Now to be sure that the specified element was really removed my call must either use the key and search for that element again or compare the size before and after the erase call. Many thanks, Ovanes http://www.boost.org/doc/libs/1_47_0/libs/multi_index/doc/reference/ord_ind… iterator erase(iterator position); *Requires:* position is a valid dereferenceable iterator of the index. *Effects:* Deletes the element pointed to by position. *Returns:* An iterator pointing to the element immediately following the one that was deleted, or end() if no such element exists. *Complexity:* O(D(n)). *Exception safety:* nothrow.

5 10

Using a part of a Jamfile.v2
by Daryle Walker 21 Feb '12

21 Feb '12

I was thinking about tweaking a library in Boost. Its BOOST_ROOT/libs/math/test/ directory is shared with some other libraries, so there's a lot of test files and the shared Jamfile.v2 used to make the tests is huge. Is there a way to restrict which targets are built from the command line? Right now, it takes MANY hours to go through every test. (When I did something similar last week with another library, I just commented out all the other libraries sharing the file. Do I have to resort to that all the time?) Daryle W.

2 1

[hash] Reliability of hash_value() for std::string
by Szymon Gatner 21 Feb '12

21 Feb '12

Hi, I am coding a simple serialization/factory lib and want to use class hash code based on std::string representing class name as a factory identifier. My question is: how portable / reliable is hash value produced by std::string hash_value() overload? Will it be the same across platforms? Across boost versions? Or maybe actual algorithm is dictated by standard? Or should I just go with explicit algo like crc32? btw. I was just very hardly bitten by the fact that hash_value() for const char* != hash_value() for std::string, (it is just being treated as regular pointer). In other words hash_value("Class") != hash_value(std::string("Class")) which gave me few more gray hair. Is that desirable? Cheers, Szymon Gatner

3 4