- Boost-users - lists.boost.org

metaclass must be exactly Boost.Python.class, not a subclass?
by william.r.chesters＠jpmorgan.com 29 Aug '07

29 Aug '07

Hi, I'd like to create a metaclass M which I can use to construct new Python classes C inheriting from a C++ base class B. B's metaclass is of course "Boost.Python.class", so I inherit M from that. It's not quite working for me though, because when I try to pass an instance of one of the C's back to C++, I get a "TypeError: No registered converter was able to extract a C++ pointer to type B from this Python object of type C". It seems that BP is testing for "type(type(obj)) is Boost.Python.class" rather than "isinstance(Boost.Python.class, type(type(obj))". Is this an easy issue to fix or work around? Many thanks, William This communication is for informational purposes only. It is not intended as an offer or solicitation for the purchase or sale of any financial instrument or as an official confirmation of any transaction. All market prices, data and other information are not warranted as to completeness or accuracy and are subject to change without notice. Any comments or statements made herein do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries and affiliates. This transmission may contain information that is privileged, confidential, legally privileged, and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is STRICTLY PROHIBITED. Although this transmission and any attachments are believed to be free of any virus or other defect that might affect any computer system into which it is received and opened, it is the responsibility of the recipient to ensure that it is virus free and no responsibility is accepted by JPMorgan Chase & Co., its subsidiaries and affiliates, as applicable, for any loss or damage arising in any way from its use. If you received this transmission in error, please immediately contact the sender and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. Please refer to http://www.jpmorgan.com/pages/disclosures for disclosures relating to UK legal entities.

1 0

serialization lib, const data, documentation
by Robbie Morrison 29 Aug '07

29 Aug '07

Hello all Regarding the Boost.Serialization library. Archiving to XML requires the use of name-value-pairs for all serialized objects, be they stand-alone or encapsulated. I wanted to confirm the following is correct. In reference to the online and bundled HTML documentation: http://www.boost.org/libs/serialization/doc/serialization.html#const file:///.../boost_1_34_1/libs/serialization/doc/serialization.html#const It would appear that the code block which reads: ar & const_cast<T &>(t); should also mention that, in the case of XML serialization: using boost::serialization::make_nvp; // for convenience ar & make_nvp("t", const_cast<T &>(t)); // cast used on load Moreover, this code could be a candidate for a macro, something like (or would this create its own can of worms?): ar & BOOST_SERIALIZATION_NVP_CONSTDATA(T, t); ar & BOOST_SERIALIZATION_CONSTDATA_NVP(T, t); --- On that note, a new page describing user Boost.Serialization macros and their usage might be helpful. For instance I have created the following list (readers should note that my comments may well be incorrect): BOOST_CLASS_EXPORT_GUID(SubClass, "MySC") // GUID is "globally unique id" BOOST_CLASS_EXPORT(SubClass) // shortcut if class name is acceptable BOOST_CLASS_VERSION(Class, 2) // reset class version, the default is zero BOOST_IS_ABSTRACT(AbstractClass) // only some compilers [which ones ?] BOOST_SERIALIZATION_NVP(data) // shortcut if object name acceptable [1] BOOST_SERIALIZATION_BASE_OBJECT_NVP(Base) // shortcut if object name acceptable [2] BOOST_SERIALIZATION_SPLIT_MEMBER() // 'serialize' to 'save'/'load' BOOST_SERIALIZATION_SHARED_PTR(Class) // depreciated [3] BOOST_CLASS_TRACKING(MyVirtualBaseClass, boost::serialization::track_always) BOOST_SHARED_POINTER_EXPORT(T) // depreciated from Boost version 1.32 BOOST_SHARED_POINTER_EXPORT_GUID(T, K) // depreciated from Boost version 1.32 BOOST_SERIALIZATION_LIBRARY_VERSION // yields a "release" number [4] [1] boost::serialization::make_nvp("data", data); [2] requires #include <boost/serialization/base_object.hpp> [3] more specifically, for types without tracking set, that is using boost::serialization::track_never also under 1.34.1 this generates code only for particular Metrowerks and Borland compilers [4] requires #include <boost/serialization/serialization.hpp> (but doesn't appear to be maintained) I don't mind drafting out such a page. Could it be part of the official web documentation or would I be better to add it to the boost wiki as an informal contribution. many thanks in advance Robbie --- Robbie Morrison PhD student -- policy-oriented energy system simulation Institute for Energy Engineering (IET) Technical University of Berlin (TU-Berlin), Germany [from IMAP client]

1 0

Program_options assertion failure
by Johnathan Bunn 29 Aug '07

29 Aug '07

im having an issue with a project using visual studio pro 2005 if it makes a difference the librarys im using are Qt, pdflib, boost, and some crystal com stuff thorugh activex my problem is that when I build in release my command line args are parsed fine and my application behaves as expected. when I build in debug mode, I get an asertion failure as follows Assertion failed: n == name.size()-2, file libs\program_options\src\options_description.cpp, line 122 the relevent code is as follows // process command line args.. po::options_description desc( "TEST" ); desc.add_options( ) ("help", "produce help message" ); ("hideSplash", "hide splash message" ) ("report", po::value<int>( ), "show requested report and quit" ) ; po::variables_map vm; po::store( po::parse_command_line( argc, argv, desc ), vm ); po::notify( vm ); if( vm.count( "help" ) ){ Log::Info::out( ) << "Help Message requested"; // show help window return true; } there is more blocks the same as the last if block that look for my other messages they are commented out now for testing though I am completely stumped and sought help on the #boost irc channel volodya and others on that channel were stumped as well if anyone has any insight into this issue I thank you in advance for your help :-D -- To follow the Path, Look to the master, Walk with the master, See through the master, Become the master

3 3

Boost Parameter Question.
by Surya Kiran Gullapalli 29 Aug '07

29 Aug '07

Hello all, I'm a newbie to boost parameter library. I had no problem in using BOOST_PARAMETER_MEMBER_FUNCTION as long as the function is non static. what is the macro to use for static member function. I gave BOOST_PARAMTER_MEMBER_FUNCTION ( (static void), ... ) But i was getting warning messages with MSVC 8.0 compiler, with boost-1.34.1 warning C4042: 'unnamed-parameter': has bad storage class. I do not know whether these are harmless or not. Please help. Thanks, Surya

2 1

[boost::format] How to not display 0 integer?
by tomasz2k＠poczta.onet.pl 29 Aug '07

29 Aug '07

Hello group I'm using boost::format heavily to format my output for user. I've run into a problem that I cannot resolve for now. I use this format: int i = 4; cout << format ("(%|1$+2|)") % i; // -> (+4) i = -4; cout << format ("(%|1$+2|)") % i; // -> (-4) That's fine and expected. But my question for the one below: int i = 0; cout << format ("(%|1$+2|)") % i; // -> ( 0) // -> ( ) how? How to avoid printing 0? I want formatter to print values in plus or in minus only. When there's value 0 - display only spaces to width. My problem is that I cannot create wrapper over that integer, I must feed formatter with int for various reasons. Does anyone know how to achive this? Maybe I override some parts of library or parsing method? Maybe some of you added your own formatts to formatter and want to share? Any hint or example would be valuable. Thanks in advance, Tomasz Kalkosiñski

1 0

warning C4180 in VS 2005 for bind.hpp
by Adam Badura 29 Aug '07

29 Aug '07

I have a large project at work. We use Visual Studio 2005 with SP 1 and Boost 1.34.0. When we use boost::bind in algorithms on collections we very often get following warning (at level 3): J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored the detailed message is following (just an example extract from compilation log, sadly I was unable to produce a simple stand-alone example - but if it is required then I will give some more work into finding one): J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1609) : see reference to class template instantiation 'boost::_bi::add_cref<Pm,I>' being compiled with [ Pm=boost::shared_ptr<nle::views::ScrollEventListener> , I=1 ] j:\sources\Common/Helpers/Observable.h(128) : see reference to class template instantiation 'boost::_bi::dm_result<Pm,A1>' being compiled with [ Pm=boost::shared_ptr<nle::views::ScrollEventListener> (__thiscall boost::weak_ptr<nle::views::ScrollEventListener>::* )(void), A1=boost::arg<1> ] j:\sources\Common/Helpers/Observable.h(120) : while compiling class template member function 'void common::helpers::Observable<T>::registerListener(const boost::weak_ptr<T> &) const' with [ T=nle::views::ScrollEventListener ] j:\sources\NLE/Views/ScrollInfo.h(65) : see reference to class template instantiation 'common::helpers::Observable<T>' being compiled with [ T=nle::views::ScrollEventListener ] One of my colleagues found a "solution" to this problem. If the modifier "const" is removed in one of typedef's the no warning are generated. However we would like a better solution since modifying the library doesn't seem a good one. What to do? Adam Badura

3 2

Graph: "extra" edge processing during BFS
by Eric Fowler 29 Aug '07

29 Aug '07

I have an algorithm that just screams breadth first search. After some thrashing around in the graph library I am confident in about 90% of the implementation. However there is one little snag: When I examine a grey vertex, I want to do some basic logic on the vertex and the "candidate" next edge to see if what I want to do with this edge and this vertex is "legitimate", according to my own definition of legitimacy(*). If the edge is good, I use it, perhaps relaxing it, if not, I just pass on it and go on to the next one. I would like to do this without tweaking or rewriting the BFS algo. Can this be done with just visitors and the various template parameters or must I write my own BFS variant (not that hard to do of course)? Eric * My idea of "legitimate" is to check to see if a path to a given vertex already contains a certain vertex. The path is built as I run the algorithm. "Parent" vertices are stored as the algorithm runs so I can follow a chain up from the examined vertex back to the root. So the basic logic is: Vertex u = Q.top(); Q.pop(); vis.examine_vertex(u, g); //this is already in BFS if(has_certain_property(u, g)) add_to_tree(u, g); else do_not_add(u, g); .... etc.

1 0

Different functions based on type of iterator?
by HighCommander4 29 Aug '07

29 Aug '07

Let's say I had a function template <typename Iterator> void function(Iterator first, Iterator last); and I want it to call two different functions based on if Iterator is a random access iterator or just a forward iterator. Essentially, I want two versions of this function: template <typename RandomAccessIterator> void function(RandomAccessIterator first, RandomAccessIterator last); template <typename ForwardIterator> void function(ForwardIterator first, ForwardIterator last); I want a function call to function(a, b) to translate to the first one if a and b are random access iterators, and the second one if they are not. I'm thinking of somehow combining enable_if and concept checking, I'm just not sure how... -- View this message in context: http://www.nabble.com/Different-functions-based-on-type-of-iterator--tf4345… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

Linking to DLLs on Windows with GCC
by John Pye 29 Aug '07

29 Aug '07

Hi all, I am having trouble linking against Boost DLLs. Even though I think I am using the correct compiler commands, I seem to end up with statically linked binaries (everything works fine, but Dependency Walker shows that I have no DLL dependency on Boost). Specifically, I am trying to link against Boost:Serialization. Because I'm using MinGW, my Boost libraries had to be compiled by me, so I used "bjam --toolset=gcc --layout=system". I then ran "bjam install" with the same options, which resulted in (over 1GB of) files in c:\Boost. Next I build my own program. It compiles against the Boost files using GCC parameters "-IC:\Boost\include" and then links against Boost::Serialisation using "-LC:\Boost\lib" -lboost_serialization-mt So, what can I do to set GCC to use dynamic linking to the DLL in this case, instead of the static linking that it appears to be using at the moment? (I am a little unclear about the purpose of the .a files currently present in the c:\Boost\lib directory. Do they refer in some way to the DLLs or are they self-contained static libraries?) Cheers JP

1 0

[ptr_container]Swap/reorder elements in ptr_vector
by Bill Buklis 28 Aug '07

28 Aug '07

What is the best or recommended way to swap or reorder elements in a ptr_vector? Naturally, I simply want to swap pointers and not copy any elements. Using swap with the "base" iterator works (see below), but it feels like I'm using an undocumented feature or something. Is there a preferred method? Or is this the best approach? typedef boost::ptr_vector<int> VEC; VEC v; v.push_back( new int(1) ); v.push_back( new int(2) ); VEC::iterator a = v.begin(); VEC::iterator b = v.begin() + 1; // Swap the two underlying pointers using the base iterator std::swap( *a.base(), *b.base() ); -- Bill --

1 0