December 2008 - Boost-users

proto for array expressions
by Daniel Oberhoff 01 Jan '09

01 Jan '09

Hi, I have skimmed over the new proto library a few times now. I am kind of excited about a clean implementation of ET. I have used blitz++ in a project before, but am disappointed, because it breaks unexpectedly in places, where it seems the et engine is just not smart enough, even though the expression makes a lot of sense. I ended up having to handcode a lot of things, avoiding which was the very reason I ever used blitz. So now I am back at hand coing in a new projects, using nothing but a few good iterators. Now with proto I was wondering how hard it would be to build an et engine for array math. Basically I am interested in general tensor notation, and the resulting expressions should be iteratable efficiently. think (advanced example): a_ij = b_i * (c_j - sum(d_ijk, k = [i-1,1]) so there should be math on the indices possible, and I should be able to iterate the expression to avoid direct index lookups wich are expensive with arrays, where indices must be multiplied by strides to get the memory address. other goodies would be optional range checks and the ability to pass subexpressions to template functions. also interesting might be transformations to automatically optimize expressions. a nice start would be maybe something simpler like just a_ij = b_i * (c_j - sum(d_ijk, k) where k is summed fully (over the whole range in d_ijk). can anymone give me an estimate on how hard this would be and where to start? Best Daniel Oberhoff Fraunhofer FIT St. Augustin Birlinghoven

4 6

[Serialization] [Serialization of stl map with pointer to POD type]
by Denis Gabaidulin 31 Dec '08

31 Dec '08

I have following class with stl container: class storage : { private: map<const char*, int, cmp> data_; friend class boost::serialization::access; template<class Archive> void save(Archive & ar, const unsigned int version) const { map<string, int> local; constDataIter iter; for (iter = data_.begin(); iter != data_.end(); iter++) { local[iter->first] = iter->second; } ar & local; } } This code works, but it slowly(full copy) and consume too much memory (x2) :-( What is the best serialization strategy with such stl containers or pointer to pod type? I cannot change const char* to string, dont ask me why :-)) Thanks you for any advice.

2 9

Idea: Chained comparison
by Michiel Helvensteijn 31 Dec '08

31 Dec '08

I just came up with this handy class. I was wondering if it might have a place in Boost somewhere. Perhaps it already exists, and I didn't find it? It creates a syntax for chained comparison expressions. Ideally you'd want to be able to do the following in C++: if (0 <= i < N <= INT_MAX) f(); But this doesn't work as expected, because it would actually be evaluated like this: if (((0 <= i) < N) <= INT_MAX) f(); You would actually need to use the following code: if (0 <= i && i < N && N <= INT_MAX) f(); The alternative I propose is this: if (c(0) <= i < N <= INT_MAX) f(); It has the advantage of being shorter, of being closer to the usual mathematical notation, and of evaluating i and N only once. This is useful if they have side-effects. It requires the following class and function (which I'm sure could be improved). What do you think? --------------------------------------------------------------------- template <class T> class Comparison { public: Comparison(T value, bool truth = true) : _last(value), _truth(truth) {} template<class U> Comparison operator==(U val) { return Comparison(val, _truth && _last == val); } template<class U> Comparison operator!=(U val) { return Comparison(val, _truth && _last != val); } template<class U> Comparison operator<(U val) { return Comparison(val, _truth && _last < val); } template<class U> Comparison operator>(U val) { return Comparison(val, _truth && _last > val); } template<class U> Comparison operator<=(U val) { return Comparison(val, _truth && _last <= val); } template<class U> Comparison operator>=(U val) { return Comparison(val, _truth && _last >= val); } operator bool() { return _truth; } private: T _last; bool _truth; }; template <class T> Comparison<T> c(T val) { return Comparison<T>(val); } --------------------------------------------------------------------- -- Michiel Helvensteijn

2 1

Re: [Boost-users] shared_ptr and signals
by Frank Mori Hess 31 Dec '08

31 Dec '08

On Tue Dec 30 12:47 , Steven Watanabe <watanabesj(a)gmail.com> sent: >James C. Sutherland wrote: >> Is it possible to use a shared_ptr object in a signal? >> >> I want something like >> >> struct A{ >> ... >> void operator()(){ ... } >> ... >> }; >> typedef boost::signal Signal; >> Signal mySignal; >> >> boost::shared_ptr a( new A() ); >> mySignal.connect(a); >Wrap the shared_ptr in a function object > >struct deref_call { > void operator()() const { > (*f)(); > } > boost::shared_ptr f; >}; bind can also be used to do this, since shared_ptr can be passed as the "this" argument when binding member functions: boost::shared_ptr a( new A() ); mySignal.connect(boost::bind(&A::operator(), a));

1 0

Re: [Boost-users] [boost] [random] A much faster Poisson variate generation
by Max 31 Dec '08

31 Dec '08

I'm not an expert in the field. IMO, all advances in theory and/or techniques are welcome and appreciated. B/Rgds Max ----- Original Message ----- From: Ross Levine To: boost(a)lists.boost.org Subject: [boost] [random] A much faster Poisson variate generation Date: 2008-12-31 03:42:23 Hello, I have modified the poisson_distribution class to use an updated version of Ahrens and Dieter's PD algorithm for generating Poisson variates, as described in "Computer Generation of Poisson Deviates from Modified Normal Distributions." Is this allowed in Boost, and is there interest? The current algorithm is O(mean), while PD is bounded time. In addition, the current implementation fails for mean >> 750. PD does, however, use a lookup table and thus has a non-trivial constructor and needs 36 * sizeof(RealType) + 1 * sizeof(IntType) bytes per poisson_distribution object. Please let me know if this is desired. Ross ------------------------------------------------------------------- 新浪空间——与朋友开心分享网络新生活！(http://space.sina.com.cn/ )

1 0

[Format] Asserts with non-default structure packing in 1.37 (MSVC)
by Daniel James 31 Dec '08

31 Dec '08

I'm using MS Visual Studio 2005 to work on some legacy code that has its origins back in 16-bit Windows, and relies on the /Zp2 switch to align fields in structures in the same way as earlier versions. One of the more recent modules uses boost::format, and this has compiled correctly and been working with Boost 1.34.1. I'm trying to bring the project up-to-date using 1.37, and I am getting BOOST_STATIC_ASSERT reports when I include boost/format.hpp. I don't have to *use* format in my code, just including boost/format.hpp in a "Hello world" program is sufficient to trigger the error when using /Zp2 Is this to be expected? Is there a workaround? I know we should really use explicit pack pragmas around the structures that have to be packed in this particular way, but there are too many of them -- and they're defined all over the code -- for this to be a simple or safe process ... /Zp2 has always worked fine in the past. Cheers, Daniel (Just a user -- not to be confused with the Boost contributor of the same name)

2 2

Re: [Boost-users] [lambda] Should the code compile?
by Max 31 Dec '08

31 Dec '08

Hello Watanabe, It seems to be a problem of my mail editor. I have used the correct static_cast grammar but some characters were silently eaten. :-( Sorry for misleading. At last, I've found the problem and the working code is as follows: #define DECLARE_PTR_AUTO_DELETOR_CONDITIONAL(CLASS, VAR_PTR, CONDITION)\ boost::scoped_ptr<CLASS> auto__deletor__for__ ## VAR_PTR (\ (boost::lambda::if_then_else_return(boost::lambda::_1, VAR_PTR,\ static_cast<CLASS*>(0))) (CONDITION)) I should use if_then_else_return instead of if_then_else. Thanks for help. B/Rgds Max ----- Original Message ----- From: Steven Watanabe <watanabesj(a)gmail.com> To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] [lambda] Should the code compile? Date: 2008-12-31 00:51:01 AMDG Max wrote: > I have the following simple code, but doesn't get compiled: > > #define DECLARE_PTR_AUTO_DELETOR_CONDITIONAL(CLASS, VAR_PTR, CONDITION)\ > boost::scoped_ptr auto__deletor__for__ ## VAR_PTR (\ > boost::lambda::ret(boost::lambda::if_then_else(_1,\ > VAR_PTR, static_cast(0)))(static_cast(CONDITION))) > That's not the correct syntax for static_cast In Christ, Steven Watanabe ------------------------------------------------------------------- 新浪空间——与朋友开心分享网络新生活！(http://space.sina.com.cn/ )

1 0

Passing signals as functions
by Remko Tronçon 30 Dec '08

30 Dec '08

Hi, I have a signal boost::signal<void ()> foo that I would like to pass to a function taking a function<void ()>. Currently, I pass the pointer to a helper function that calls foo(), but I suspect there's an easier way. Same goes for the more advanced case, where i have a signal boost::signal<void (Bar)> foo which i would like to the same function (taking a function<void ()>), after binding the signal parameter to a specific value. Does anyone know how to do this as well? thanks, Remko

2 3

[unit test] enable_if in MSVC
by n e 30 Dec '08

30 Dec '08

Hello, I am using static link to boost.test in my project. Using that gives about 50 errors in MSVC compiler. I tried excluding enable_if header with defining that before unit_test inclusion and that seems to work. Is there any better ways to handle that? By the way I use version 1.35 .lib file linked to project. Thanks in advance.

2 2

Re: [Boost-users] [Serialization]Can not export polymorphic types
by Daniel Roberts 30 Dec '08

30 Dec '08

That works. Some things I notice: 1) The xml file created has an attribute in the object tag called "class_name", containing the name of the derived type. This attribute does not exist in an xml file created using derived type pointers. I guess this is the critical piece of information needed to reconstruct a derived type using a base class pointer. 2) It also works if you create a derived class pointer/object, static_cast the derived type pointer to a base class pointer and pass the base class pointer to the serialization. 3) someInXMLArchive >> boost::serialization::make_nvp("", someTypePtr); works, which is surprising since I am passing an empty string for a tag. It appears serializing in ignores tag names and just processes data until the class is constructed or a stream error occurs. 4) Calling close on the output file stream (e.g. std::ofstream::close) causes the closing boost_serialization tag, </boost_serialization> to NOT be written to the file. Is this an issue that needs to be resolved or am I doing something else wrong? Thanks a bunch for all the help. Hopefully I'll be running on my own now. Please feel free to comment on my observations above, however. Regards, Daniel Roberts What happens if you try the following: TestBase * pT1 = new TestType1; TestBase* pT2 = new TestType2; testOutXML << boost::serialization::make_nvp("pT1", pT1); testOutXML << boost::serialization::make_nvp("pT2", pT2); testTypeOFS.close(); delete pT1; delete pT2; ... testInXML >> boost::serialization::make_nvp("pT1", pTB); testInXML >> boost::serialization::make_nvp("pT2", pTB); Robert Ramey

2 1