February 2009 - Boost-users - lists.preview.boost.org

[Proto] implementing an computer algebra system with proto
by Kim Kuen Tang 09 Feb '09

09 Feb '09

Hi all, considering the following expression created with proto: "3 *var_+5=2; .. The tree would be : = / \ + 2 / \ * 5 / \ 3 var_ Here the term "var_" is somewhat a placeholder. I wonder wheter it is possible now to change the structure of the tree. That is i want to define a grammar that is able to produce another expression like this. "var_=(2-5)/3;" In tree: = / \ var_ div / \ - 3 / \ 2 5 With this expression i can now define a context that return the final result of the right hand side. But is this possible in proto with the existing functions? This situation is somewhat different from the examples in the documentation. Here i dont want to replace the child node. The whole structure of the input tree should be transformed in another one. Cheers Kim Tang

6 24

boost accumulators - kth central moments? [and slow compiles]
by Gordon Anderson 09 Feb '09

09 Feb '09

I have a couple Qns about boost accumulators in 1_37_0 - 1) Central moments ? - There are accumulators for simple kth moments - boost/accumulators/statistics/moment.hpp Are central kth moments [relative to the mean] there somewhere? Obviously you can construct one from the other with a weighted sum of the k moments on the same data. 2) Slow build ! - It is a bit slow to compile code using eg. moment<6> accumulators.. It takes around 12 seconds on my machine to build sample code with 'moment<12>' exponential recursive template evaluation?? I'm using gcc 4.2.4 cheers, gord.

3 2

Re: [Boost-users] Cycles of shared_ptr
by Donovan Parks 09 Feb '09

09 Feb '09

Hello Alan, Thanks for the response. I appreciate what you are saying and realize that we need to draw up some diagrams which indicate ownership rules. It turns out my problem was actually unrelated to cycles of shared_ptr, but it is still good to be thinking about this possibility. Cheers, Donovan

1 0

Building on OS X
by Stuart Dootson 09 Feb '09

09 Feb '09

So, I've installed Xcode 3.0, followed by Xcode 3.1.2. gcc is defaulting to 4.0, which builds Boost fine (using the configure & make install OR bjam routes). However, I would like to build boost with gcc 4.2, as installed with Xcode 3.1.2 - I've not managed to find the correct incantation yet to persuade Boost to use 4.2 vs 4.0 - help!!!! I guess I could just alter the gcc and g++ soft links in /usr/bin, but that seems...wrong. Assistance would be greatly appreciated. TIA!!! Stuart Dootson

4 5

Re: [Boost-users] Cycles of shared_ptr
by Alan M. Carroll 09 Feb '09

09 Feb '09

Detecting cycles at run time is expensive and rarely worth the effort. It is much better to have a clear understanding of your data structure and awareness of where cycles could arise and then break them using Boost.weak_ptr, as weak_ptr was created to deal with exactly this problem. For instance, if you have a tree that needs parent to child and child to parent links, then you can use shared_ptr in the parent class and weak_ptr in the child struct. The important thing is to be clear on the ownership semantics. There are so many possibilities that it's hard to be more specific without a lot of detail. At 12:01 PM 2/9/2009, you wrote: >I have recently started using boost::shared_ptr in my code. I >understand that cycles of shared_ptr will cause a problem. Now, I >strongly believe that such a situation has arisen in my code and am >wondering if there is any advice/tips on how one can detect when a >cycles of shared_ptr has occured. The code I am working on is >developed by several people and the shared_ptr of interest is passed >to many objects.

1 0

Cycles of shared_ptr
by Donovan Parks 09 Feb '09

09 Feb '09

Hello, I have recently started using boost::shared_ptr in my code. I understand that cycles of shared_ptr will cause a problem. Now, I strongly believe that such a situation has arisen in my code and am wondering if there is any advice/tips on how one can detect when a cycles of shared_ptr has occured. The code I am working on is developed by several people and the shared_ptr of interest is passed to many objects. Thanks for any and all help. Cheers, Donovan PS: This is my first time posting to boost-users.

1 0

Building boost regex on Vxworks compiler
by Raghavender Boorgapally 09 Feb '09

09 Feb '09

Hi, I am assigned the job of building the boost regex on VXworks gcc compiler. >From the boost regex documentation, I see that we need to have a make file corresponding to the platform to build the library. Can some one tell me how to build it for Vxworks gcc, if they have built the library earlier.

2 1

[regex] character sets
by Detlef Meyer-Eltz 09 Feb '09

09 Feb '09

I'm working with CBuilder6 and boost boost_1_34_1 and found, that regex expression1("[[:print:]]+"); matches with regex_search the whole string: testtext = "ab\r\ncd ef"; including the linebreak and the white space. I had assumed, that the print-class would consist exactly in the set of characters for which the standard function isprint returns true but isprint returns false for the space characters. Is there there any documentation with the exact enumerations of the different character sets? With Kind Regards, Detlef Meyer-Eltz

2 2

[Smart Ptr] make_shared for arrays
by Alexander 09 Feb '09

09 Feb '09

Why isn't there ? It looks easier to implement than for single objects as there would be only unary version which would take array size.

1 0

boost serialization not working unless you have at least one virtual function in base
by Martin Seiler 09 Feb '09

09 Feb '09

Hi all, Does anybody know why this is the case? Is there a way to make a class "polymorphic" without adding a dummy virtual function? Is this some "silly" optimization done by C++ compilers? Not generating vtables? It took me quite some time to figure this one out. It is mentioned in the documentation (2 lines, easily missed). http://www.boost.org/doc/libs/1_37_0/libs/serialization/doc/serialization.h… Maybe add a big huge WARNING? Also please mention this fact in the examples! (like in demo.cpp?) regards M. Seiler #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> #include <boost/serialization/nvp.hpp> #include <boost/serialization/base_object.hpp> #include <boost/serialization/type_info_implementation.hpp> #include <boost/serialization/export.hpp> #include <boost/archive/archive_exception.hpp> #include <boost/archive/basic_xml_archive.hpp> class A { public: A() : test(33) { } int test; virtual void dummy(){} // if you comment this line it won't work }; class B : public A { public: B() : test_b(52) { } int test_b; virtual void foo(){}; }; namespace boost { namespace serialization { template<class Archive> void serialize(Archive & ar, A & o, const unsigned int version) { ar & make_nvp("test", o.test); } template<class Archive> void serialize(Archive & ar, B & o, const unsigned int version) { ar & boost::serialization::make_nvp("A", boost::serialization::base_object<A>(o)); ar & make_nvp("test_b", o.test_b); } } } BOOST_CLASS_EXPORT(B); // Should register B for all archives such that it can be serialized of the basepointers void main() { A* a = new B(); boost::archive::xml_oarchive archive(std::cout); archive.template register_type<B>(); // Should register B for this archive archive & boost::serialization::make_nvp ("a", a); // If you check the output, you see that `{B}.test_b' was not serialized int i; std::cin >> i; }

4 3