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

Re: [Boost-users] [Announce] Boost Installer for Visual C++ 7.1/8.0
by Jeff Flinn 18 May '06

18 May '06

David Abrahams wrote: > Boost Consulting is pleased to announce the availability of a Windows > installer for the current Boost release (1.33.1). See: > > http://www.boost-consulting.com/download.html Is this to be the final location? > Thanks to everyone who made this possible, including Rene Rivera, Eric > Niebler, Devin Smith, and especially to Daniel Wallin, who applied > most of the elbow grease. <raucus applause> :-) </raucus applause> > Enjoy! Our customers certainly will. Thanks, Jeff

2 2

[Announce] Boost Installer for Visual C++ 7.1/8.0
by David Abrahams 18 May '06

18 May '06

Boost Consulting is pleased to announce the availability of a Windows installer for the current Boost release (1.33.1). See: http://www.boost-consulting.com/download.html Thanks to everyone who made this possible, including Rene Rivera, Eric Niebler, Devin Smith, and especially to Daniel Wallin, who applied most of the elbow grease. Enjoy! -- Dave Abrahams Boost Consulting www.boost-consulting.com

2 3

Re: [Boost-users] How to write a block-wise filter...
by EMalenfant＠interstarinc.com 18 May '06

18 May '06

Hi Stephen, From: Stephen Webb [mailto:swebb75ky@gmail.com] > My trouble is > knowing when the end-of-input has arrived. Is there a well > defined way that my output filter is notified of a close > event or something similar? I did see close functions > defined in the example filters, but mine isn't getting > called. Is you filter's category convertible to "closable_tag"? A filter's category has to be tagged as "closable" for its close function to be called: http://www.boost.org/libs/iostreams/doc/index.html?path=4.1.3.1 HTH, Éric

1 0

[program_options] Positional options doc/bug in add()
by Ian McCulloch 18 May '06

18 May '06

Hi, Following the example from program_options/overview.html, I tried positional_options_description pd; pd.add("output-file", 2).add_optional("input-file", -1); This doesn't work: positional_options_description::add() returns 'void', and add_optional() is probably a typo for add() ? Currently, the only way to get this to work seems to be positional_options_description pd; pd.add("output-file", 2); pd.add("input-file", -1); But following the style of the rest of the library, I expect it was intended that positional_options_description::add() would return a reference to *this ? Using boost 1.33.1, the CVS version of positional_options.hpp is unchanged. Cheers, Ian McCulloch

2 1

Exposing Class with Private ctor and dtor
by Anand, Kumar 18 May '06

18 May '06

Hi, A class has static method to create and a static method to destroy. And the constructors and destructor of the class is marked as private. How do I expose such a class in Python? Eg. class A { private: static A* instance; A() { }; ~A {}; public: static A* create () { If(instance != 0) instance = new A(); return instance; } static void destroy() { delete instance; instance = 0; } }; In Boost Python, I tried the following but the code doesn't compile unless I make the destructor of class A as public. namespace bp = boost::python; bp::class_<A> ("A", bp::no_init) .def("create", &A::create) .def("destroy", &A::destroy) .staticmethod("create") .staticmethod("destroy") ; Any suggestions? Thanks Kumar

2 1

[boost-users]how to use iostreams::mapped_file
by Sunil Kumar N 18 May '06

18 May '06

Could someone point me to a link for a sample program of using iostreams::mapped_file(except the default tutorial)...I gave a shot at it and snippet is as follows... io::mapped_file_params aparams("c:\\mf.txt"); aparams.new_file_size = 1e8; aparams.mode = BOOST_IOS::out; io::stream_buffer<io::mapped_file> mfStrmBuf(aparams); ostream out(&mfStrmBuf); out << .... keep on writing to file... it works but I need to know how one create a mapped_file that has to grow beyond new_file_size!! That means I dont know the size of the file to be created at compile time!! How to flush explicitly? I am in the process of understanding as to which one offers better performance with respect to time taken when some thousands of lines have to be written to a file.....ofstream vs mapped_file? Thanks in advance :)

1 0

How to write a block-wise filter...
by Stephen Webb 17 May '06

17 May '06

I don't really have my head around the Boost IO filter architecture yet, but I have managed to get a very basic filter working. I'm having some trouble with end-of-input issues. I think what I'm trying to do comes up frequently, so hopefully this will be a straightforward answer... What I'm trying to do is write a filter that takes in input one character at a time (for simplicity -- for now), and outputs encrypted characters. The block size is 4 bytes, so every 4 characters I can encrypt, and then spit out the result. If I hit the end of input, then I want to pad the plaintext, encrypt and output the final block. My trouble is knowing when the end-of-input has arrived. Is there a well defined way that my output filter is notified of a close event or something similar? I did see close functions defined in the example filters, but mine isn't getting called. I tried to look at the base64 encode/decode headers (which would have to deal with the same issue) but I couldn't really follow it. Any help is appreciated. Thanks -Steve

1 0

Re: [Boost-users] [date_time] Need help with a couple simple problems
by Scott 17 May '06

17 May '06

> #include "boost/date_time/local_time_adjustor.hpp" > #include "boost/date_time/c_local_time_adjustor.hpp" > > ptime local_var > =date_time::c_local_adjustor<ptime>::utc_to_local(utc_var); > > What do you mean by the file is in binary format? Thanks for the help. What I mean is I would like to read/write the value of the ptime in the most efficient way possible. So if the ptime is stored as 2 ints, I would like to (de)serialize 8 bytes (4 bytes each int). Portability isn't an issue. What I would rather *not* do is use something like to_iso_string(ptime) and from_iso_string(ptime). The string representation is rather long. But if that's the only way to do it, guess I'll have to make do. Thanks, Scott

2 1

[multi_array] Useful algorithms (from Re: Multi_array: A "proper" iterator over a N>1 Multi_array?)
by Beth Jacobson 17 May '06

17 May '06

On a recent thread, Bruno Martinez provided a std::for_each-like function for multi_arrays which lets users apply a function to each element in a multi_array or view without having to know anything about the dimensions. I've been playing around with it (on MSVC 7.1) and it works perfectly. It seems like a good utility for anyone using boost.multi_array. Working from Bruno's code, I wrote a similar std::accumulate-like function, which could also be generally useful. I'm fairly new to template programming, so if someone could look over the code below and see if I've done something dangerous or stupid, it would be much appreciated. Also, is there some standard Boost way of making this sort of code generally available? It probably isn't at a level where it could be included in the library, but it seems like a shame to just let it languish in the bowels of the mailing list archive. Anyway, here are both algorithms: My accumulate-like function: template<class MA, class T, class F, int dim> struct accum_impl { F f; T &val; accum_impl(T &val, F f) : f(f), val(val) {} T operator()(MA& ma) { std::for_each(ma.begin(), ma.end(), accum_impl<typename boost::add_const<MA::const_reference>::type, T, F, dim-1>(val, f)); return val; } }; template<class Ref, class T, class F> struct accum_impl<Ref, T, F, 0> { F f; T &val; accum_impl(T &val, F &f) : f(f), val(val) {} T operator()(Ref r) { val = f(val, r); return val; } }; template<class MA, class T, class F> T multi_accum(MA& ma, T init, F f) { accum_impl<MA, T, F, MA::dimensionality> impl(init, f); return impl(ma); } And a test: typedef boost::multi_array<int, 3> array_type; typedef array_type::index_range range; typedef array_type::index index; array_type B(boost::extents[4][2][3]); int values = 0; for(index i = 0; i != 4; ++i) for(index j = 0; j != 2; ++j) for(index k = 0; k != 3; ++k) B[i][j][k] = values++; array_type::array_view<2>::type myview = B[ boost::indices[range(1,3)][1][range(0,2)] ]; cout << "Sum: " << multi_accum(myview, 0, std::plus<int>()) << endl; cout << "Product: " << multi_accum(myview, 1, std::multiplies<int>()) << endl; Bruno Martinez's for_each-like function: template<class MA, class F, int dim> struct for_each_impl { F f; for_each_impl(F f) : f(f) {} void operator()(MA& ma) { std::for_each(ma.begin(), ma.end(), for_each_impl<typename MA::reference, F, dim-1>(f)); } }; template<class Ref, class F> struct for_each_impl<Ref, F, 0> { F f; for_each_impl(F f) : f(f) {} void operator()(Ref r) { f(r); } }; template<class MA, class F> void multi_for_each(MA& ma, F f) { for_each_impl<MA, F, MA::dimensionality> impl(f); impl(ma); } And a test: typedef boost::multi_array<double, 3> array_type; typedef array_type::index index; array_type myarray(boost::extents[3][4][2]); typedef array_type::index_range range; array_type::array_view<3>::type myview = myarray[ boost::indices[range(0,2)][range(1,3)][range(0,4,2)] ]; multi_for_each(myview, _1 = 56);

5 6

How to link (properly)
by Jason House 17 May '06

17 May '06

When I run bjam install on my laptop running Mandriva 2006, it places a lot of files in /usr/local/lib with names like libboost_XXX-gcc.a While I know that I can pass -lboost_XXX-gcc and it links, the -gcc bit seems to make things a bit too platform dependent for my liking. What's the proper way to handle that extra bit at the end of the library name? The project I'm working on uses autoconf/configure/make as its build environment.

2 2