February 2007 - Boost-users

question about pool::get_next_size/set_next_size
by Stefan Felkel 21 Feb '07

21 Feb '07

Hi, is there a way to access pool::get_next_size, pool::set_next_size from my own fast_pool_allocator without modifying the boost source (1.33.1). This my_fast_pool_allocator is a copy of fast_pool_allocator that tries to allocate the requested memory with next_size divided by 2 if malloc returns NULL. I added static member functions get_next_size/set_next_size to singleton_pool which works, but is there a way without modifying the boost source. If not, would it make sense to add these two methods in a future boost release. This is the allocate method of my_fast_pool_alloctor: template <typename T, typename UserAllocator, typename Mutex, unsigned NextSize> class my_fast_pool_allocator { ... static pointer allocate(const size_type n) { typedef boost::singleton_pool<boost::fast_pool_allocator_tag, sizeof(T), UserAllocator, Mutex, NextSize> mypool; pointer ret = NULL; for( ; !ret; ) { if (n == 1 ) { ret = static_cast<pointer>( mypool::malloc() ); } else { ret = static_cast<pointer>( mypool::ordered_malloc(n) ); } size_t nsize = mypool::get_next_size(); if( !ret ) // not enough memory available { if( nsize <= n ) // even less than we really need break; nsize >>= 1; // request half the size nsize = std::max( nsize, n ); // at least as much as we really need } mypool::set_next_size(nsize); } if (ret == 0) throw std::bad_alloc(); return ret; } ... }; Regards, Stefan

1 0

question about boost::interval
by Marco Correia 21 Feb '07

21 Feb '07

Hi, I'm a little bit confused about the documentation in http://www.boost.org/libs/numeric/interval/doc/rounding.htm, in the subsection "Transcendental functions". Does the "Please note" paragraph means that the provided classes for transcendal functions (_opp and _std) do not guarantee the inclusion property? From what I understood it says that it depends on the implementation of the std::* math functions, that, in most cases, don't obey current rounding mode. Am I getting this right? In case I am, does anyone knows if I'm safe with gcc std library (perhaps by using the -frounding-math switch)? Thanks! Marco -- Marco Correia <mvc(a)di.fct.unl.pt>

2 2

missing files?
by Jan Van den bosch 21 Feb '07

21 Feb '07

Hi, I'm trying to build the examples of the Channel event framework ( channel.sourceforge.net) Apparently, the Boost archives at sourceforge (.zip, .tar.gz and .tar.bz2 files) are missing several necessary files, like serialization/array.hpp. When I do a CVS checkout of Boost however, those files are included. Are they still part of Boost? Thanks in advance, Jan

1 0

split_iterator<string::iterator> example
by Meryl Silverburgh 21 Feb '07

21 Feb '07

hi, Can someone please tell me where I can find example of split_iterator<string::iterator> If I have 2 split_iterator<string::iterator>s, how can I compare the 2 strings that the split_iterator pointing to? Thank you.

2 13

[Boost][bind] problem using bind with non-copyable class members?
by Darren Garvey 21 Feb '07

21 Feb '07

Hi, I'm trying to use bood::bind and I'm having a problem with compilation when I'm passing an object that holds a std::stringstream. The errors are to do with std::ios_base and std::streambuf's copy constructors being private, but as far as I knew (and understand from the docs), boost::bind passes by reference, so nothing should be copied. I'm sure I'm missing something fundamental here, so any enlightening would be very welcome. My setup is gcc 4.1.2 on Ubuntu 6.10. Here's an example program that fails: #include <sstream> > #include <boost/bind.hpp> > #include <boost/thread.hpp> > > using namespace boost; // sorry > using namespace std; // oops > > > struct holder > { > std::ostringstream oss; > }; > > void sub_main( holder& h ) > { > return; > } > > class event_handler > { > public: > void operator() ( holder& h ) > { > thread t( bind( &sub_main, h ) ); > t.join(); > } > }; > > int main( int, char** ) > { > event_handler e_h; > holder h; > > e_h(h); > > return 0; > } Regards, Darren

2 2

[signals] building sandbox documentation?
by Frank Mori Hess 20 Feb '07

20 Feb '07

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Timmo, Do you know a bjam command to build the signals documentation in boost-sandbox/libs/signals/doc ? I'd like to update them to reflect the changes we've made for thread safety. Do I have to copy the files over to a real boost source tree to get them to build? I am not very familiar with bjam. - -- Frank -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFF2xd85vihyNWuA4URAkBBAJ43BFzalbQSbzMaKNWuaEPR3tEpUQCdE67i 7ozVU/0NcLWKVW8gNv4KR5E= =WB61 -----END PGP SIGNATURE-----

2 2

Re: [Boost-users] Cannot build boost on MSVS8 per instructions
by DAVID CRAVEY 20 Feb '07

20 Feb '07

Hi All, I am having the same errors as in this post: http://lists.boost.org/boost-users/2006/09/22001.php I have tried downloading the boost-jam both binary: boost-jam-3.1.13-1-ntx86.zip (I get the same error as the above post) And from source: boost-jam-3.1.13.zip I get a "Debug Assertion Failed!" error when I run build.bat Program: C:\Boost\boost-jam-3.1.13\bootstrap\jam0.exe File: dtoxtm64.c Line: 67 Expression: ( ( ( long )( yr - 1990) >= BASE_YEAR) && ( ( long )( yr - 1990 ) <= _MAX_YEAR64 ) ) Running the prebuilt bjam.exe file in the Threads build directory (running from Visual Studio 2005 Command Prompt): C:\Boost\boost_1_33_1\libs\thread\build>bjam "-sTOOLS=vc-8_0" Built three library files. So I think the bjam.exe file probably is not corrupt. If I run (from Visual Studio 2005 Command Prompt): C:\Boost\boost_1_33_1>bjam "-sTOOLS=vc-8_0 --with-threads" I get the same 00012f2a error as before. Other Notes: I have Visual C++ 2003 Standard and Visual Studio 2005 Pro Installed. Starting with the appropriate command prompt, I get the same error on both at Offset 0001f2fa. My machine is an Dell Latitude Centrinio Running Windows XP Pro SP2. Any help would be greatly appreciated! Thanks, David Cravey

3 2

Boost serilaization onexit problem
by gast128 20 Feb '07

20 Feb '07

Dear all, We use (too many) singletons in our program and now the program crashes in Boost Serialization because in an onexit function a static member (oserializer::instantiate) of Boost Serialization is already destroyed. Is there a way to force an intialisation of these static members so that they are destroyed after my singleton destroys? The code can be reproduced by the follwing snippet: // #define BOOST_ALL_DYN_LINK #define BOOST_LIB_DIAGNOSTIC #include <fstream> #include <boost/archive/text_oarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/serialization/base_object.hpp> #include <boost/serialization/version.hpp> struct Foo { Foo() : m_i(0) {} template <class Archive> void serialize(Archive& ar, const unsigned int /*version*/) { ar & BOOST_SERIALIZATION_NVP(m_i); } int m_i; }; void f (); void test_serialize (); //---------------------------------------------------------------------------- // Function main //---------------------------------------------------------------------------- // Description : main //---------------------------------------------------------------------------- int main() { atexit(&f); test_serialize(); return 0; } void f() { test_serialize(); } void test_serialize() { std::ofstream ofs("test.xml"); boost::archive::xml_oarchive oa(ofs); const Foo foo; oa << BOOST_SERIALIZATION_NVP(foo); } wkr, me

2 2

[Review][Bimap] Reviewers needed for Bimap review
by Ion Gaztañaga 20 Feb '07

20 Feb '07

Hi to all, Bimap review started 5 days ago and we're having some good comments but no thorough reviews. I would like to encourage people to write a Bimap review, because we still have five more days and I would propose extending that if that's necessary. Container and STL experts, we need you! I think that Bimap is a great library and I'm pretty sure you'll be happy to review it once you start reading the docs. Start reading the documentation here: http://tinyurl.com/y95qng And download the library here: http://boost-consulting.com/vault/index.php?&directory=Containers Regards, Ion

1 0

Using Boost string library
by Meryl Silverburgh 20 Feb '07

20 Feb '07

Hi, I need to compare 2 strings in this format (a integer separator by '/': /1/1 /1/1/2 /1/1/3 so "/1/1" is less than "/1/1/2" and "/1/1/2" is less than " "/1/1/3". Is there anything in Boost string library to help me to write such a comparison? Is there a string iterator which literates base on the separator '/'? My idea is to find an iterator to loop thru the string based on separator '/'. Thanks for any idea.

3 14