February 2010 - Boost-users

array of small world graphs
by Vasilis Hatzopoulos 04 Feb '10

04 Feb '10

Hi all, new to boost and experiencing a bit of an issue with generating a collection of small worlds. Basically i want for each pass of the loop to generate a small world with a certain pRewire. Here's some code typedef small_world_iterator<mt19937, Graph> SWGen; base_generator_type generator(1u); for(int i = 0; i < num_graphs; i++){ Graph g( SWGen( generator, N, k, pRewire, SWGen(), N ); . . some operations on g, i.e. path lengths e.t.c. } As it happens all graphs g are the same graph while what i want is a collection of different small worlds with the same pRewire. Why does the state of the generator not advance after the first Graph is generated? Can anyone please help me with this. Thanks for your time, Vasilis.

2 1

building boost under mingw-w64
by Alexander Kalinin 04 Feb '10

04 Feb '10

Hi, Here some details about building boost under mingw-w64. I am using only "threads" and "date_time" libraries from boost and I started to building boost by command: bjam --prefix=c:\boost toolset=gcc address-model=64 variant=debug,release link=static,shared threading=multi --with-thread --with-date_time install and I got an error: ./boost/thread/win32/basic_ timed_mutex.hpp:152: error: cast from 'void*' to 'long int' loses precision This caused by the macros BOOST_INTERLOCKED_COMPARE_EXCHANGE_POINTER defined in interlocked.hpp. This macros has several definitions that depends on building defines. So I add BOOST_USE_WINDOWS_H define into the building command: bjam --prefix=c:\boost toolset=gcc address-model=64 variant=debug,release link=static,shared threading=multi --with-thread --with-date_time define=BOOST_USE_WINDOWS_H install add boost was compiled successful. The question is. Is it possible to build boost under mingw-w64 without BOOST_USE_WINDOWS_H define?

1 0

[BGL] Kolmogorov max flow
by Olivier Tournaire 04 Feb '10

04 Feb '10

Hi all, I am currently trying to use kolmogorov_max_flow algorithm. I am building a graph from another lib, namely CGAL. I first build an arrangement of the plane, and use the dual arrangement representation provided by CGAL (see http://www.cgal.org/Manual/3.3/doc_html/cgal_manual/Arrangement_2/Chapter_m…) So, each vertex of the graph is a face of the arrangement, and each edge in the graph is an edge of the arrangement. CGAL provides a graph_traits to use BGL algorithm. However, I do not know how to use kolmogorov_max_flow with this graph_traits. The doc states that reversed edges must be included in the graph. The graph provided by CGAL already has reverse edges. How can I deal with that ? Could you please give me the way to call kolmogorov_max_flow withthe graph provided by CGAL ? Best regards, Olivier

2 3

multi_array: how can I efficiently have a column selected?
by petros mamales 04 Feb '10

04 Feb '10

Hi, I am a real newbie here, so please excuse the naivity of the question. I, say, have a 3d array (fortran storage), and want to create a view to it so that I select say: grid[ indices[ [range(9,10)] [3], [range(9,10)] ]; I am not clear on exactly what this creates: from the documentation, I see this creating a 3d view. Can I use stl-like iterators on this, to use stl type algorithms? Is there any other repository of examples that would use high-level constructs like this so that I could look into a few examples? Again, I apologize for the bredth of the question. Best Regards, Petros

2 4

Re: [Boost-users] Thread crash on interruption mingw
by Chris Spencer 04 Feb '10

04 Feb '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Finally got it sorted. I recompiled gcc to use Dwarf exceptions instead of SJLJ (added '--disable-sjlj-exceptions --enable-dwarf2' to the configure line) and it works perfectly now. Sorry for the spam, maybe someone in the future with the same problem will find it useful.. Chris. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAktqE8UACgkQ6iqRtkAADZixkgCfVy6RHLU3tj0xRCAMqGl8ukoH DnIAn1gpYWAZH01M2/74J96Ovt/YfS/q =aEbg -----END PGP SIGNATURE-----

1 0

Re: [Boost-users] Thread crash on interruption mingw
by Chris Spencer 03 Feb '10

03 Feb '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Well to make things more interesting I have discovered that using - -static-libgcc it works as intended, but if I do that it means exceptions won't traverse DLL boundaries, which needs -shared-libgcc to work... Chris. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAktqA2IACgkQ6iqRtkAADZi1NwCfSzzcdsDZAJcbVuUvvdZjMh1T Rt8AoKSZs0dZTYYLl9DEzLO1pWMGQfgi =AFv4 -----END PGP SIGNATURE-----

1 0

Protected binds
by Zachary Turner 03 Feb '10

03 Feb '10

Performance issues aside, is there any reason why i should prefer to use boost::bind() by itself as opposed to boost::protect(boost::bind ())? In other words, is there a situation where given a certain functional requirement, boost::bind produces correct results but boost::protect(boost::bind()) doesnt? Back to performance, just how much extra work is protect going to do in an optimized build? Zach Sent from my iPhone

1 0

Thread crash on interruption mingw
by Chris Spencer 03 Feb '10

03 Feb '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all, I'm having some trouble with the thread library. When I interrupt a thread it is causing my application to crash. Here is a simple test file: #include <boost/thread.hpp> #include <iostream> #ifdef _WIN32 #include <windows.h> #endif using namespace boost; using namespace boost::posix_time; using namespace std; void threadFunction() { try { boost::this_thread::sleep(seconds(10)); } catch (thread_interrupted&) { cerr << "interrupted" << endl; } } int main() { thread myThread(threadFunction); #ifdef _WIN32 Sleep(1000); #else sleep(1); #endif myThread.interrupt(); myThread.join(); return 0; } If I compile this on Linux, or on Windows with MSVC it works perfectly every time, but if I compile with mingw it *usually* crashes after printing 'interrupted'. If I put a sleep in between the interrupt and join though it never crashes, so there appears to be a race condition somewhere.. Compiled with: i386-pc-mingw32-g++ -mthreads thread_test.cpp - -lboost_thread -o thread_test.exe $ i386-pc-mingw32-g++ --version i386-pc-mingw32-g++ (GCC) 4.4.3 Boost version 1.42 Any help would be appreciated. Chris. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAktp+iAACgkQ6iqRtkAADZiiWwCgrd3R/G2yCstU0ruuipCPbISo IAcAoOCE93g+alGcSUf7ZedHYC7Mwq3u =BHbA -----END PGP SIGNATURE-----

1 0

Re: [Boost-users] Boost release 1.42.0 available
by Tom Kent 03 Feb '10

03 Feb '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I've built windows libraries (vc7.1, vc8, vc9) off of this, if you want them, you can get them from my website: http://boost.teeks99.com Tom > Subject: [Boost-users] Boost release 1.42.0 available > From: Beman Dawes (bdawes_at_[hidden]) > Date: 2010-02-02 16:52:50 > > Release 1.42.0 of the Boost C++ Libraries is now available. > > These open-source libraries work well with the C++ Standard Library, > and are usable across a broad spectrum of applications. The Boost > license encourages both commercial and non-commercial use. > > This release contains one new library and numerous bug fixes for > existing libraries. For details, including download links, see > http://www.boost.org/users/news/version_1_42_0 > > The release can also be downloaded directly from SourceForge. See > http://sourceforge.net/projects/boost/files/ > > To install this release on your system, see > http://www.boost.org/doc/libs/1_42_0/more/getting_started/index.html > > Thanks, > > --The Boost release team -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAktp80AACgkQccZrtZUgfJA55wCghynRoR4MYtt1EJvcFyoTIrUY obgAnigLXnvjDd6AS2h3UqkvD1NkrdOR =hHFS -----END PGP SIGNATURE-----

1 0

Re: [Boost-users] [fusion] xlC 6.0 building math lib - Errors
by Jerry Jeremiah 03 Feb '10

03 Feb '10

> Date: Wed, 3 Feb 2010 12:13:56 -0000 > From: "John Maddock" <john(a)johnmaddock.co.uk> > > > Here is the entire stack trace for the math library error I get: > > Sigh... those are still truncated and don't give the full template > parameters :-( > > John. > > Date: Wed, 3 Feb 2010 12:15:46 -0000 > From: "John Maddock" <john(a)johnmaddock.co.uk> > > > info.cpp > > > "./boost/optional/optional.hpp", line 439.66: 1540-0210 (S) "T" is not a > > base class of "locale". > > "./boost/optional/optional.hpp", line 439.10: 1540-0700 (I) The previous > > message was produced while processing > > "boost::optional_detail::optional_base<std::locale>::destroy_impl > > (is_not_reference_tag)". > > :-( > > I admit to being a bit perplexed by this, particularly as the IBM AIX > tests are mostly passing with xlC 10.1, options used are here: > http://beta.boost.org/development/tests/trunk/developer/output/ > IBM_Canada_Ltd-boost-bin-v2-libs-config-test-config_info-test- > vacpp-debug.html > > John. Hi John, I knew it compiled with xlC 10. And clearly IBM has improved their compiler significantly since I bought xlC 6. I figure xlC 6 is like MSVC 6 - it does templates but not really. It makes sense - xlC 6 is four versions behind and so is MSVC 6. It's just too old. Sorry about the truncated error messages - I have no idea how to make them contain more detail. Jerry

1 0