August 2010 - Boost-users - lists.preview.boost.org

[1.43.0] Release notice
by Beman Dawes 19 Aug '10

19 Aug '10

Release 1.43.0 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 two new libraries (functional/factory, functional/forward), a major range library upgrade, plus numerous enhancements and bug fixes for other libraries. For details, including download links, see http://www.boost.org/users/news/version_1_43_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_43_0/more/getting_started/index.html Thanks to all who contributed, --The Boost release team

4 5

[BGL] adjacency_list.hpp requires to link with serialization lib in 1.44?
by Oliver Koenig 19 Aug '10

19 Aug '10

Hello, when trying to update Boost from 1.43 to 1.44 in our code we noticed that in 1.44 we need to link to the serialization lib as soon as boost/graph/adjacency_list.hpp is included. Is this correct behavior or is this a bug? To reproduce this it is sufficient to build the following test program hello_adjacency_list.cpp (tested with MSVC-9.0): #include <iostream> #include <boost/graph/adjacency_list.hpp> int main(){ std::cout << "hello adjacency_list\n"; return 0; } With the boost build rule "exe hello_adjacency_list : hello_adjacency_list.cpp ;" compilation fails with: compile-c-c++ ..\..\..\bin\even_utils\test\graph\msvc-9.0\release\address-model-64\threading-multi\hello_adjacency_list.obj hello_adjacency_list.cpp msvc.link ..\..\..\bin\even_utils\test\graph\msvc-9.0\release\address-model-64\threading-multi\hello_adjacency_list.exe LINK : fatal error LNK1104: cannot open file 'libboost_serialization-vc90-mt-1_44.lib' Using the build rule "exe hello_adjacency_list : hello_adjacency_list.cpp /boost//serialization : <link>shared<define>BOOST_SERIALIZATION_DYN_LINK ;" everything works fine. Sincerely ok -- -------------------------------------------- Oliver Koenig Dr.sc.techn. ETH Email: koenig(a)even-ag.ch EVEN - Evolutionary Engineering AG Technoparkstrasse 1 CH-8005 Zurich Tel : +41 (0)44 500 93 60 Fax : +41 (0)44 500 93 61 Mail : info(a)even-ag.ch WWW : www.even-ag.ch -------------------------------------------- -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

5 6

how to detect TCP communication problems on time with synchronous operations
by Gözde Bahar KOŞ 19 Aug '10

19 Aug '10

Hello, I'm using synchronous TCP communication via asio sockets. I have following problems: First problem; When I unplug ethernet cable, "eth0: Link is Down" is written in the console, but socket.write_some command does not detect link down failure. Second problem; During socket.connect command, if I plug ethernet cable, connect command does not detect this till "Connection timed out" error occurs. If I run socket.connect again, it detects the connection. How can I detect these plug/unplug events as soon as the event occurs? Best Regards, Gözde Koş ###################################################################### Dikkat: Bu elektronik posta mesaji kisisel ve ozeldir. Eger size gonderilmediyse lutfen gondericiyi bilgilendirip mesaji siliniz. Firmamiza gelen ve giden mesajlar virus taramasindan gecirilmekte, guvenlik nedeni ile kontrol edilerek saklanmaktadir. Mesajdaki gorusler ve bakis acisi gondericiye ait olup Aselsan A.S. resmi gorusu olmak zorunda degildir. ###################################################################### Attention: This e-mail message is privileged and confidential. If you are not the intended recipient please delete the message and notify the sender. E-mails to and from the company are monitored for operational reasons and in accordance with lawful business practices. Any views or opinions presented are solely those of the author and do not necessarily represent the views of the company. ######################################################################

2 1

Re: [Boost-users] Placeholders and iterator arithmetic in MPL
by Rob Platt 18 Aug '10

18 Aug '10

Steven Watanabe wrote: > After I included all the right headers, it compiled for me. > The only problem is that you're dereferencing past the > end of the sequence. Indeed, thank you! One of my previous attempts was written to avoid the past-the-end, or so I thought... But [slaps forehead] I somehow was thinking that reverse_iter_fold would give reverse_iterators, like those found in the STL. Whereas, next<> and prior<> always operate forwards. So, I had them mixed up. This little example does the job neatly: template<class T> struct works : reverse_iter_fold< T, vector_c<long>, push_front<_1, if_<empty<_1>, long_<0>, deref< next<_2> > > > > {}; It substitutes a long_<0> for the first step of the reverse_iter_fold, when next<> would be past the end. It could be adapted to iter_fold when using prior<>. Dave Abrahams wrote: > stylistically, may I suggest > > struct fails : reverse_iter_fold< .... > {}; You're right to correct me -- I should know that since I have the (your) book! Thank you both for your help, Rob

1 0

Dynamic Properties and Removing Edges/Vertex
by ef 18 Aug '10

18 Aug '10

Hello, I am pretty new and inexperienced so forgive me if any of this is incorrect, any corrections would be appreciated. I am currently using Dot/Graphviz libaries of boost to create a graph. However my graph is huge, to the point where memory will run out. I have devised an algorithm to remove edges/and vertex that are unused when I am done with them. When an vertex/edge is added using read_graphviz function the following happens: the vertex/edge is added to some data structures, as well as the function add_vertex and add_edge are called. I think memory management is pretty straight forward for those situation. The issues that I think will be of some concern is that in order to map names and properties of these vertex, data is added via dynamic properties. >From what I have seen there is no way of removing information from these dynamic properties. (So even if I remove the vertex/edges there is still some residue left in the dp). My question is this correct? Are there functions to remove a certain vertex/edge dynamic property or is this done using the function remove_vertex/edge. From the source code I have seen there is no such thing. Do I need to edit the dynamic property class and make a special delete function, or am I incorrect in my thinking? Any Advice is appreciated, Thanks! EF

2 1

Placeholders and iterator arithmetic in MPL
by Rob Platt 18 Aug '10

18 Aug '10

Hello folks. I'm trying to use the iter_fold and reverse_iter_fold algorithms in MPL, since I want to operate on sub-ranges of an input sequence. Now if I do just the trivial case of dereferencing the current element, then all is well: struct works { typedef typename reverse_iter_fold< T, vector_c<long>, push_front<_1, deref< _2 > > >::type type; }; I can successfully instantiate this with, for example, vector_c<int,1,2,3,4>. But when I do iterator arithmetic (e.g. prior, next or begin) I run into trouble: struct fails { typedef typename reverse_iter_fold< T, vector_c<long>, push_front<_1, deref< next<_2> > > >::type type; }; -- error messages about arg 3 - I suspect that prior is being applied to the placeholder arg type itself, not to the iterator. I'm wondering if I should be using lambda/apply, but I've tried that without success. What's a would-be template metaprogrammer to do? Regards Rob

3 2

[1.44] PDF build of documentation is now available
by John Maddock 18 Aug '10

18 Aug '10

The PDF build of the documentation for Boost-1.44 is now available for download from https://sourceforge.net/projects/boost/files/ The direct links are: https://sourceforge.net/projects/boost/files/boost-docs/1.44.0/boost_1_44_p… https://sourceforge.net/projects/boost/files/boost-docs/1.44.0/boost_1_44_p… https://sourceforge.net/projects/boost/files/boost-docs/1.44.0/boost_1_44_p… https://sourceforge.net/projects/boost/files/boost-docs/1.44.0/boost_1_44_p… As usual please note that this only includes the subset of Boost documentation that is authored in quickbook/boostbook/docbook and so has an automatic conversion to PDF format. Enjoy! Regards, John.

1 0

boost asio sequential?
by Lloyd 18 Aug '10

18 Aug '10

hi, I have a simple boost asio server (copied from asio sample), multiple clients can connect to the server, but the operation performed only in a sequential manner. that is only after completing the first reqest, the second one is performed. Is this the defined behavior? Thanks, Lloyd ______________________________________ Scanned and protected by Email scanner

3 4

thread problem - incorrect checksum for freed object - object was probably modified after being freed
by Kraus Philipp 17 Aug '10

17 Aug '10

Hi, I create some boost::threads: std::vector< worker<T,D> > l_worker; boost::thread_group l_threadgroup; for(std::size_t i=0; i < p_threads; ++i) { l_worker.push_back( worker<T,D>(p_iteration, p_stepsize, m_fulltable, m_derivation, m_optimize, m_static, p_batch) ); l_threadgroup.create_thread( boost::bind( &worker<T,D>::optimize, l_worker[i] ) ); } l_threadgroup.join_all(); In the constructor of my worker-object makes a copy of the parameters ( p_iteration = std::size_t, p_stepsize = T (double), p_batch = std::vector and the others are std::map): ... ) : m_iteration( p_iteration ), m_stepsize( p_stepsize ), m_symbols(), m_derivation(), m_initvalues(), m_staticvalues(), m_batch() { std::copy( p_derivation.begin(), p_derivation.end(), std::inserter(m_derivation, m_derivation.begin())); std::copy( p_initvalues.begin(), p_initvalues.end(), std::inserter(m_initvalues, m_initvalues.begin())); std::copy( p_syms.begin(), p_syms.end(), std::inserter(m_symbols, m_symbols.begin())); } in the optimize-method I use this code: GiNaC::exmap l_dynamic; // exmap is == std::map for(typename std::map<std::string, std::pair<T,T> >::iterator it = m_initvalues.begin(); it != m_initvalues.end(); ++it) l_dynamic[ m_symbols[it->first] ] = it->second.first; If I create only one thread, everything works fine. On more thread I'll get main(4398,0xa07ad720) malloc: *** error for object 0xd2b880: incorrect checksum for freed object - object was probably modified after being freed. *** set a breakpoint in malloc_error_break to debug My idea is to create p_threads number of my worker objects and add some data with the constructor parameters. Than I call the join_all method so every object can work with their own data and after join_all is finished I receive the data from every object. Can someone please explain to me what I did wrong. Thanks Phil

1 0

Problem with nested bind
by przemyslaw.sliwa＠uk.bnpparibas.com 17 Aug '10

17 Aug '10

Hello, I am trying to make a simple nested bind in boost. I have the fillowing structure: struct eventInterval { double start, finish, weight; }; and a typedef typedef boost::shared_ptr<eventInterval> eventIntervalPtr; and then I am trying to copy (actually push back to a vector) all elements which have weight > 0.0. I cannot sompile this bit of code: std::vector<eventIntervalPtr> posIntervals, negIntervals, intervals(eventPtr->decomposeIntervals(anchorDate, today)); std::transform(intervals.begin(), intervals.end(), std::back_inserter(posIntervals), boost::bind(std::greater<double>(), boost::bind(&eventInterval::weight, _1), 0.0)); Could someone help me with this issue? Thanks ___________________________________________________________ This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited. Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-… for additional disclosures.

2 1