- Boost-users - lists.boost.org

[Boost.MPI] SendingReceiving an array "view"
by Joel Falcou 27 May '09

27 May '09

I have a small class view<T> that containes a T* and a size. When I send/receive view using Boost.MPI, I want transfert the actual n data pointed by the view pointer and send the size. On receiving, i will receive the size, checks if they are compatible and receive the data into the zone pointed by the pointer of the receiving view. Of course, I may do something more complex than send/Receive, like all_2_all etc. So i was thinking to give view<T> a proper serialize function so boost.serialization may kick in. As the behavior of sending and receiving is different, I use the load/save idiom to separate the behavior. The question is then, how to serialize the n elements poitned by a raw pointer to those data in an efficient way ? Regards

4 4

Integral type with no upper limit?
by Robert Dailey 27 May '09

27 May '09

Is there an integral type in Boost that has no upper limit? The only physical limit would be the amount of memory on the system. And of course, the larger the number, the heavier the math is on such an object, because more cycles are required.

4 4

transform_iterator causing strange problems
by Lindley M French 27 May '09

27 May '09

I'm not sure if this is the fault of transform_iterator----I really don't see how it could be----but I'm really confused here. I've got an unordered_map<GroupID, Group> (typedefed to GroupMap) in my class. The specifics of those types don't matter except that GroupID is an integral type, and Group contains a std::vector vec. I'm trying to use a transform_iterator so that my class may be viewed as a group container: typedef boost::transform_iterator<GroupSelector, GroupMap::const_iterator> const_iterator; const_iterator begin() const { return boost::make_transform_iterator(currgroups.begin(),GroupSelector()); } const_iterator end() const { return boost::make_transform_iterator(currgroups.end(),GroupSelector()); } std::pair<const_iterator, const_iterator> groups() const { const Group &grp = *(begin()); return std::make_pair(begin(),end()); } Where my GroupSelector class is defined as struct GroupSelector { typedef const Group& result_type; const Group& operator()(const std::pair<GroupID, Group> &p) const { std::cout << p.second.vec.size() << " " << &p.second << std::endl; return p.second; } }; Note the output statement. Now, the usage in my main function is for (tie(iter,end) = final.groups(); iter != end; ++iter) { const Group& grp = *iter; cerr << &grp << endl; cerr << grp.segs.size() << endl; ... My output is 10 0012F50C 10 0012F65C 0012F65C 1243176 which is strange for two reasons. The first line comes from GroupSelector when I dereference begin() in the groups() function. The second comes from GroupSelector when I dereference iter in main. As these should both correspond to the same iterator, I can't understand why the address of the returned Group reference is different. The 3rd and 4th lines are outputted in main. Here's the second confusion: The address for the Group object is the same as in the most recent GroupSelector output, yet now the size of the vector within the Group comes out wrong. HUH?

5 8

Re: [Boost-users] [BGL] Trying to use subgraph
by Lindley M French 27 May '09

27 May '09

My inbox was full so I had to grab the response off the archive, hence the lack of quote. However, this: typedef boost::property<edge_index_t, size_t, EdgeProperties> EdgePropWithIndex; .... typedef boost::subgraph< boost::adjacency_list< boost::vecS,// Use list for edges boost::vecS,// Use vector for nodes boost::undirectedS,// Undirected graph SegIndex,// Nodes are segments EdgePropWithIndex// Edges have weights > > BoostGraph; still gives me compile errors: error C2065: 'edge_index_t' : undeclared identifier error C3203: 'property' : unspecialized class template can't be used as a template argument for template parameter 'EdgeProperty', expected a real type On a (possibly) related note, I'm trying to use the above subgraph in kruskal_minimum_spanning_tree(), but it's giving me an error about error C2039: 'vertex_bundled' : is not a member of 'boost::subgraph<Graph>' I'm quite sure this error didn't appear before I started trying to use the subgraph adapter.....any idea what's wrong?

2 19

symmetric_matrix operation
by Archie14 26 May '09

26 May '09

I am very new to the whole Ublas library. I have two symmetric matrices and I need to divide each element of the first matrix to the element in the second matrix. I am now doing smth like this: int _matrixsize; symmetric_matrix<double> mfirst(_matrixsize); symmetric_matrix<double> msecond(_matrixsize); symmetric_matrix<double> ratio (_matrixsize); for (int ifirst = 0; ifirst < _matrixsize; ifirst++) for (int isecond = 0; isecond < _matrixsize; isecond++) ratio(ifirst, isecond) = mfirst(ifirst, isecond) / msecond (ifirst, isecond); My questions are: 1. Can I use iterators (begin1, begin2) and if yes, then how? Examples in the doc use above method of iterating through elements, without using iterators. 2. I did not find built-in operators in the symmetric_matrix to do this division, but may be I overlooked smth or did not understand the docs? 3. If matrix "msecond" contains zeros - what is the "boost-way" to deal with this?

2 2

Re: [Boost-users] [Serialization] Guard against attempting to deserialize the wrong type
by Dan Eaton 26 May '09

26 May '09

Robert, Could you please clarify what you mean by that? Do you mean the first and last tags of the archive are checked? I can't reproduce that behaviour either. int foo = 1; unsigned int bar =1; // create stream & output xml archive oarchive << BOOST_SERIALIZATION_NVP(foo); // create input xml archive from stream iarchive >> BOOST_SERIALIZATION_NVP(bar); This deserializes happily. Thanks, Danny From: Dan Eaton Sent: Monday, May 25, 2009 1:30 PM To: 'boost-users(a)lists.boost.org' Subject: [Serialization] Guard against attempting to deserialize the wrong type Hello, I'm just wondering if NVP/XML archives check the name during deserialization? I just had a problem whereby I tried deserializing an object to one type, but the archive actually contained a different (larger ) object type. The first serialized element of the object just so happened to be an unsigned int, and I was trying to deserialize an enum, so the output actually seemed sensible, and didn't cause any problems until I later deserialized the full object from the same archive, which caused a segmentation fault in the boost::serialize source (trying to get ETI from a null pointer). I was using a polymorphic text archive, and as I mentioned above, am wondering if I switched to an XML archive if this sort of user-error would be detected. Thanks, Dan Eaton Software Developer Acceleware Corp. (TSX-V:AXE) www.acceleware.com<http://www.acceleware.com/> Phone: +1.403.249.9099 Ext.295 Fax: +1.403.249.9881 Email: daniel.eaton(a)acceleware.com<mailto:daniel.eaton@acceleware.com> Get superpowered! Acceleware gets your products to market faster, better and stronger! This e-mail may contain information that is privileged and confidential and subject to legal restrictions and penalties regarding its unauthorized disclosure or use. You are prohibited from copying, distributing or otherwise using this information if you are not the intended recipient. If you have received this e-mail in error, please notify us immediately by return e-mail and delete this e-mail and its attachments from your system. Thank you. (c) 2008 Acceleware, Inc. All Rights Reserved

2 1

Question about enable_if / disable_if
by Zachary Turner 26 May '09

26 May '09

Suppose I have a template function template<typename T> void f(const std::string&, T); which I want there to be a couple of different variants of. Version 1 - Enabled if and only if boost::is_same<T, bool> Version 2 - Enabled if and only if boost::is_same<T, foo> (foo is a class) Version 3 - Enabled if and only if boost::is_enum<T> Version 4 - Otherwise Is there an easy way to express this? It seems I run into a bunch of overload ambiguity problems. I'm sure I can get around this by adding to every version the negation of all the other conditions, but this seems unnecessarily complicated.

4 3

Second call: The Boost Bug Sprint is coming!!
by Marshall Clow 26 May '09

26 May '09

The boost bug sprint is coming! It starts _this_ Friday, May 28th, and runs through Sunday June 7th. The goal is to address a large number of open bugs/patches/feature requests in the Trac system at <http://svn.boost.org/trac/report/1>. There are currently 791 open tickets; we'd like to knock that number down _a lot_. What can you do? 1) You can fix bugs. * If you are a library maintainer, you can commit your fixes. ;-) * If you are not a library maintainer, you can fix a problem on your local machine (and test it!!), and then attach your fix as a patch to the Trac ticket for the library maintainer to commit. If you do this, please provide a test case that exercises the problem; i.e, one that fails w/o your patch, and succeeds afterwards. 2) You can characterize bugs. * You can attempt to replicate the bug on your system(s), and attach information that you discover to the Trac ticket. This will make it easier for other people to solve the problem, even if you cannot. * You can write test cases demonstrating the problem, and attach them to the Trac ticket. * You can determine if the bug still exists; sometimes they get fixed, and the ticket never gets closed. In this case, try to identify a checkin that addresses the bug; it's always suspicious when bugs go away "by accident." 3) You can test patches. * Some of the tickets in the Trac system contain patches - some are quite old. You can install these patches on your system, and determine if they (a) build, and (b) work. * You can write test cases for the patches, and attach them to the ticket. 4) You can help implement new features. * There are a whole set of "feature requests" in the Trac system. If you have an idea about one of these, you can add comments to the Trac system. You can attempt to implement these new features, and attach the patch to the ticket. (Don't forget the test cases!!). Every day during the bug sprint, I'll post the # of tickets modified/closed/opened on the previous day, as well as the total number of open tickets. --------- Sign-up sheet -------- The following people have committed to participate in the bug sprint. Please add you name here, along with what you can do. 1) Marshall Clow (coordination, Mac OS X characterization) 2) Beman Dawes (filesystem, system, timer) 3) Doug Gregor (function, graph, logic, mpi, numeric, property_map, signals) 4) "joel falcou" <joel.falcou(a)u-psud.fr> ( linux gcc testing/VC 2008 testing, patching) 5) Michael Caisse <boost(a)objectmodelingdesigns.com> (bug fixing) 6) Jeremiah Willcock <jewillco(a)osl.iu.edu> (will test and commit graph patches) 7) David Abrahams (iterator, python) 8) Sohail Somani (patches for g++ 4.2.4) -- -- Marshall Marshall Clow Idio Software <mailto:marshall@idio.com> It is by caffeine alone I set my mind in motion. It is by the beans of Java that thoughts acquire speed, the hands acquire shaking, the shaking becomes a warning. It is by caffeine alone I set my mind in motion.

1 0

Units library and wstring
by Macumber, Daniel 26 May '09

26 May '09

Hello, my group has been simultaneously incorporating the boost units library into our code as well as converting all internal strings from std::string to std::wstring. I am having difficulty merging the changes, it seems that the boost units library has hard coded std::string as the return type for the methods which print units, powers, etc. Is there an easy way to template use of the units library on wchar_t instead of char? Are there any other suggestions on how to use the units library in a program which uses wstring for everything else? Thanks a lot, Dan

2 1

[BGL] dijkstra_shortest_paths on dynamically growing and shrinking graphs
by Ingo Maindorfer 26 May '09

26 May '09

Hello, consider the follwoing Graph: struct EP { double weight; }; typedef boost::adjacency_list< boost::listS, boost::vecS, boost::directedS, boost::no_property, EP > Graph; The call of dijkstra_shortest_paths is clear to me on that graph: boost::dijkstra_shortest_paths( m_graph, a_start_idx, weight_map( m_weight_map ).distance_map( &distance_vector[0]).predecessor_map( &parent_vector[0] ) ); where parent_vector and distance_vector are plain std::vector of the according size. Over time, I have to remove the "oldest" edges and vertices with lower index via remove_vertex() and remove_edge() from the graph and add some "newer" edges and vertices to the graph. My problem is, that the dijkstra_shortest_paths method dosen't work when I provide a plain std::vector to the distance_map() and predecessor_map() parameter of the dijkstra_shortest_paths call because of the mapping from the vertex descriptor to position in the vector. Can anybody tell me how to provide the "right" type of parameter for distance_map() and predecessor_map() ? Maybe something like boost::property_map< Graph, double EP::*>::type m_weight_map = boost::get( &EP::weight, graph ); that I'm already have done for the weight_map parameter but I'm looking for something that can handle the mapping stuff of vertex descriptors correctly. Performance is not a subject because of a few hundred vertices and edges. I'm on Boost 1.39. Kindly, Ingo

2 1