- Boost-users - lists.boost.org

Newbie and his template for comments.
by Kamil Burzynski 07 Aug '02

07 Aug '02

Hello. I'm a C++ developer, who is learning STL recently, and loving it :) I've encountered Boost library, which seems to extending STL funcationality a lot :) I've also created a sampel template for usage in my company, and I would like to get comments about it: is this code ok, exception safe, somebody did it better things like that. In my projects I'm often using structures like map<T1, T2 *>, e.g. map<int, Point *> or map<string, Person*> or something like this. So, often I want to do something like that: map<int, Point *> my_map; map<int, Point *>::iterator it; /* .. */ for( it=my_map.begin();it!=my_map.end();++it ) { it->second->Reset(); //sets x,y to 0,0 } example may be stupid, but shows what I'm trying to explain. To improve it I've created a template (given below), so I ca use it like that: for_each( my_map.begin(), my_map.end(), map_fun( &Point::Reset ) ); Which i think is much better. My template is as follows (style is kept from GCC includes, so it may be not as readable as it could be :) template <class _Ret, class _Tp> class map_fun_t : public unary_function<_Tp*,_Ret> { public: explicit map_fun_t(_Ret (_Tp::*__pf)()) : _M_f(__pf) {} template< class _Mp> _Ret operator()(_Mp &__p) const { return ((__p.second)->*_M_f)(); } private: _Ret (_Tp::*_M_f)(); }; template <class _Ret, class _Tp> inline map_fun_t<_Ret,_Tp> map_fun(_Ret (_Tp::*__f)()) { return map_fun_t<_Ret,_Tp>(__f); } What do you think about it? Has it any flaws? Somebody done it already? Is it usable for anyone except me? :) Thanks in advance. -- Best regards from Kamil Burzynski Senior Design Engineer Advanced Digital Broadcast Poland, LTD. - - "Yes, I'm criminal. My crime is that of curiosity."

4 6

Regex++ link problem with VC6 when building with _UNICODE defined
by Edward Diener 07 Aug '02

07 Aug '02

If I build boost 1.28 regex with _UNICODE defined and then attempt to link the static mssd.lib version with a _UNICODE application using the library, I get: boost_regex_vc6_mssd.lib(w32_regex_traits.obj) : error LNK2005: _iswspace already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswalpha already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswupper already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswlower already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswdigit already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswxdigit already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswspace already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswpunct already defined in libcmtd.lib(_wctype.obj) boost_regex_vc6_mssd.lib(c_regex_traits.obj) : error LNK2005: _iswcntrl already defined in libcmtd.lib(_wctype.obj) Debug/TestRegularExpressionGUIStatic.exe : fatal error LNK1169: one or more multiply defined symbols found Error executing link.exe. This only occurs in the debug multi-threaded static library version. Has anyone else seen this or know why it might occur ? If I change the build options for this library from /Od /Gz to /O2, everything works fine and no linker errors occur.

2 1

[BGL] Common children for vertices in DAG
by Erik Arner 07 Aug '02

07 Aug '02

Is there a fast way, with or without preprocessing of the graph, for checking if two non-adjacent vertices in a DAG have one or more common children? Any pointers to components of the BGL that implement, or can be used to implement, this task would be much appreciated. Thanks, Erik

2 4

Tuple from cons list
by j1k1cki 07 Aug '02

07 Aug '02

Hi, Is there a simple way of obtaining a tuple type out of cons list? It also seems natural to build tuple type from two MPL iterators. Is such code available anywhere already? Best regards

1 0

BGL: Problem with parallel edges and filtered_graph
by Michael Kettner 06 Aug '02

06 Aug '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Jeremy, I think I discovered a problem concerning filtered_graphs and parallel edges: Assume I have a graph containing two edges E1 and E2 connecting node A and node B. Now I crate a boost::filtered_graph (myFilteredGraph) with a filter that returns true for only E1. When I now try to get E1 by calling edge = boost::edge( A, B, myFilteredGraph ).first; the result is undefined. Looking at the code (filtered_graph.hpp) I found out that the edge is return by: edge(typename filtered_graph<G, EP, VP>::vertex_descriptor u, typename filtered_graph<G, EP, VP>::vertex_descriptor v, const filtered_graph<G, EP, VP>& g) { typename G::edge_descriptor e; bool exists; tie(e, exists) = edge(u, v, g.m_g); return std::make_pair(e, exists && g.m_edge_pred(e)); } Here, the line tie(e, exists) = edge(u, v, g.m_g); returns _some_ edge, but not necessarily the _right_ one. The call edge(...) is not called for the filtered graph and g.m_edge_pred(e) is called _after_ getting the wrong edge (resulting in a return value FALSE). One solution would be to call boost::edge_range(A, B, myFilteredGraph) on my filtered_graph to get all visible edges, but this function is not provided. Any ideas how to solve my problem? Michael - -- http://www.ive.uni-hannover.de # kettner(a)ive.uni-hannover.de -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE9TqUAkCdGnb0kVFMRAlsLAJwKJJ2NSLRu2HUv/H4sAhr4W5WdkACdHXZ6 3kz8fBDvysWKc1vxd5rf/CE= =l+NV -----END PGP SIGNATURE-----

2 3

BGL: dijkstra_shortest_paths and parallel edges
by Michael Kettner 06 Aug '02

06 Aug '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, my former posting leads me to another question: How does the dijkstra_shortest_paths behave if applied to a multigraph? And how does the dijkstra_shortest_paths behave if applied to a filtered_(multi)graph containing parallel edges, but with only _one_ of the parallel edges visible respectively? Michael - -- http://www.ive.uni-hannover.de # kettner(a)ive.uni-hannover.de -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE9TqprkCdGnb0kVFMRApGUAJ9nzr3Y4ijbrNnWENzxYIwcdLy8iQCfX2fx dl0M/IUt2bWFL0H9ovwajm4= =+FTJ -----END PGP SIGNATURE-----

2 1

BGL example boost_web_graph.expected
by Arjun Bijanki 06 Aug '02

06 Aug '02

When I compile and run Boost_1_28_0/libs/graph/example/boost_web_graph.cpp my output differs from the expected output: 1c1 < The diameter of the boost web-site graph is 2 --- > The diameter of the boost web-site graph is 3 Is the expected file current? Thanks, Arjun Bijanki Microsoft VC++ [Non-text portions of this message have been removed]

2 1

Why stop building with error "Thread support is unavailable!"
by xuejm 06 Aug '02

06 Aug '02

Hi, I installed boost1.28.0 with microsoft visual c++ .NET and regex worked very well, but when I build a very simple example to test boost.thread, it stop building with error "Thread support is unavailable!". I found out that it's because that there's not a pre-define variable defined. Can I define the variable? How can I use boost.thread in my vc++ .net? TIA Phipps Xue

2 1

Re: [Boost-Users] Re: Logical AND
by dan.ellis＠sophos.com 05 Aug '02

05 Aug '02

On 08/03/2002 05:04:26 PM news wrote: > > I need a match on "a AND is" when searching a sentence like "This is a > > test". >Regular expressions match sequences of characters. Your match says to match >a string when both of 2 different matching possibilities exist anywhere in >the string. To do this you must use 2 separate regular expression matches >separating them by the normal C++ "&&" operator. Maybe I'm misunderstanding what's wanted, but wouldn't is.*a do what he wants? Or if they need to be matched in either order: (is.*a)|(a.*is) It doesn't scale well, of course. -- Dan Ellis, Software Engineer, Sophos Anti-Virus Email: dan.ellis(a)sophos.com, DDI: 01235 540204 Real Business/CBI Growing Business Awards: Company of the Year Tel: 01235 559933, Web: http://www.sophos.com/

2 1

uBLAS Observation: terrible execution speed for prod(matrix,prod(matrix,prod(matrix,matrix)))
by Walter Bowen 05 Aug '02

05 Aug '02

Hi, I am new to uBLAS. I am looking to see if it will be useful for my application (I develop 6-Degrees-of-freedom flight simulations used to analyze avionics systems.) I have benchmarked a few lines of code with RogueWave's Math.h++ (testing for execution speed) and made the following observation. Observation 1) For matrix products of two small (~3-by-3) matrices ( i.e., prod(smallMat1, smallMat2) ) uBLAS executes MUCH FASTER than Math.h++. Observation 2) For matrix produces of two large matrices (100 by 100) uBLAS and Math.h++ give similar execuation speed. Obervation 3) (This is the kicker...) for strung-together matrix products like "Mat1 * Mat2 * Mat3 * Mat4 * Mat5", which I am implementing as "prod(Mat1, prod(Mat2, prod( Mat3, prod(Mat4,Mat5) ) ) ) Math.h++ is MUCH MUCH FASTER than uBLAS. -- by ~2 orders OF MAGNITUDE! I believe that this is a characteristic of the scalability of template expressions. My question is this: Is there a way I an instruct uBLAS not to use template expressions when for strung-together matrix products? Regards, Walt

5 4