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

Unexpected result in parsing a regular expression
by abhinkar 27 Oct '02

27 Oct '02

Hi, I recently downloaded the boost regex library and am trying to use the search and replace (regex_merge) API. But I am not getting the expected result. Here is what I am trying to do. I am trying to replace a single quoted string with something. But, the catch here is that the string in single quotes can have zero or many 2 consecutive single quotes. For example: Input (Ignore the angle brackets>: <foo='string with '' consecutive '' single ''quotes' bar='string with no single quotes'> To be replaced by 'XXX'. So the output that I want: <foo=XXX bar=XXX> Here is the regular expression that I wrote in Perl: /(\'[^']*(\'\')*[^']*\')/ I get the expected result in Perl. So, to test this with regex library I wrote a sample program to use the regex_merge API and try the replace functionality. ---------------------------replace.cc-------------------------------- #include <boost/regex.hpp> #include <string> #include <iostream> #include <sstream> #include <iterator> using namespace std; using namespace boost; const std::string formatstr("XXX"); std::string processexp(string resp) { regex expression("(\'[^']*(\'\')*[^']*\')"); std::string result; std::ostringstream t; std::ostream_iterator<char, char> oi(t); cout << "Response:" << resp << endl; result = boost::regex_merge(resp, expression, formatstr, boost::format_all); return result; } int main(int argc) { std::string res; res = processexp("foo='string with '' consecutive '' single ''quotes' bar='string with no single quotes'"); cout << "results=" << res << endl; return 0; } --------------------------------------------------------------------- The output I get is: foo=REPLACEDREPLACED bar=REPLACED I tried several ways, but couldn't get the regular expression to work to return the expected result. I WOULD really appreciate if someone could help me. Thanks in advance, Sameer

2 1

vecS versus listS
by aaronyinyong 27 Oct '02

27 Oct '02

compare the following codes: 1.First I create Graph using vecS parameter: typedef boost::adjacency_list<vecS,vecS,directedS,VertexProperty,EdgeProperty> Graph; ..... E edge_array[num_edges]={E(0,1),E(0,2),E(0,3),E(2,3)}; boost::add_edge(edge_array[i].first,edge_array[i].second,g); ..... This is OK! 2.Then I changed the parameter "vecS" to "listS". typedef boost::adjacency_list<listS,listS,directedS,VertexProperty,EdgePropert y> Graph; ..... E edge_array[num_edges]={E(0,1),E(0,2),E(0,3),E(2,3)}; boost::add_edge(edge_array[i].first,edge_array[i].second,g); It can't be compiled.The VC7 tells me "int can't be converted to void*". So what's wrong? Is there any difference which could be caused by "vecS" and "listS"?

2 1

[BGL] Making GraphvizDigraph bidirectional
by kweeheong 27 Oct '02

27 Oct '02

Hi, For convenience, I am using Graphviz digraphs, reading them into BGL, running my algorithms and writing them out to view the results (with Graphviz dotty ). My compiling environment is VC6. After some compilation failures, I realized that Graphviz digraphs did not support in_edges and other functionality of the bidirectional graphs. Graphviz digraphs is currently defined in graphviz.hpp as : typedef subgraph<adjacency_list<boost::vecS, boost::vecS, boost::directedS, GraphvizVertexProperty, GraphvizEdgeProperty, GraphvizGraphProperty> > GraphvizDigraph; There are 2 potential solutions I see ( there may be others ): Solution 1 : ------------ - Read in Graphviz digraph and then copy it into a working bidirectional graph to run the algorithms. - Copy back to Graphviz digraph to view the results. Solution 2 : ------------ - Change the definition of GraphvizDigraph to : typedef subgraph<adjacency_list<boost::vecS, boost::vecS, boost::bidirectonalS, GraphvizVertexProperty, GraphvizEdgeProperty, GraphvizGraphProperty> > GraphvizDigraph; - Recompile bglviz.lib - Now GraphvizDigraph is bidirectional I prefer Solution 2 because it leverages the generic way that BGL has been written. I am able to rebuild bglviz.lib and my minimal testing shows the bidrectional GraphvizDigraph seems to work so far. I would like feedback whether there are any gotchas with Solution 2. Thanks,

2 1

RE: [Boost-Users] are examples on maual page ready to run?
by Mark Storer 25 Oct '02

25 Oct '02

Which error? What's the compiler saying? We (okay: I) need more info for a proper diagnosis. --Mark Storer Software Engineer Cardiff Software #include <disclaimer> typedef std::disclaimer<Cardiff> Discard;

1 0

are examples on maual page ready to run?
by J-S Lin 25 Oct '02

25 Oct '02

I just download boost and tried to get a hang of it by running examples out of the doc page. But I am not sure if these examples are ready to run or simply like a pseudo-code? A point in the case: This simple matrix example int main () { using namespace boost::numeric::ublas; matrix<double> m (3, 3); for (int i = 0; i < m.size1 (); ++ i) for (int j = 0; j < m.size2 (); ++ j) m (i, j) = 3 * i + j; std::cout << m << std::endl; } I copied it and added the following includes: #include <boost/numeric/ublas/matrix.hpp> #include <iostream> But the compiler (VC7) flag an error on cout. So is this line " std::cout << m << std::endl;" supposed to work, or it is a pseudo-code. Thanks. J-S Lin

1 0

How do I use boost::multi_array with a templated argument?
by Björn Lindberg 25 Oct '02

25 Oct '02

I cannot get the code below to compile with g++-3.2 on GNU/Linux. The error message is: exempel.cpp: In member function `void functor<T>::operator()() [with T = P]': exempel.cpp:26: instantiated from here exempel.cpp:18: `nodetab' has incomplete type exempel.cpp:18: confused by earlier errors, bailing out Since everything is visible at compile time, I can't understand why I'm getting this error message, or what to do about it. ------------- #include <boost/multi_array.hpp> struct P {}; template<typename T> struct functor { typedef boost::multi_array<T, 3> table_t; typedef typename table_t::array_view<2>::type nodetab_t; typedef typename table_t::index_range range; void operator()() { table_t table(boost::extents[5][6][7]); nodetab_t nodetab = table[boost::indices[1][range()][range()]]; } }; int main() { functor<P> vis; vis(); return 0; } ------------- Björn

1 0

Borland 6 Graph example 1
by chongwhite001 25 Oct '02

25 Oct '02

Hi All, I am attempting to run the 1st example (file dependancy) of Boost Graph library with Borland C++Builder 6. It compiles ok but crashes when run in exception.hpp with a call, or rather THROW to negative_edge exception. If anyone can provide any light into this I would be most grateful. (details below) Thanks Christian _________________________________________________________ Fault occurs in: dijkstra_shortest_paths.hpp: void examine_edge(Edge e, Graph& g) { if (m_compare(get(m_weight, e), m_zero)) throw negative_edge(); m_vis.examine_edge(e, g); } The actual fault is in exception.hpp at the point: struct negative_edge : public bad_graph { negative_edge() : bad_graph("The graph may not contain an edge with negative weight."){ } }; The error is: --------------------------- Debugger Exception Notification --------------------------- Project Boost_Ex1_Console.exe raised exception class boost::negative_edge with message 'Exception Object Address: 0x896912'. Process stopped. Use Step or Run to continue. --------------------------- OK Help ---------------------------

2 3

Fwd: uBLAS vs MTL for sparse matrices
by Paul C. Leopardi 25 Oct '02

25 Oct '02

Hi, I posted the following message to boost and have so far had no response. Maybe boost-users is the correct mailing list? Hi, Thanks to Joerg Walter's original conversion of GluCat ( http://glucat.sf.net ) to uBLAS, I have fairly painlessly been able to get GluCat going with the version of uBLAS now promoted to Boost. I'm using sparse matrices: boost::numeric::ublas::sparse_matrix< Scalar_T, boost::numeric::ublas::row_major> and comparing performance to MTL: mtl::matrix< Scalar_T, mtl::rectangle<>, mtl::compressed<>, mtl::row_major >::type The operation I'm performing is quite complicated. Basically, I am generating a large number of sparse matrices, and computing the inner product of a given matrix with each of these matrices in turn. ( If you have a copy of GluCat, see basis_element() in matrix_multi_imp.h ) Because these sparse matrices have only one non-zero per row, operations on an n x n matrix are typically O(n), and seem to be so using MTL: Size vs speed for MTL: Size Time (ms) 32 10 64 20 128 80 256 160 512 700 But with uBLAS, essentially the same operations seem to be O(n^2): Size Time (ms) Slowdown 32 30 3 64 70 3.5 128 830 10.4 256 1940 12.1 512 31770 45.4 o Is there a more appropriate data type in uBLAS? o Do you have any tips on where to tweak uBLAS to get better (ie. O(n) ) performance? o What is BOOST_UBLAS_USE_CANONICAL_ITERATOR and what effect should it have on performance? Thanks

2 1

assign a scalar to boost::numeric::ublas::matrix?
by Michael Na Li 24 Oct '02

24 Oct '02

Hi, having tried various matrix classes, I was excited to see boost finally includes a "standard" matrix class. Thanks for the good work. I have one question, though. How do I initialize a matrix to a scalar value? There is no constructor of the form 'matrix (nr, nc, value)' nor 'matrix::operator=( T value )'. I think both would be nice to have. Michael -- ---------------------------------------------------------------------------- Michael Na Li Email: lina(a)u.washington.edu Department of Biostatistics, Box 357232 University of Washington, Seattle, WA 98195 ---------------------------------------------------------------------------

1 0

[BUG] Macos.hpp configuration woes
by Bobby Thomale 24 Oct '02

24 Oct '02

Yesterday I was trying to help my wife get the Boost libraries to compile in her application. It reminded me of an issue I thought I brought up a while ago (maybe I forgot) that still exists in the boost Macos.hpp config file. In the MacOS.hpp file, it checks to see if you are compiling a Mach-O target. If you are, it reasons, you must be using the Apple BSD headers as well. This logic is even documented in a comment: // If __MACH__, we're using the BSD standard C library, not the MSL: #if __MACH__ The problem with this is that, starting with CW Pro 8, it is actually possible to build a Mach-O target and use the MSL in place of the BSD libraries. Lots of people are apparently doing this, including my wife. (I am still on CFM and have not made the transition to Mach-o yet, but I will eventually and I will potentially be in the same boat. The newer Carbon APIs are Mach-o only so that is where all the carbon programmers will eventually go, and as long as Codewarrior is around, MSL will be there as an option for people.) There are also people who use other STL implementations, such as STLPORT. When it defines all of the stuff it does in there, for the BSD libraries, the mach-o MSL code does not compile. Instead, I get hundreds of compile errors. The only way I could get this to compile for her was to force it to use the TARGET_CARBON stuff at the bottom, after the #else (meaning "not mach-o"), by commenting the __MACH__ stuff out. Obviously, that was a "see if this is really her problem" solution, also to get her up and running, and not a long-term fix. :-) It worked by the way. The config.hpp stuff seperates out your platform settings from your standard library settings. Is there a reason that standard library implementation-specific settings are being set by platform and binary format? If there is magic that needs to happen that is specific to Apple's BSD headers, shouldn't there maybe be a "BSD.hpp" or even "AppleBSD.hpp" for BOOST_STDLIB_CONFIG instead? Or am I missing something? Anyway. That is just a question/suggestion, take it or leave it. It makes sense that the TARGET_CARBON stuff at the bottom is platform specific, but the stuff above it seems like it should go somewhere else. In any case, whether that stuff is moved or not, something needs to change to alleviate the pain for those using Mach-o and MSL together, because right now boost is broken for those users. (Am I reporting this to the right list by the way?) -- Bobby --------------------------------------------------------------------- Bobby Thomale Senior Software Developer Inoveon Corporation http://www.inoveon.com/ ---------------------------------------------------------------------

3 6