August 2004 - Boost-users - lists.preview.boost.org

Regex help needed, please help.
by Kosta Adarchenko 15 Aug '04

15 Aug '04

boost-users-request(a)lists.boost.org wrote: Send Boost-users mailing list submissions to boost-users(a)lists.boost.org To subscribe or unsubscribe via the World Wide Web, visit http://www.lists.boost.org/mailman/listinfo.cgi/boost-users or, via email, send a message with subject or body 'help' to boost-users-request(a)lists.boost.org You can reach the person managing the list at boost-users-owner(a)lists.boost.org When replying, please edit your Subject line so it is more specific than "Re: Contents of Boost-users digest..." Today's Topics: 1. Re: Regex problems, please help (John Maddock) ---------------------------------------------------------------------- Message: 1 Date: Fri, 13 Aug 2004 11:39:54 +0100 From: "John Maddock" Subject: Re: [Boost-users] Regex problems, please help To: Message-ID: <01d401c48125$ee4e79f0$3c520352@fuji> Content-Type: text/plain; charset="iso-8859-1" > After some seven thousands hits the match algo throughs the memory exhausted exception. > So I cought the exception and moved the iterator by one, printed the location it points to in my searched file to view the problematic section for the regex find algorithm. > I have noticed that the problem happens not with all large files, but with quite a bit of them. > After I have received the problematic section in my file, I ran the program on it and it did not raise any exceptions. > Please , help me I have a Graduate project to complete within 3 weeks from now, and this is one of the major bugs I have not resolved. You don't say *exactly* what the regex is that's causing the problem - it makes a difference - problems like this are usually caused by an expression that behaves badly with specific input (because regexes are NP-Hard to match in the general case). Boost.Regex will throw an exception rather than go on thrashing forever trying to find a match, but to determine whether this is the case here, you need to trap the exception at the point it's being thrown, and see if the regex lib is indeed the culprit, if it is, then refactoring the expression that you're using so that it's less ambiguous is usually the answer. John. ------------------------------ _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://www.lists.boost.org/mailman/listinfo.cgi/boost-users End of Boost-users Digest, Vol 279, Issue 2 ******************************************* John, I have stated the exact regex, it's \d\d\d\d\d\d , a simple regex matching 6 numbers, but the same happens even with one \d regex. I don't see a way to refactor the \d expression. --------------------------------- Do you Yahoo!? New and Improved Yahoo! Mail - Send 10MB messages!

2 1

[boost-users]_BGL:_copy_component_bugs
by Erik Jenelius 13 Aug '04

13 Aug '04

Hi. In boost/graph/copy.hpp, besides the copy_graph function, I noticed a nice, undocumented function called copy_component which uses breadth-first search to copy a component from a graph to a new graph. Unfortunately, it doesn't work. I think there are at least two bugs in it: 1. The vertices are copied at the examine_vertex event point and the edges are copied at the examine_edge event point. However, the target of the edge has not yet been examined at that point, which leads to strange results. If the edges are copied at the black_target event point instead, the algorithm seems to work correctly. 2. The std::vector orig2copy is templated on the vertex descriptor type of the input graph, while it should be the output graph. The function copy_graph is correct here. Also, both copy_graph and copy_component by default assume the vertices have a vertex_all property and the edges an edge_all property, which are copied to the output graph. Wouldn't it be more intuitive to not assume and copy any properties by default? - Erik Jenelius

1 0

Regex problems, please help
by Kosta Adarchenko 13 Aug '04

13 Aug '04

Here is the problem: I want to find a simple regular _expression something like \\d\\d\\d\\d\\d\\d in a large file (15 mb) using VC++ 7.1 . I use Spirit's File Iterator as the input iterator for the search algorithm. After some seven thousands hits the match algo throughs the memory exhausted exception. So I cought the exception and moved the iterator by one, printed the location it points to in my searched file to view the problematic section for the regex find algorithm. After the "problematic section" of some 300 chars ends the algorithm keeps working fine. here is the code. //------------------------------------------------------------------------------------------- void TheFunction() { //print the problematic section here ofstream of; of.open("d:\\err1.txt"); boost::spirit::file_iterator< > begin("d:\\log1.txt"); boost::spirit::file_iterator<> beginning_saved(begin); //save the beginning in order to compare later boost::spirit::file_iterator< > ending; ending= begin.make_end(); boost::match_results<boost::spirit::file_iterator<char > > what; boost::match_flag_type flags = boost::match_default; long matchescounter=0; std::exception e; int i; bool found; found=false; try { found=boost::regex_search(begin, ending, what, _expression, flags); if (found) { begin =what[0].second; cout<<what[0]; matchescounter=1; } } catch (std::exception &e) { cout<<e.what(); } while(found) { try { found=boost::regex_search(begin, ending, what, _expression, flags); begin =what[0].second; if (found) { cout<<matchescounter<<"----------"<<what[0]<<"-------------------"<<what[0].second-beginning_saved<<"\n"; matchescounter++; } } catch (std::exception &e) { cout<<*begin; of<<*begin; begin+=1; } } cout<<"---------------the end------------\n"; cout<<matchescounter; of.close(); } //--------------------------------------------------------------------------------------- I have noticed that the problem happens not with all large files, but with quite a bit of them. After I have received the problematic section in my file, I ran the program on it and it did not raise any exceptions. Please , help me I have a Graduate project to complete within 3 weeks from now, and this is one of the major bugs I have not resolved. Yours truely, Kosta. Desparatly waiting for your reply. --------------------------------- Do you Yahoo!? Y! Messenger - Communicate in real time. Download now.

2 1

RE: [Boost-users] polymorphic container/boost::any
by Klaus Nowikow 13 Aug '04

13 Aug '04

> dp12-ipvi(a)spamex.com wrote: > > Use std::list<boost::shared_ptr> std::list<boost::shared_ptr<tree> > to be exact ;-) > > I hope I spelled all that correctly. > > ++PLS -- Klaus > > -----Original Message----- > From: goochrules! [mailto:goochrules@gmail.com] > Sent: Thursday, August 12, 2004 2:02 PM > To: boost-users(a)lists.boost.org > Subject: [Boost-users] polymorphic container/boost::any > > * Replies will be sent through Spamex to boost-users(a)lists.boost.org > * For additional info click -> http://www.spamex.com/i/?v=4259897 > I'm quite sure that this problem has been addressed many > times before, so I > suspect it as a better name and I suspect that boost::any could be the > solution. > I want to store related types in the same stl container, for > instance, this > is the class heirachy I want: > class tree { public: void grow() =0; }; > class oak : public tree { public: void grow() { std::cout << > "oak grows > slowly" << std::endl; }; class pine : public tree { public: > void grow() { > std::cout << "pine grows quickly" << std::endl; }; Then, I > try to put them > in a std::list: > std::list< tree > forest; > forest.push_back(oak()); > forest.push_back(pine()); > ...and I find out that I can't define tree::grow as abstract. > So I give it > a dummy implentation -- void grow() { std::cout << "tree grows" << > std::endl; --, and try to iterate over the objects: > for(std::list< tree >::iterator iter = forest.begin(); > forest.end() != iter; > ++iter) { > iter->grow(); > } > ...and I don't get what I expect: > tree grows > tree grows > I then think about trying to use a std::list< tree& > and > discover that it > won't compile. I think briefly consider std::list< tree* > > but decide this > dosen't meet requirements of memory management. > So I do some searching for 'polymorphic container' and 'polymorphic > iterator' and come up with boost::any, but I've no luck in > making it work. > Attached is my code. > In the end, I'm looking for a way to add related types to a > stl container > and not have to do anything complicated to call a method on the type's > abstract interface and to not have to wory about memory > management. Maybe > this question is better suited for comp.lang.c++[.moderated]? > -- Matthew Peltzer > -- goochrules(a)gmail.com > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

1 0

compiling boost using Intel C++ 8.0 under Windows
by Lars Schouw 13 Aug '04

13 Aug '04

I am havng problems building boost using the intel C++ 8.0 compiler. It looks as if bjam can't fine the icl.exe compiler executable itself. This file is installed under this directory C:\Program Files\Intel\CPP\Compiler80\Ia32\Bin Any ideas what I do wrong? I try this: run the iclvars.bat batch file. Then I use bjam like this: C:\dev\boost\boost_1_31_0_icc>bjam "-sTOOLS=intel-win32" "-sINTEL_VERSION=80" "-sINTEL_BASE_MSVC_TOOLSET=vc7.1" But when compiling I get these errors from each command to the compiler: /Zm800 -nologo -GX -c -DNDEBUG -DNDEBUG -DBOOST_DATE_TIME_POSIX_TIME_STD_CONFIG -DBOOST_ALL_DYN_LINK=1 /Ogity /O2 /Gs /Ob2 /GX /GR /MD /Op /Zc:forScope /Zc:wchar_t /Qwn5 /Qwd985 -Qoption,c,--arg_dep_lookup /Qansi_alias -D_NATIVE_WCHAR_T_DEFINED /Qvc7.1 -I"bin\boost\libs\date_time\build" -I"C:\dev\boost\boost_1_31_0_icc" -I"C:\Program Files\Microsoft Visual Studio .NET 2003\VC7\include" -Fo"bin\boost\libs\date_time\build\boost_date_time.dll\intel-win32\release\greg_weekday.obj" -Tp"C:\dev\boost\boost_1_31_0_icc\libs\date_time\build\../src/gregorian/greg_weekday.cpp" ...failed vc-C++ bin\boost\libs\date_time\build\boost_date_time.dll\intel-win32\release\greg_weekday.obj... vc-C++ bin\boost\libs\date_time\build\boost_date_time.dll\intel-win32\release\date_generators.obj '/Zm800' is not recognized as an internal or external command,operable program or batch file. Lars --------------------------------- Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish.

1 0

RE: [Boost-users] polymorphic container/boost::any
by dp12-ipvi＠spamex.com 12 Aug '04

12 Aug '04

Use std::list<boost::shared_ptr> I hope I spelled all that correctly. ++PLS -----Original Message----- From: goochrules! [mailto:goochrules@gmail.com] Sent: Thursday, August 12, 2004 2:02 PM To: boost-users(a)lists.boost.org Subject: [Boost-users] polymorphic container/boost::any * Replies will be sent through Spamex to boost-users(a)lists.boost.org * For additional info click -> http://www.spamex.com/i/?v=4259897 I'm quite sure that this problem has been addressed many times before, so I suspect it as a better name and I suspect that boost::any could be the solution. I want to store related types in the same stl container, for instance, this is the class heirachy I want: class tree { public: void grow() =0; }; class oak : public tree { public: void grow() { std::cout << "oak grows slowly" << std::endl; }; class pine : public tree { public: void grow() { std::cout << "pine grows quickly" << std::endl; }; Then, I try to put them in a std::list: std::list< tree > forest; forest.push_back(oak()); forest.push_back(pine()); ...and I find out that I can't define tree::grow as abstract. So I give it a dummy implentation -- void grow() { std::cout << "tree grows" << std::endl; --, and try to iterate over the objects: for(std::list< tree >::iterator iter = forest.begin(); forest.end() != iter; ++iter) { iter->grow(); } ...and I don't get what I expect: tree grows tree grows I then think about trying to use a std::list< tree& > and discover that it won't compile. I think briefly consider std::list< tree* > but decide this dosen't meet requirements of memory management. So I do some searching for 'polymorphic container' and 'polymorphic iterator' and come up with boost::any, but I've no luck in making it work. Attached is my code. In the end, I'm looking for a way to add related types to a stl container and not have to do anything complicated to call a method on the type's abstract interface and to not have to wory about memory management. Maybe this question is better suited for comp.lang.c++[.moderated]? -- Matthew Peltzer -- goochrules(a)gmail.com

1 0

Random Library
by Andy Tompkins 12 Aug '04

12 Aug '04

Hi, I have a question about the random library. I'm using vc7.1 and I'm getting a warning when I include the random library. The warning I get when I, #include <boost\random.hpp> is: d:\includes\boost_1_31_0\boost\random\linear_congruential.hpp(233) : warning C4244: 'return' : conversion from 'boost::uint64_t' to 'boost::int32_t', possible loss of data When I look at the file, rand48::operator() returns an int32_t, but lcf is defined with uint64_t. Thanks in advance for any help. Andy.

1 0

RE: [Boost-users] Regex problem
by Evgeny Benediktov 12 Aug '04

12 Aug '04

It was the case. Thank you. Evgeny Benediktov -----Original Message----- From: John Maddock [mailto:john@johnmaddock.co.uk] Sent: Thursday, August 12, 2004 1:04 PM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Regex problem >When extensive usage of regex_match and regex_search functions is done numerous times, then after several thousands iterations >Regex library starts throwing "Regular expression too big" exception on the same regular expressions it worked perfectly with at >several thousands previous iterations. >Looks like Regex fails to do some vital clean up. >If anyone saw this problem before, or has any suggestion about it, this may be very helpful. It's most likely that those exceptions are coming from the matching algorithms (regex_match/regex_search)? If so then it's because the combination of the particular regular expression, and the particular text being matched, were becoming "pathological", so an exception is thrown to prevent the match taking forever to obtain (regexes are NP-Hard to match in the worst case). Often these issues can be resolved by refactoring the regular expression, but you really need to track down which specific expression, against what specific text is causing the problem - try setting your debugger to break when an exception is thrown, and then work back up the stack to see what caused the issue. John Maddock. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users ***************************** Finjan Software This e-mail and any attached files are confidential and may be legally privileged. The unauthorized use, disclosure or copying of this email or any information contained within it is strictly prohibited. This also confirms that Finjan Software's Vital Securityfor E-Mail has scanned this message for the presence of known viruses and potentially malicious code. Finjan Software - Prevention is the Best Cure! *****************************

1 0

uBLAS - How to use a range of a matrix row?
by Pascal Bauer 12 Aug '04

12 Aug '04

Hi, I have some (strange?) problems with the ublas: I'm writing a SPH-Code for fluids and brittle solids. I use a uBLAS matrix to store all the particle-Data. Now, since the SPH code itself includes a lot of physics, i wrote a "particle" class, that give me access to the values stored in the matrix in an intuitive way. As template argument, i'll give a matrix, so i do no have to take care about matrix internals. It looks somehow thatlike: template <typename T> class Particle { static T& SPHData; //Matrix where the values are stored typename T::size_type GID; //Particle ID (...) matrix_row<T> LData; // Proxy to the particle data vector_range<matrix_row<T> > Pos; // Position vector, for // calculations... (...) typename T::value_type& mass; // Reference to physical values... Particle(typename T::size_type ParticleID); // Constructor, needs // Particle ID } template <typename T> Particle<T>::Particle(typename T::size_type ParticleID) : mass(SPHData(ParticleID, MASS)) // Set the references... { LData.reset(SPHData, ParticleID); // OK, works fine... Pos.reset(LData, range(0, 2)); // Give strange error message: // (at the end of this mail...) } I tried different methods to get the Pos to the right place in my matrix, but if i use the "=" operator, i've got some strange run-time assertion errors. My thing is, that i'd like to have a range of a matrix_row or column, that i can use later for vector-operations. How can I do that? - Pascal Bauer PS: Compile Error for my example code. */home/pbauer/oosph/src/particlecontainer.h:103: instantiated from `sph::Particle<T>::Particle(typename T::size_type) [with T = ParticleTest()::Matrix]' */home/pbauer/oosph/src/test.cpp:36: instantiated from here */usr/include/boost/numeric/ublas/vector_proxy.hpp:126: error: no matching function for call to ` boost::numeric::ublas::matrix_row<boost::numeric::ublas::matrix<double, boost::numeric::ublas::row_major, boost::numeric::ublas::unbounded_array<double> > >::reset( boost::numeric::ublas::matrix_row<boost::numeric::ublas::matrix<double, boost::numeric::ublas::row_major, boost::numeric::ublas::unbounded_array<double> > >&)' */usr/include/boost/numeric/ublas/matrix_proxy.hpp:112: error: candidates are: void boost::numeric::ublas::matrix_row<E>::reset(M&) [with M = boost::numeric::ublas::matrix<double, boost::numeric::ublas::row_major, boost::numeric::ublas::unbounded_array<double> >] */usr/include/boost/numeric/ublas/matrix_proxy.hpp:117: error: void boost::numeric::ublas::matrix_row<E>::reset(M&, typename M::size_type) [with M = boost::numeric::ublas::matrix<double, boost::numeric::ublas::row_major, boost::numeric::ublas::unbounded_array<double> >] -- \ | Pascal Bauer | | Space Research & Planetary Sciences | Research Division | | University of Bern | | Phone: +41316314416 | | e-mail: mailto:pascal.bauer@phim.unibe.ch /

1 0

RE: [Boost-users] Feature for using C++ as script language
by Persson Jonas 12 Aug '04

12 Aug '04

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org]On Behalf Of Jeff Garland > Sent: den 11 augusti 2004 17:19 > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] Feature for using C++ as script language > > > On Wed, 11 Aug 2004 11:33:50 -0300, Maurício wrote > > Hi, > > > > After I discovered Boost, which has regex and filesystem > > libraries, I decided to use C++ for tasks that I usually do with > > Ruby (a Perl-like language), because I know C++ a lot better than > > other languages. However, there's a feature of those languages I > > miss: the ability to call external programs and get the output as a > > string. Does any of Boost libraries (or other free libraries you > > know) have this capability? > > There isn't a boost library, but it would be a nice addition. > Turns out the > 'system' call is available on both windows and unix so you > can do something > like this: > > #include <stdlib.h> > > std::string command = "...your command here..." > std::string tmpfile_name("...your temporary file name here..."); > std::string cmd = command + "> " + tmpfile_name; > system(cmd.c_str()); > std::ifstream output(tmpfile_name.c_str()); > > Of course this has many limits an issues. If you want > something more formal > pstreams might work for you: > > http://pstreams.sourceforge.net/ > > HTH, > > Jeff > Another one is here: http://libexecstream.sourceforge.net/ / Jonas

1 0