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

iterators
by jens-theisen＠gmx.de 09 Jan '06

09 Jan '06

Hello, Say a library programmer wants to make a virtual interface to his class with a function that returns a collection of things. Usually one end up writing something silly like class interface { ... virtual std::vector< ... > give_stuff() const; }; which has some disadvantages, the most important being: - The implementing function must calculate the whole result, even if some of it might not be needed. Especially infinite collections of things are not possible in this way. - The vector is a container and that's a poorly defined concept. As a consequence, we can't do very much with it. For instance, you can't write give_stuff() + give_more_stuff() to concatenate. Though it's easy to implement that, we really don't want to do this on the basis of containers in general and we also don't want to do it again and again in every single project for std::vector in particular. - As the library gets used, it's poor interface is carved in stone, and so are the other two problems for it. The first problem can be addressed by using an iterator to some container. We obviouly can't use an iterator to a vector since that wouldn't solve any problems. It has to be a custom iterator class lib_iterator_impl : ... { ... virtual ... dereference() const; virtual ... advance() const; }; class lib_iterator { ... lib_iterator(lib_iterator_impl const* impl); }; class interface { ... virtual lib_iterator give_stuff() const; }; where lib_iterator_impl than has to be extended by the client of the library. Actually, there is no real need for lib_iterator to be a proper iterator in the standard sense. Especially, such iterators could be aware of there ends (or even their beginnings) to make it unnecessary to have always two of them. Both the virtualisation and halving the number of those iterators would also reduce the complexity of error messages. On such a framework, many small building blocks could be build, like concatenation, filtering, mapping, cartesian products, etc. In my opinion, at least the very low-level parts of such a framework are so important that they belong in boost. My question at this point is in what regard people share this opinion and to what extend we already have these things. I have heard of any_iterator in the past, but I can't find anything about it in the boost malings list archives or google, which would be related to this. Other than this, I'm not aware of anything in this regard. Cheers, Jens

4 3

Re: [Boost-users] Please don't reply to digests
by Phil Endecott 08 Jan '06

08 Jan '06

David Abrahams wrote: > Please don't reply to digests. > > http://www.boost.org/more/discussion_policy.htm#quoting Which says: Do not reply to digests if you are a digest delivery subscriber. Your reply will not be properly threaded and will probably have the wrong subject line. Instead, you can reply through the GMane web interface. The GMane interface doesn't work for posting messages, I've tried it; the mailing list rejects messages from non-subscribers, including those that come via GMane. There is a trade-off here: - On one hand, replying to digests means that threaded views using references: headers will not work properly. (There is no need for the subject line to be wrong, but it isn't difficult to forget to change it when replying.) - On the other hand, prohibiting digest subscribers (like me) from sending replies means that you won't hear from me again. You may or may not consider that a good thing. I do try to reply to questions on the rare occassions that someone even newer to Boost than me asks a question that I have an answer to, and I do so in the hope that this lets the more expert readers (Dave!) concentrate on more important things. I'd like to encourage a debate - which is considered more important, better threading or more contributions? If you don't want to hear from me again, that's fine, I'll unsubscribe until I next have a question of my own to ask. Regards, --Phil.

2 1

installing BGL-python
by Jeff Abrahamson 08 Jan '06

08 Jan '06

I'm installing the Boost Graph Library python bindings on a Debian testing machine. BGL-python appears not to be available as a .deb, so I am installing from source. http://www.osl.iu.edu/~dgregor/bgl-python/ The readme suggests I run $BOOST_ROOT/tools/build/jam_src/build.sh. This directory exists, but the file does not. It also suggests I run "bjam <options>". I can't think of any options I want a priori [ ;-) ] so I run bjam with no options, but to no avail: jeff@krypton:bgl-python-0.9 $ bjam Unable to load Boost.Build: could not find "boost-build.jam" --------------------------------------------------------------- Attempted search from /home/jeff/tmp/bgl-python-0.9 up to the root and in these directories from BOOST_BUILD_PATH and BOOST_ROOT: /usr/share/boost-build. Please consult the documentation at 'http://www.boost.org'. jeff@krypton:bgl-python-0.9 $ BTW, Boost.Build does not exist on my machine. I'm not clear on what should cause it to exist. Googling largely points me to the above. I'm a bit confused, clearly. Can anyone offer a pointer? Thanks much. -- Jeff Jeff Abrahamson <http://www.purple.com/jeff/> +1 215/837-2287 GPG fingerprint: 1A1A BA95 D082 A558 A276 63C6 16BF 8C4C 0D1D AE4B

5 5

[Boost.Python] newbie: Lifetime of objects
by Sergey Sikorskiy 08 Jan '06

08 Jan '06

Hello everyone. I'm trying to map into python classes shown below. class A { friend class B; public: ~A(void) {} private: A(B& b) : m_B(&b) { } B* m_B; }; class B { friend class C; public: ~B(void) {} A* make_a(void) { return new A(*this); } private: B(C& c) : m_C(&c) { } C* m_C; }; class C { public: ~C(void) {} B* make_b(void) { return new B(*this); } }; C* make_c() { return new C(); } Objects are supposed to be created in the order C, B, A and destroyed in the order A, B, C. I do not want to map constructors of A, B and C. I can see three ways to do what I want: 1) class_<A>("A", no_init); class_<B>("B", no_init) .def("make_a", &B::make_a, return_internal_reference<>()) ; class_<C>("C", no_init) .def("make_b", &C::make_b, return_internal_reference<>()) ; def("make_c", make_c, return_value_policy<manage_new_object>()); 2) class_<A>("A", no_init); class_<B>("B", no_init) .def("make_a", &B::make_a, return_value_policy<manage_new_object>()) ; class_<C>("C", no_init) .def("make_b", &C::make_b, return_value_policy<manage_new_object>()) ; def("make_c", make_c, return_value_policy<manage_new_object>()); 3) class_<A, shared_ptr<A>, noncopyable>("A", no_init); class_<B, shared_ptr<B>, noncopyable>("B", no_init) .def("make_a", &B::make_a, return_value_policy<manage_new_object>()) ; class_<C, shared_ptr<C>, noncopyable>("C", no_init) .def("make_b", &C::make_b, return_value_policy<manage_new_object>()) ; def("make_c", make_c, return_value_policy<manage_new_object>()); C::make_b and B::make_a are supposed to return shared_ptr<B> and shared_ptr<A> respectively in this case. After debugging this code I've got strange results: - In case 1 objects aren't get deleted at all; - In case 2 objects are get deleted in wrong order; - In case 3 objects are get deleted in wrong order; Cases 2 and three seem to be identical to me. The only difference I can see is that in case 3 I should manage objects lifetime by myself from C++ code. In opposite, in case 2 objects lifetime will be managed by python. I couldn't get an answer to my problem neither from documentation nor from Boost.Python tests nor from Google. Any explanations or suggestions will be greatly appreciated. Thank you. -- Sergey Sikorskiy

2 1

Re: [Boost-users] Boost-users Digest, Vol 760, Issue 2
by Ryan McGann 08 Jan '06

08 Jan '06

On Dec 19, 2005, at 9:00 AM, boost-users-request(a)lists.boost.org wrote: >> 1) The build.sh file for the boost bjam tool in the 1.33.0 and 1.33.1 >> release seems to have DOS line endings. I know there are prebuilt >> binaries of bjam up on the Web, but our build system builds it >> manually, and it choked on Darwin because sh didn't like the DOS line >> endings. I resaved the file with Unix line endings and it built fine. > > Did you download the tar.gz distribution? That should have Unix line > endings throughout where as the .zip file has DOS line endings > throughout > (intensionally). Well our build system uses whichever archive is checked in. I imagine the PC folks downloaded the .zip version. I can check in a different version for the Unix folks. >> 2) I get errors for multiply defined symbols for the placeholder >> symbols declared in boost/bind/placeholders.hpp. Every file that >> #include <boost/bind.hpp> contains actual instances of _1, _2, etc. >> Changing the declarations of these in boost/bind.hpp to static (I >> added a __APPLE__ to the #ifdef in the middle of the file) fixes this >> issue, but then I get warnings for unused static symbols if I don't >> use the placeholder objects. This issue in particular is quite >> annoying. > > What compiler are you using? The first #if says: > > #if defined(__BORLANDC__) || defined(__GNUC__) > > isn't this the branch that gets taken? Hm in my boost/placeholders.hpp, I see #if defined(__BORLANDC__) ... #elif defined(BOOST_MSVC) || (defined(__DECCXX_VER) && __DECCXX_VER <= 60590031) || defined(__MWERKS__) ... #else ... #endif I don't have a _GCC__ anywhere in there. This is from 1.33.1 (which I just newly expanded from the .tar.gz, not the .zip). If I add a defined(__GCC__) to the first or second set of #ifdefs, it works fine. >> 3) The header in boost/compatibility/cpp_c_headers/cstdio >> does not match up with the cstdio file included with the Mac OS X dev >> tools. >> In particular it is missing a using ::snprintf, which causes problems >> because this cstdio file is being included BEFORE the Mac OS X >> version of the same file, and other C++ header files rely on >> std::snprtinf being declared. I fixed this at first by deleting the >> boost/compatibility/cpp_c_headers folder, but then ending up just >> deleting that single file so that OS X always used the built-in >> header file. > > Why are those headers getting included? They were intended for use > only > with certain broken compilers that didn't have the c-prefixed std lib > headers. They're known not to work with several std conforming > compilers > (but shouldn't be included for these anyway). I'm not sure why they're being included. I'm guessing it's because we pass -I<path to boost> to GCC, which causes it to search inside of that directory before the standard directories get searched. I'll have to investigate that further. Thanks, Ryan

3 2

[Graph] Maximum common subgraph
by Andreas Sæbjørnsen 08 Jan '06

08 Jan '06

Hi, Lately I have been looking for a good C/C++ implementation of an algorithm for finding the maximum common subgraph in Boost::Graph. The algorithms I have looked for is mainly Bron-Kerbosch and the MCS algorithm, but I am very open for alternatives. The implementation must be able to handle directed graphs containing self-loops (it is possible to have two edges at the same time between two nodes i->j, j->i). Do you know of any implementation? Thanks Andreas Saebjornsen

6 9

Bjam problems with intel compiler on Windows when building Boost1.3.2
by Miller, Peter 08 Jan '06

08 Jan '06

Hi, list members. I am trying to build Boost (1_32_0) using intel/Windows compiler, but no dice. Following the steps in the "Getting started" guide, I chose to set up the environment variables, i.e. approach 2.1 So I pull up a command window, navigate to the intel directory, and run iclvars to establish the intel directory is in the PATH. It adds a bunch of items to the path, so it seems that worked. Then, I navigate to boost\boost_1_32_0\boost\1_32_0 where bjam.exe lives. I use it as-received, as it seemed there was no point in building it locally. Invocation to make Boost libraries I used was: bjam "-sTOOLS=intel-win32" install What I get is odd: first, a notice: skipping Boost.Python library build due to missing or incorrect configuration couldn't find python.h in "c:/tools/python/include" THis is puzzling to me: there is no Python setup on this machine - and why should Boost (a c++ package) be looking to create Python libraries? anyhow, there is a bit more info about various PYTHON_xxxx variables that I could set to configure the python installation. Then, it gets into the c++ compilation, and all these fail. The invocation that is shown in the command window is (for example): vc-C++ bin\boost\libs\date_time\build\libboost_date_time.lib\intel-win32 \debug\runtime-link-static\greg_month.obj '/Zm800' is not recognized as an internal or external comand, operable program or batch file. /Zm800 -nologo /EHsc -c -DBOOST_DATE_TIME_STATIC_LINK .Z7 etc. To me (who doesn't understand bjam), it seems that bjam has invoked the VC compiler rather than the intel compiler (which has an invocation of icl, IIRC). Anyhow, it looks like the compiler is hollering about the /Zm800 option, which is not a legal option for the intel compiler, though it is for VS 7.1. This error then repeats several hundred times, once for each file for which compilation is attempted. I suspect there is some very simple thing I am overlooking, but I've tried to follow every step in the "Getting Started" guide. Does anyone suggestions on what is wrong here? I'm stumped, and would appreciate any help or thoughts on how to chase this further. Thanks! Peter

2 1

boost and VC++ toolkit 2003
by Raphael Bousquet 08 Jan '06

08 Jan '06

I was wondering the Visual C++ Toolkit 2003 was going to be a future platform for boost. I believe no port is necessary just the appropriate bjam configuration. I have tried using the "vc-7_1" option but even though most of the libraries are created, when you try to use them, files such as msvcptr.lib are requested (though this does not generate any error during the compilation)... Maybe there is a workaround already? If anyone has any experience with that... The best link I have found so far is: http://groups.google.com/group/boost-list/browse_thread/thread/d9e9ba8e6c9c5 fa0/7a963d8e2eee9a4d?lnk=st&q=vc%2B%2B+toolkit+boost&rnum=1&hl=en#7a963d8e2e ee9a4d Thanks

3 4

Problem with program_options.
by Simen Kvaal 06 Jan '06

06 Jan '06

Dear all, I have problems with the program_options library. I am using boost 1.33.1 and gcc 3.2.3, and I notice that there are others with the same problem. However, I cannot find a reference to with a solution anywhere. Stated simply: My programs (even the example programs included with boost 1.33.1) using program_options do not link! For example (gcc 3.2.3): $ g++ -o first.x first.cpp /tmp/ccM0BLhg.o(.text+0x4a): In function `main': : undefined reference to `boost::program_options::options_description::options_description(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, unsigned int)' [ ... and a dozen more similar errors ... ] (As far as I can see, all header files in boost/program_options/ are included in boost/program_options.hpp.) I would appreciate any help with this. Regards, Simen Kvaal. -- ---- Simen Kvaal -- Ph.D student --------+------------------------- Centre of Mathematics for Applications | cell: +47 90199552 Room 1039, N.H.Abel's House, UiO | office: +47 22855681 http://folk.uio.no/simenkva/ | simen.kvaal(a)cma.uio.no

4 3

iostreams: example code of walk into zip file's filename/subdir
by David Xiao 06 Jan '06

06 Jan '06

Hi Friends: I am looking for a piece of code which can: input a .zip file name and provide a callback facility which allow me walking in file structure, as well as survive from corrupted zip file. I had read the example code and test code of zlib part in boost::iostreams, it extract data like this: //-- copy from iostreams/filter/test.hpp --// template<typename OutputFilter, typename InputFilter> bool test_filter_pair( OutputFilter out, InputFilter in, const std::string& data, mpl::true_ ) { for ( int inc = default_increment; inc <= default_increment * 40; inc += default_increment ) { array_source src(data.data(), data.data() + data.size()); std::string temp; std::string dest; iostreams::copy(src, compose(out, non_blocking_sink(temp, inc))); iostreams::copy( compose(in, non_blocking_source(temp, inc)), iostreams::back_inserter(dest) ); if (dest != data) return false; } return true; } I think it is not gentle because it try catch all data into memory. So there must be some better way to do? Rgds, David Xiao

1 0