- Boost-users - lists.boost.org

[serialization] Storing something that could be primitive or non-default constructable.
by Noah Roberts 13 Jul '11

13 Jul '11

template < typename T > struct storing { storing(T const& t) : val(t) {} T val; }; namespace boost { namespace serialization { template < typename Archive, typename T > void save_construct_data(Archive & ar, storing<T> const* ptr, unsigned int const) { // what to put here? } template < typename Archive, typename T > void load_construct_data(Archive & ar, storing<T> * ptr, unsigned int const) { // what here? } }} If I try to store val as a pointer then I can't instantiate with anything serialization considers "primitive". For instance, if I try to serialize a storing<std::string> then I'll get errors about std::string not having a serialization routine even if I include boost/serialization/string.hpp. If I try to load into a temporary variable then I run into the problem of having to construct one, and sometimes I need extra information to do so (I have the serialization routines dealing with this in such cases). Any ideas?

1 0

Re: [Boost-users] [Range] sub_range and list not working together
by Brian Wood 13 Jul '11

13 Jul '11

Neil Groves: > This compiler error is to be expected because your code is calling the > size() member function. The size() member function requires that the > underlying range is a model of the RandomAccessRange Concept. Since you are > using a sub_range of a std::list and the std::list has bidirectional > iterators, the sub_range is a model of the BidirectionalRange Concept. > > This has changed between boost versions. The boost::size(Range&) function > was modified to only work with ranges that model the RandomAccessRange > Concept because the performance complexity was inconsistent with typical > expectations. For example, in your case the sub_range size would have been > O(N) despite the container probably having an O(1) implementation of size(). > The complexity of std::list size() is not mandated by the standard to be > O(1), even though it is recommended. If one wishes to pay the price and > wishes to have a more general 'size' calculation I recommend using > boost::distance(Range&). Thanks. I found this thread which is closely related to what I'm wondering. http://boost.2283326.n4.nabble.com/Distance-size-of-boost-ranges-for-std-se… I haven't checked into this much, but would using boost::distance when the underlying container is a deque or vector, be slower than using the container's size? Regards, Brian Wood Ebenezer Enterprises http://webEbenezer.net

2 1

[io::streams] Using array device and std::output_iterator
by Ovanes Markarian 13 Jul '11

13 Jul '11

Hello *, I tried to use an array-device based stream and wantet to pass the stream to std::output_iterator or std::istream_iterator, but unfortunately, I get a compilation error with gcc 4.3.5. Boost::IOStreams docu states that the io::stream is either derived from std::basic_istream or std::basic_ostream or both dependent on the underlying device category. array has seekable category, so I'd expect io::stream to derive from both and be compatible with ostream_iterator or istream_iterator. Here is the code snippet: typedef unsigned char byte; typedef io::basic_array<byte> array_device; typedef io::stream<array_device> array_stream; array_stream ios_(ptr_to_byte_array); std::istream_iterator<byte> in(ios_); And the last line results in the error stating: src/chunk_reading_iterator_test.cpp: In member function 'void chunk_reading_iterator_suite::ctor::test_method()': src/chunk_reading_iterator_test.cpp:86: error: no matching function for call to 'std::istream_iterator<unsigned char, char, std::char_traits<char>, int>::istream_iterator(chunk_reading_iterator_suite::<unnamed>::array_stream&)' Thanks for any suggestions, Ovanes

1 0

[regex] Segmentation fault on regex_search only on CentOS
by Alessandro Candini 13 Jul '11

13 Jul '11

I have compiled the same C++ source on two machines: 1 - Ubuntu 11.04, 2.6.38-10-generic #46-Ubuntu SMP Tue Jun 28 15:07:17 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux 2 - CentOS release 5.3, 2.6.18-128.1.14.el5 #1 SMP Wed Jun 17 06:40:54 EDT 2009 i686 i686 i386 GNU/Linux On Ubuntu I have g++-4.5.2, on CentOs g++-4.4.4 and on both I have boost-1.42.0. My source compiles perfectly on both machines, but on the CentOS one I get a Segmentetion Fault during the execution. Using some prints, I found out where it happens: ... boost::regex grepGPP("Monthly Ecosystem GPP", boost::regex::egrep); cerr << "DEBUG BEFORE!" << endl; boost::regex_search (currGPP, grepGPP) cerr << "DEBUG AFTER!" << endl; ... The segmentation fault comes during the execution of regex_search, beacause "DEBUG AFTER" is not printed. Any ideas on how to solve this problem? Thanks in advance. -- Alessandro Candini MEEO S.r.l. Via Saragat 9 I-44122 Ferrara, Italy Tel: +39 0532 1861501 Fax: +39 0532 1861637 http://www.meeo.it ======================================== "ATTENZIONE:le informazioni contenute in questo messaggio sono da considerarsi confidenziali ed il loro utilizzo è riservato unicamente al destinatario sopra indicato. Chi dovesse ricevere questo messaggio per errore è tenuto ad informare il mittente ed a rimuoverlo definitivamente da ogni supporto elettronico o cartaceo." "WARNING:This message contains confidential and/or proprietary information which may be subject to privilege or immunity and which is intended for use of its addressee only. Should you receive this message in error, you are kindly requested to inform the sender and to definitively remove it from any paper or electronic format."

3 2

[regex] Is it possible to feed boost regex with NFA/DFA
by Chandan Nilange 13 Jul '11

13 Jul '11

We are developing DFA/NFA based hardware regex engine. I intend to use regex boost as goldenized model. Can we feed the regex boost with NFA/DFA states instead of regular expressions syntax Regards, Chandan Nilange

2 1

Serialization writing an old version of a class
by Don 13 Jul '11

13 Jul '11

Using boost serialization, I would like to be able to write an old version of a class, so that I can save data in an "old" format to allow other users with old versions of my program to read the file. I see from the archives that this has come up before, and Robert Ramey indicated it was doable and on the todo list, but I don't see it in the release notes or documentation. Did anything every come of this? Or, has anyone else done this? Thanks, Don.

1 0

Segmentation fault of boost::format with Intel compiler on OS X
by Daniel Langr 13 Jul '11

13 Jul '11

Hi, compiling the following code with Intel's icpc on OS X: int i = 10; std::string s = boost::str(boost::format("%d") % i); we got segmentation fault when running the executable. With g++ everything works fine. Versions: Boost 1.46.1, Intel icpc 11.1/0.67, OS X 10.8.0 Please help, thanks in advance, Daniel

1 0

[Range] sub_range and list not working together
by Brian Wood 13 Jul '11

13 Jul '11

B"H I've been testing my marshalling support for sub_range and have a question. In reviewing some old posts and notes, it looks like the following used to compile: boost::sub_range<std::list<int32_t> > rng; Today though I get the following error message: g++ -Os -s -W -Wall -I. -I ./loki-0.1.7/include/loki/flex -I /home/brian/boost_1_47_0_beta1 -c -o local_messages_d.cg.o local_messages_d.cg.cc In file included from /home/brian/boost_1_47_0_beta1/boost/range/iterator_range.hpp:13:0, from /home/brian/boost_1_47_0_beta1/boost/range/sub_range.hpp:23, from ./local_messages_d.cg.hh:5, from local_messages_d.cg.cc:1: /home/brian/boost_1_47_0_beta1/boost/range/iterator_range_core.hpp: In member function ‘boost::iterator_range<IteratorT>::difference_type boost::iterator_range<IteratorT>::size() const [with IteratorT = std::_List_iterator<int>, boost::iterator_range<IteratorT>::difference_type = int]’: /home/brian/boost_1_47_0_beta1/boost/range/sub_range.hpp:117:58: instantiated from ‘boost::sub_range<ForwardRange>::difference_type boost::sub_range<ForwardRange>::size() const [with ForwardRange = std::list<int>, boost::sub_range<ForwardRange>::difference_type = int]’ local_messages_d.cg.cc:22:33: instantiated from here /home/brian/boost_1_47_0_beta1/boost/range/iterator_range_core.hpp:252:32: error: no match for ‘operator-’ in ‘((const boost::iterator_range<std::_List_iterator<int> >*)this)->boost::iterator_range<std::_List_iterator<int> >::m_End - ((const boost::iterator_range<std::_List_iterator<int> >*)this)->boost::iterator_range<std::_List_iterator<int> >::m_Begin’ /usr/lib/gcc/i686-redhat-linux/4.5.1/../../../../include/c++/4.5.1/bits/stl_bvector.h:179:3: note: candidate is: ptrdiff_t std::operator-(const std::_Bit_iterator_base&, const std::_Bit_iterator_base&) /home/brian/boost_1_47_0_beta1/boost/range/iterator_range_core.hpp:253:13: warning: control reaches end of non-void function Is this to be expected? I don't know if a specialization for std::list in this regard would be useful, but seems possible. If I change the list to a vector, everything compiles fine. boost::sub_range<std::vector<int32_t> > rng; The following sentence from the docs doesn't make sense to me. "counting_range is a function to generator that generates an iterator_rangewrapping a counting_iterator (from Boost.Iterator). " Regards, Brian Wood Ebenezer Enterprises http://webEbenezer.net

2 1

boost regex 64 bit build
by Lloyd 13 Jul '11

13 Jul '11

Hi, I want to build boost regex 64 bit for a program on a Windows 7 64bit machine. I have downloaded boost 1.45 and built "bjam" by running "bootstarp.bat". >From "bjam" I built the regex library using the following command bjam --build-type=minimal --with-regex using=msvc --address-model=64 I am compiling using visual studio 2010. When I try to link my program with boost::regex, i get the error unresoleved external symbol " private: class boost::basic_regex<char, ...... What could be the problem? Thanks a lot, Lloyd

2 3

[iostreams] Issue when compiling with zlib, and suggestion
by Sean Farrow 13 Jul '11

13 Jul '11

Hi all: I'm currently trying to compile boost 1.47. I need to use the zlib features in the ioStreams library hence am using the following command line: bjam toolset=msvc --stagedir=stage/vc/v10/32-bit -sZLIB_SOURCE="c:\progra~1\zlib" -sZLIB_LIBPATH="c:\progra~1\zlib\boost_zlib.lib" -sBZIP2_SOURCE="c:\progra~1\bzip2" --build-type=complete >boostv10.txt this fails telling me it carn't open boost_zlib.lib. it appears from the docs that this command line is correct. I'm compiling on windows xp, vc 2010. Also, whilst I think about it, without modifying the default jam file the boost.iostreams library fails to compile, the file gzio.c is no longer in the zlib distribution as of version 1.2.3.9-could this be rectified for future boost releases? Regards Sean.

2 1