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

[thread] stupid question
by Johan Råde 29 Aug '08

29 Aug '08

What is the correct syntax to make the current thread sleep for 5 seconds? --Johan Råde

2 1

[logging] Library status?
by Allen Cronce 29 Aug '08

29 Aug '08

Hi all, I'm experimenting with John Torjo's proposed logging library. According to a post by John to this mailing list in May, he was interested in having the third review by August. Did the review happen? What's the status of the logging library? Best, -- Allen Cronce

3 3

boost - newbie question - remove files
by H H 29 Aug '08

29 Aug '08

Hi, I'm new to boost. Can I remove some files from the boost directory and keep only what I need? Will that work and it's license for? Thanks. boost\lib boost/function.hpp functional boost/preprocessor/iterate.hpp boost/detail/workaround.hpp boost/bind.hpp boost/shared_ptr.hpp boost/regex.hpp boost::smatch boost::regex_search boost::regex boost::regbase::no_except

3 2

multi_array, resize and remove a row
by gast128 29 Aug '08

29 Aug '08

Dear all, this is yet another post on multi_array. It seems possible that the multi_array's are resizable (using default construction :( ). However we have no the problem that a specific row should be removed (which is not necessarily the last row). Using the resize option it seems not possbile. Am I correct? Wkr.

2 2

[spirit2] push_back not fully understood - vc8 rejects code, why?
by Markus Werle 29 Aug '08

29 Aug '08

> Hi! Though I appreciate simple examples in the docs it is indeed helpful to have examples with fully qualified names. I am having hard times figuring out which versions of "_1", "at_c", etc. are in use. the fusion/phoenix/proto/mpl chaos is hard to keep track of. In summary: using declaratives are nice for toy examples but drive you nuts, once you #include more files and have ADL destroying the idyll of simplicity. Below is a piece of code that nearly compiles. One line has a problem: A semantic action, namely a push_back fails under vc8 and the error message is far from being helpful. So could you please be so nice and give me some hint on how to fix this. (BTW, Joel: "iter+30" directives from the spirit2 examples are assert-catched by the VC8 debugging STL, so I changed that part a little bit to first check the distance and doing iterator arithmetics afterwards. See below.) best regards, Markus ---snip--- #define BOOST_SPIRIT_DEBUG_OUT std::cerr #define BOOST_SPIRIT_DEBUG #include "Enforce.h" #include <boost/tr1/functional.hpp> #include <boost/tr1/tuple.hpp> #include <boost/mpl/assert.hpp> #include <boost/mpl/at.hpp> #include <boost/config/warning_disable.hpp> #include <boost/spirit/include/qi.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/spirit/include/phoenix_operator.hpp> #include <boost/spirit/include/phoenix_fusion.hpp> #include <boost/spirit/include/phoenix_stl.hpp> #include <boost/spirit/include/phoenix_container.hpp> #include <boost/fusion/include/adapt_struct.hpp> #include <boost/variant/recursive_variant.hpp> #include <boost/foreach.hpp> #include <boost/preprocessor/repetition.hpp> #include <boost/preprocessor/arithmetic/sub.hpp> #include <boost/preprocessor/punctuation/comma_if.hpp> #include <list> #include <string> #include <vector> #include <iostream> #include <iterator> using namespace boost::spirit; using namespace boost::spirit::qi; using namespace boost::spirit::ascii; using namespace boost::spirit::arg_names; using boost::phoenix::push_back; namespace fusion = boost::fusion; namespace phoenix = boost::phoenix; namespace { //////////////////////////////////////////////////////////////////////////////// // parser traits struct parser_traits { template <typename T, int Dummy = 0> struct parser; #define PARSER_TRAITS(TYPE, PARSER_T) \ template <int Dummy> struct parser<TYPE, Dummy> \ { \ typedef PARSER_T type; \ \ static inline \ PARSER_T p() \ { \ PARSER_T result; \ return result; \ } \ } // int PARSER_TRAITS(int, int_type); // unsigned int PARSER_TRAITS(unsigned int, uint_type); // double PARSER_TRAITS(double, double_type); }; //////////////////////////////////////////////////////////////////////////////// // parser template <typename Iterator, typename Tuple> struct tuple_grammar; template <typename Iterator, typename T1, typename T2> struct tuple_grammar<Iterator, std::tr1::tuple<T1, T2> > : grammar<Iterator, std::list<std::tr1::tuple<T1, T2> >(), space_type> { typename parser_traits::parser<T1>::type p1; typename parser_traits::parser<T2>::type p2; typedef typename std::tr1::tuple<T1, T2> tuple_t; tuple_grammar() : tuple_grammar::base_type(rule_tuple_list) { using phoenix::at_c; using boost::spirit::arg_names::_1; // using boost::spirit::container::push_back; using boost::phoenix::push_back; rule_tuple = '(' >> p1[at_c<0>(_val) = _1] >> ',' >> p2[at_c<1>(_val) = _1] >> ')' ; rule_tuple_list = // COMPILER FALS ON THIS SEMANTIC ACTION - WHY? *(rule_tuple[push_back(ref(_val), _1)]) ; } rule<Iterator, tuple_t(), space_type> rule_tuple; rule<Iterator, std::list<tuple_t>(), space_type> rule_tuple_list; }; //////////////////////////////////////////////////////////////////////////////// // string_to_list_of_tuples template <typename Tuple> struct string_to_list_of_tuples; template <typename T1, typename T2> struct string_to_list_of_tuples<std::tr1::tuple<T1, T2> > { typedef std::tr1::tuple<T1, T2> tuple_t; static std::list<tuple_t> apply(std::string const & s) { std::list<tuple_t> result; typedef tuple_grammar<std::string::const_iterator, tuple_t> tuple_grammar; tuple_grammar tg; std::string::const_iterator iter = s.begin(); std::string::const_iterator end = s.end(); bool r = phrase_parse(iter, end, tg, result, space); if (r && iter == end) { std::cout << "-------------------------\n"; std::cout << "Parsing succeeded\n"; std::cout << "-------------------------\n"; } else { std::size_t dist = std::distance(iter, end); std::string::const_iterator some = dist > 30 ? iter+30 : end; std::string context(iter, some); std::cout << "-------------------------\n"; std::cout << "Parsing failed\n"; std::cout << "stopped at: \"" << context << "...\"\n"; std::cout << "-------------------------\n"; } return result; } }; } // namespace int main() { try { { typedef std::tr1::tuple<int, int> tuple_t; std::list<tuple_t> test = string_to_list_of_tuples<tuple_t>::apply (" ( 1, 2 ) ( 3, 4)(5,6)"); std::copy(test.begin(), test.end(), std::ostream_iterator<tuple_t>(std::cout, "\n")); } // { // typedef std::tr1::tuple<unsigned int, int, double> tuple_t; // std::list<tuple_t> test = // string_to_list_of_tuples<tuple_t>::apply // (" ( 1, -2 , 5.123) ( 3, 4,7.9)(5,6,8.6789)"); // std::copy(test.begin(), test.end(), // std::ostream_iterator<tuple_t>(std::cout, "\n")); // } } catch (std::exception const & e) { std::cerr << e.what() << std::endl; } }

2 3

[filesystem] Catching permissions errors?
by Allen Cronce 29 Aug '08

29 Aug '08

Hi all, I need to programmatically test for permissions errors during path iteration. I think what I need to do is catch the filesystem exception, then drill out the error code and compare it to the permission_denied error value. Here's a code fragment written under boost 1.35 that demonstrates what I'm talking about: try { // Force a permissions error on *nix machines when not running as root. // Note that this code is is for demonstration purposes only. Assume that // in the real implementation I'm doing something else here with the // filesystem that *might* throw a permissions error. boost::filesystem::path testDirPath( "/usr/local/cant_touch_this" ); create_directory(testDirPath); } catch(const boost::filesystem::filesystem_error& e) { // Check the error code to see if it's a permissions error if (boost::system::posix_error::permission_denied == e.code().value()) { std::cout << "The error code indicates a permissions error, as expected. " << std::endl; } else { std::cout << "Hmm... I didn't get a permissions error. That's wrong." << std::endl; } } Is this basically the right approach? Note that one concern I have is maintainability. A coworker wrote similar code to the above, but under boost 1.33.1. That code broke under 1.35 due to significant changes to the filesystem exceptions mechanism. I understand that there are no guarantees moving forward to new boost versions. I knew the job was dangerous when I took it ;-) But if there's anything I can do today to future-proof a little bit, that would be great. Thanks in advance for any feedback or suggestions. Best, -- Allen Cronce

1 1

[exception] RTTI
by Anteru 28 Aug '08

28 Aug '08

Hi, I just upgrade to Boost 1.36, and while I'm not using Boost::exception, it gets included somewhere (gotta investigate that). That's no problem, but as my project does not use RTTI, but Boost::exception does, I'm getting warnings that the typeid() operator is used. Happens at two lines, Q:\Boost\include\boost/exception/exception.hpp(74) : warning C4541: 'typeid' used on polymorphic type 'boost::exception' with /GR-; unpredictable behavior may result Q:\Boost\include\boost/exception/exception.hpp(81) : warning C4541: 'typeid' used on polymorphic type 'boost::exception' with /GR-; unpredictable behavior may result Is there any way to disable this? After all, I'm not using boost::exception specifically, and I don't want to enable RTTI just to get rid of the warning. I'm fine if the diagnostic message becomes garbage then, after all, one usage is just to get the name of the class (typeid(*this).name), which may even return bogus results as typeid is not required to return some meaningful name. Cheers, Anteru

7 18

Visual studio 2003 settings
by Sephiroth 28 Aug '08

28 Aug '08

Hello I'm new to boost and c++. I wanted to know where I can find the settings for Visual Studio 2003. Expecially the linking additional dependencies. Thank for the help Chris

2 1

new user with compile problems
by Brenda Orlin 28 Aug '08

28 Aug '08

I am trying to use Boost version 1.36 with Microsoft Visual Studio 2005. When I try to include the thread library, I get all kinds of compile errors. One example is: Error 1 error C2632: 'float' followed by 'float' is illegal C:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h 155 This does not seem right, however, I don't know what to do about it. Please help! _________________________________________________________________ Get ideas on sharing photos from people like you. Find new ways to share. http://www.windowslive.com/explore/photogallery/posts?ocid=TXT_TAGLM_WL_Pho…

2 6

boost::reference_wrapper & polymorphism
by Robert Dailey 28 Aug '08

28 Aug '08

Does boost::reference_wrapper provide polymorphic behavior? For example, if I have a boost::reference_wrapper to a base type, and I assign a derived type to it, will the function calls be polymorphic?

2 3

[TR1] patch for tr1/detail/config_all.hpp ?
by Williams, Alan B 28 Aug '08

28 Aug '08

Hello, I'm using boost.TR1 in 1.36.0, and found that tr1/detail/config_all.hpp needs a little tweak in order to work with the PGI 7.1.4 compiler. Should I send a proposed diff or patch to this mailing list, or to some individual? Thanks, Alan

2 1

[logging] Milliseconds not logged on Windows
by Allen Cronce 28 Aug '08

28 Aug '08

Hi all, I'm using the latest boost logging library from the sandbox, in conjunction with boost 1.35.0. I'm seeing problems with millisecond strings in the log under Windows VS2005. The following formatter works fine under posix (specifically, Mac OS X): g_log_dbg()->writer().add_formatter( formatter::time("$hh:$mm.$ss. $mili ") ); Here's some sample output: 19:33.54.4196908 hello, world 19:33.54.4197788 this is so cool (but only in the debug build) 1 But the same code running under Windows always outputs a consistent, bogus millisecond value: 18:41.47.-858993460 hello, world 18:41.47.-858993460 this is so cool (but only in the debug build) 1 Looks like a sign extension problem. But I have to admit that I have not invested the time to determine whether this is a bug in the logging library proper, or a problem with the underlying boost 1.35.0 precision time and string utilities used by the logging library. Regardless I can probably live without milliseconds in my Windows logs, so I'll work around it with a platform conditional that specifies an alternate format string under Windows. Best, -- Allen Cronce

1 0

[logging] Build errors under VS2005 Unicode
by Allen Cronce 28 Aug '08

28 Aug '08

Hi all, I'm using the boost logging library from the sandbox, in conjunction with boost 1.35.0. I'm seeing problems building basic sample code under Windows VS2005 in projects where UNICODE and _UNICODE is defined. Here's an example line of problematic code: g_log_err()->writer().add_formatter( formatter::idx(), "[%] " ); Here's the error generated: --snip-- 1>c:\sources\logutils.cpp(56) : error C2664: 'void boost ::logging ::writer ::format_write < formatter_base ,destination_base ,lock_resource ,apply_format_and_write ,router_type ,formatter_array ,destination_array >::add_formatter<boost::logging::formatter::idx_t<>>(formatter,const boost::logging::char_type *)' : cannot convert parameter 2 from 'const char [5]' to 'const boost::logging::char_type *' 1> with 1> [ 1> formatter_base = boost ::logging ::formatter::base<boost::logging::default_,boost::logging::default_>, 1> destination_base = boost ::logging ::destination::base<boost::logging::default_,boost::logging::default_>, 1> lock_resource=boost::logging::lock_resource_finder::single_thread, 1> apply_format_and_write = boost ::logging ::format_and_write ::simple<boost::logging::optimize::cache_string_one_str<>>, 1> router_type=boost::logging::msg_route::simple<boost::logging::formatter::base<boost::logging::default_,boost::logging::default_>,boost::logging::destination::base<boost::logging::default_,boost::logging::default_>,boost::logging::lock_resource_finder::single_thread>, 1> formatter_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::formatter ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex>, 1> destination_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::destination ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex>, 1> formatter=boost::logging::formatter::idx_t<> 1> ] 1> Types pointed to are unrelated; conversion requires reinterpret_cast, C-style cast or function-style cast 1>c:\sources\logutils.cpp(58) : error C2440: '<function-style-cast>' : cannot convert from 'const char [19]' to 'boost::logging::formatter::time' 1> No constructor could take the source type, or constructor overload resolution was ambiguous 1>c:\sources\logutils.cpp(58) : error C2780: 'void boost ::logging ::writer ::format_write < formatter_base ,destination_base ,lock_resource ,apply_format_and_write ,router_type ,formatter_array,destination_array>::add_formatter(formatter,const boost::logging::char_type *)' : expects 2 arguments - 1 provided 1> with 1> [ 1> formatter_base = boost ::logging ::formatter::base<boost::logging::default_,boost::logging::default_>, 1> destination_base = boost ::logging ::destination::base<boost::logging::default_,boost::logging::default_>, 1> lock_resource=boost::logging::lock_resource_finder::single_thread, 1> apply_format_and_write = boost ::logging ::format_and_write ::simple<boost::logging::optimize::cache_string_one_str<>>, 1> router_type=boost::logging::msg_route::simple<boost::logging::formatter::base<boost::logging::default_,boost::logging::default_>,boost::logging::destination::base<boost::logging::default_,boost::logging::default_>,boost::logging::lock_resource_finder::single_thread>, 1> formatter_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::formatter ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex>, 1> destination_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::destination ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex> 1> ] 1> c:\sources\thirdpartylibs\installedwin32\individual\boost \include\boost\logging\detail\format_write_detail.hpp(232) : see declaration of 'boost ::logging ::writer ::format_write < formatter_base ,destination_base ,lock_resource ,apply_format_and_write ,router_type,formatter_array,destination_array>::add_formatter' 1> with 1> [ 1> formatter_base = boost ::logging ::formatter::base<boost::logging::default_,boost::logging::default_>, 1> destination_base = boost ::logging ::destination::base<boost::logging::default_,boost::logging::default_>, 1> lock_resource=boost::logging::lock_resource_finder::single_thread, 1> apply_format_and_write = boost ::logging ::format_and_write ::simple<boost::logging::optimize::cache_string_one_str<>>, 1> router_type=boost::logging::msg_route::simple<boost::logging::formatter::base<boost::logging::default_,boost::logging::default_>,boost::logging::destination::base<boost::logging::default_,boost::logging::default_>,boost::logging::lock_resource_finder::single_thread>, 1> formatter_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::formatter ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex>, 1> destination_array = boost ::logging ::array ::shared_ptr_holder < boost ::logging ::destination ::base < boost ::logging ::default_,boost::logging::default_>,boost::logging::threading::mutex> 1> ] --snip-- According to the documentation, the library should automatically sense the UNICODE definition under Windows and do the right thing. But that doesn't seem to work right. I played around with trying to get this to work, but found that I had to always pass non-wide strings to the various interfaces. Ultimately I gave up and disabled the logging of wide character strings by defining the following: #define BOOST_LOG_DONOT_USE_WCHAR_T For my particular needs, logging 8 bit characters probably makes more sense anyway. But it's somewhat bothersome that the logging library did not work out of the box as documented. Sorry in advance if I missed some other configuration or nuance needed to get the logging library to work under UNICODE. Best, -- Allen Cronce

1 0

Boost.Log database backends
by Mika Heiskanen 28 Aug '08

28 Aug '08

Hi, Does anyone know if there is a database backend available for Boost.Log? I couldn't find any matches from Google with any sensible keyword combinations. --> Mika Heiskanen / Finnish Meteorological Insititute

1 0

[date_time]: Easter calculation
by Jean-Pierre Bergamin 28 Aug '08

28 Aug '08

Dear Boost-Users Is there a year_based_generator for easter dates somewhere lying around, waiting to be used by others? Thanks James

3 4

Boost at Dev-Cpp 5 (Windows XP)
by Berno Wiegel 28 Aug '08

28 Aug '08

Hello all, I have problems with installing Boost. I downloaded boost_1_36_0.zip and extract it to c:\boost\. After that I downloaded boost-jam-3.1.16-1-ntx86(2).zip and extract bjam.exe to c:\boost\boost_1_36_0\. Now I start cmd and goes to c:\boost\boost_1_36_0\! There I execut bjam with following parameters bjam.exe "-sMINGW_ROOT_DIRECTORY=C:\Dev-Cpp" "-sTOOLS=mingw" install This is my Console: C:\BOOST\boost_1_36_0>bjam.exe "-sMINGW_ROOT_DIRECTORY=C:\Dev-Cpp" "-sTOOLS=min w" install warning: Graph library does not contain optional GraphML reader. note: to enable GraphML support, set EXPAT_INCLUDE and EXPAT_LIBPATH to the note: directories containing the Expat headers and libraries, respectively. warning: skipping optional Message Passing Interface (MPI) library. note: to enable MPI support, add "using mpi ;" to user-config.jam. note: to suppress this message, pass "--without-mpi" to bjam. note: otherwise, you can safely ignore this message. WARNING: No python installation configured and autoconfiguration failed. See http://www.boost.org/libs/python/doc/building.html for configuration instructions or pass --without-python to suppress this message and silently skip all Boost.Python targets Building Boost.Regex with the optional Unicode/ICU support disabled. Please refer to the Boost.Regex documentation for more information (don't panic: this is a strictly optional feature). warning: No toolsets are configured. warning: Configuring default toolset "msvc". warning: If the default is wrong, your build may not work correctly. warning: Use the "--toolset=xxxxx" option to override our guess. warning: For more configuration options, please consult warning: http://boost.org/boost-build2/doc/html/bbv2/advanced/configuration.htm C:/BOOST/boost_1_36_0/tools/build/v2/tools\msvc.jam:320: in configure-really *** argument error * rule path.make ( native ) * called with: ( ) * missing argument native C:/BOOST/boost_1_36_0/tools/build/v2/util\path.jam:44:see definition of rule 'm ke' being called C:/BOOST/boost_1_36_0/tools/build/v2/tools\msvc.jam:172: in configure C:/BOOST/boost_1_36_0/tools/build/v2/tools\msvc.jam:145: in msvc.init C:/BOOST/boost_1_36_0/tools/build/v2/build\toolset.jam:38: in toolset.using C:/BOOST/boost_1_36_0/tools/build/v2\build-system.jam:553: in load C:\BOOST\boost_1_36_0\tools\build\v2/kernel\modules.jam:281: in import C:\BOOST\boost_1_36_0\tools\build\v2/kernel/bootstrap.jam:128: in boost-build C:\BOOST\boost_1_36_0\boost-build.jam:16: in module scope C:\BOOST\boost_1_36_0> I followed this guide (http://wagjo.com/index.php?id=27) It is espacially for Dev-Cpp users, but ther is no installation. I dont know what to do. Please help me.

4 9

Simply including boost/bind.hpp causes compilation error
by Ken Smith 28 Aug '08

28 Aug '08

Here is my environment. % g++ -v Using built-in specs. Target: i686-apple-darwin9 Configured with: /var/tmp/gcc/gcc-5483~1/src/configure --disable-checking -enable-werror --prefix=/usr --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.0/ --with-gxx-include-dir=/include/c++/4.0.0 --with-slibdir=/usr/lib --build=i686-apple-darwin9 --with-arch=apple --with-tune=generic --host=i686-apple-darwin9 --target=i686-apple-darwin9 Thread model: posix gcc version 4.0.1 (Apple Inc. build 5483) Boost 1.35.0 (also produced same issue with Boost 1.34.1) Here is the code. 1 #include <boost/lambda/lambda.hpp> 2 #include <boost/lambda/construct.hpp> 3 #if defined(INCLUDE_BOOST_BIND) 4 # include <boost/bind.hpp> 5 #endif 6 #include <iostream> 7 #include <iterator> 8 #include <algorithm> 30 31 int main() 32 { 33 using namespace std; 34 using namespace boost::lambda; 35 36 typedef std::istream_iterator<int> in; 37 std::for_each(in(std::cin), in(), std::cout << (_1 * 3) << " " ); 76 } (The astute reader will note discontinuity in the line numbering. This is a playground for my exploration with boost. All elided lines are #if 0'ed. Here is the compile command line and the compilation error, both edited for clarity. g++ -c -g -Wall -fexceptions -DINCLUDE_BOOST_BIND -I/path/to/boost-1.35.0/include main.cpp -o main.o main.cpp: In function 'int main()': main.cpp:37: error: '_1' was not declared in this scope I have seen these related articles from the archives of this list but none of the suggestions eliminated the error. I added the inclusion of boost/lambda/construct.hpp per one of the suggestions there. Its inclusion or exclusion doesn't change the behavior. http://thread.gmane.org/gmane.comp.lib.boost.user/35128 http://thread.gmane.org/gmane.comp.lib.boost.user/16630 http://thread.gmane.org/gmane.comp.lib.boost.user/16887 As you may imagine, the elided (and commented out) lines contain actual uses of boost/bind.hpp but it is a mystery to me that simple inclusion of a header file would cause otherwise working code to stop compiling. This compile command (same as above but with -D removed): g++ -c -g -Wall -fexceptions -I/path/to/boost-1.35.0/include main.cpp -o main.o runs to completion and the subsequent link is successful. Any ideas? Thanks in advance, Ken

4 4

[filesystem] Should fstream really 'force' in+out mode?
by Sachin Garg 27 Aug '08

27 Aug '08

I want to open a read only file in ios::in mode using fstream but its open forces in+out mode, causing my open to fail when it could have succeeded. In boost::filesystem::fstream::open, the file open mode is modified to mode | std::ios_base::in | std::ios_base::out Is this really necessary or can this be changed in future? Atleast with MSVC 9.0, std::fstream isn't forcing such behavior. SG

1 1

[program_options] format_paragraph chops words
by Soren Soe 27 Aug '08

27 Aug '08

Hi, I have run into a problem with the program_options library chopping words when it formats help messages. I have created a small code fragment that will duplicate the problem: int main() { std::string p = "\n1234 6789 1234 6789 1234 6789 1234"; format_description(std::cout, p, 20, 44); return 0; } Results in this message: 1234 6789 1234 6789 123 4 6789 1234 In case formatting is not preserved, the 3'rd instance of '1234' is chopped into '123' with '4' on the next line. I think there might be a bug in format_paragraph in the block that is supposed to prevent word chopping. Specifically, this piece of code: // is last_space within the second half ot the // current line if ((unsigned)distance(last_space, line_end) < (line_length - indent) / 2) { line_end = last_space; } should probably not subtract 'indent' from 'line_length' since this was already done at the very beginning of the function. Thanks, Soren Soe. This email and any attachments are intended for the sole use of the named recipient(s) and contain(s) confidential information that may be proprietary, privileged or copyrighted under applicable law. If you are not the intended recipient, do not read, copy, or forward this email message or any attachments. Delete this email message and any attachments immediately.

1 0

[range] STL algorithm wrappers accepting range?
by Nat Goodspeed 27 Aug '08

27 Aug '08

Apologies if I'm overlooking something obvious, or if this is a FAQ. Has anyone published a suite of wrappers for common STL algorithms that accept Boost.Range instances instead of iterator pairs? Or must we each write our own? Seems like redundant effort...

4 5

[graph] [serialization] loading adjacency_list
by Mojmir Svoboda 27 Aug '08

27 Aug '08

hello, i have 1.36.0 now and i have troubles serializing adjacency list. graph itself saves okay in whatever oarchive and looks okay, but when loading graph, his properties are not set.. in fact defaults are used instead what serialization is trying to load. is there something that changed from 1.35.0 and 1.36.0 in this area? my (uneducated!) guess is that the serialization loads data into temporary that is discarded afterwards. i traced it to file adj_list_serialize.hpp around line 90: now i want the line ar >> serialization::make_nvp("vertex_property", get(vertex_all_t(), graph, v)); to be transformed into form: something & vp = get(vertex_all_t(), graph, v); ar >> serialization::make_nvp("vertex_property", vp); at least for debugging purposes.. i'm wondering that this "something" could be.. i tried property_map<...>::type etc, but failed ultimately. the same apply to edge: what is the return type of get(edge_all... when my graph is defined by: typedef boost::adjacency_list< boost::vecS, boost::vecS, boost::undirectedS, VertexProp, boost::property<boost::edge_index_t, uint16_t, EdgeProp> > T_Graph; many thanks, mojmir // please excuse incomplete information, i'm under pressure :/

1 1

base64_from_binary Q
by Terdale, Shantibhushan 27 Aug '08

27 Aug '08

I am using this sample to prove myself that base64_from_binary and binary_from_base64 can be used in my app. But I am having little problem and not able to understand whats missing Here is the sample #include "boost/archive/iterators/base64_from_binary.hpp" #include "boost/archive/iterators/binary_from_base64.hpp" #include "boost/archive/iterators/transform_width.hpp" #include <string> #include <iostream> using namespace std; using namespace boost::archive::iterators; typedef base64_from_binary< transform_width<std::string::iterator, 6, sizeof(char) * 8> > base64_t; typedef transform_width< binary_from_base64<std::string::iterator>, sizeof(char) * 8, 6 > binary_t; int main() { string str = "Hello, world!"; cout << str << endl; string enc(base64_t(str.begin()), base64_t(str.end())); // Problem here. It works when I specify "str.end()-1" instead of "str.end()" but I loose last character after decoding. cout << enc << endl; string dec(binary_t(enc.begin()), binary_t(enc.end())); cout << dec << endl; return 0; } Will appreciate any help. I tried std::copy as well and have same problem.

4 5

question about time_facet and date/time formatting
by Axel 27 Aug '08

27 Aug '08

Hello Following the documentation, the time_facet::put function take 3 common parameters : "an iterator pointing to the next item in the stream, an ios_base object, and the fill character.".But this is not clear for me, i m totally new to facets. I m using it like that and it works well : stringstream log_stream ; ptime time = second_clock::local_time() ; time_facet facet("%Y-%m-%d %H:%M:%S") ; facet.put(log_stream, log_stream, ' ' , time) ; cout << log_stream.str() ; But I don't understand WHY it works and I wonder some things : What does mean the "iterator to the next item" from the documentation since I consider the target as an input stream ? Why use the same object for the "ios_base object" parameter ? This means I could use a different stream, but what could be the utility of such an usage of put() ? Finally, what is the use of the "fill character" ? Thanks in advance.

1 0

a little Graph guidance (looking for hints)
by Carsten Hoyer 27 Aug '08

27 Aug '08

Hi Just turned to Boost - What a fantastic library :) I have a question about the construction of a Graph. I can't figure out how to use boost to represent a graph where each Vertex has several outs. F.ex Vertex 1 has out 1, 2 and 3. Vertex A has in a, b and c. I can then make edges going from 1-a, 1-b, 2-b etc. Like F.ex http://messe.harmony-central.com/Musikmesse05/Content/Native_Instruments/PR… Reaktor Any hints? Thank you Carsten -- View this message in context: http://www.nabble.com/a-little-Graph-guidance-%28looking-for-hints%29-tp190… Sent from the Boost - Users mailing list archive at Nabble.com.

4 5

MAC Address
by Jean-Sebastien Stoezel 27 Aug '08

27 Aug '08

Hello, How to get the MAC address of an ethernet interface, using the boost C++ library? Thanks, Jean

3 4

serialization of large complex structure
by Diane 27 Aug '08

27 Aug '08

I have a large, complex struct that is made up of many levels of structs, getting eventually to int, char*, enum, etc. There are also pointers to structs and arrays of pointers. There are no STL types within these structures. I am trying to wrap my head around how I can use boost.serialization for this, without modifying the structures themselves (I can't make changes to the definitions). Does anyone know the best way to do this? I saw several mentions of wrapper classes for serializing a float and figure I could use that approach, but I think I would have to make a mirror structure that's made up of my struct wrappers and copy the data over from the real struct into these classes. Or I guess I could use pointers, but I would still have to create this second monster. Is that really the best possibility? Any help is appreciated! Thanks, Diane

3 4

linker errors on 1.36.0
by Olaf Peter 27 Aug '08

27 Aug '08

Hi, I build release 1.36.0 on linux using jam successfully. By using this, I've got linker errors: interpreter.cpp.o: In function `void boost::call_once<void (*)()>(boost::once_flag&, void (*)())': interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x16): undefined reference to `boost::detail::get_once_per_thread_epoch()' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x4f): undefined reference to `boost::detail::once_epoch_mutex' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x89): undefined reference to `boost::detail::once_epoch_mutex' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xc1): undefined reference to `boost::detail::once_global_epoch' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xc7): undefined reference to `boost::detail::once_global_epoch' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xd2): undefined reference to `boost::detail::once_global_epoch' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xd8): undefined reference to `boost::detail::once_global_epoch' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xdd): undefined reference to `boost::detail::once_global_epoch' CMakeFiles/eda.dir/src/python/interpreter.cpp.o:interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xe3): more undefined references to `boost::detail::once_global_epoch' follow CMakeFiles/eda.dir/src/python/interpreter.cpp.o: In function `void boost::call_once<void (*)()>(boost::once_flag&, void (*)())': interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0xf2): undefined reference to `boost::detail::once_epoch_cv' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x125): undefined reference to `boost::detail::once_epoch_cv' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x198): undefined reference to `boost::detail::once_epoch_mutex' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x19f): undefined reference to `boost::detail::once_epoch_cv' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x20e): undefined reference to `boost::detail::once_global_epoch' interpreter.cpp:(.text._ZN5boost9call_onceIPFvvEEEvRNS_9once_flagET_[void boost::call_once<void (*)()>(boost::once_flag&, void (*)())]+0x214): undefined reference to `boost::detail::once_global_epoch' main.cpp.o: In function `main': main.cpp:(.text+0x312): undefined reference to `boost::program_options::options_description::m_default_line_length' main.cpp:(.text+0x3aa): undefined reference to `boost::program_options::options_description::m_default_line_length' main.cpp:(.text+0x478): undefined reference to `boost::program_options::options_description::m_default_line_length' main.cpp:(.text+0x4c6): undefined reference to `boost::program_options::options_description::m_default_line_length' main.cpp:(.text+0x55e): undefined reference to `boost::program_options::options_description::m_default_line_length' Any ideas what gone wrong? linker flags where: -rdynamic -L/usr/local/boost_1_36_0/lib ... -lboost_python -lboost_program_options -lboost_thread -Wl,-rpath,/usr/local/boost_1_36_0/lib:/usr/lib/python2.5/config Thanks, Olaf

2 4

Re: [Boost-users] [filesystem] Changes in Boost 1.36 not documented?
by anthony giardullo 27 Aug '08

27 Aug '08

On Tue, Aug 26, 2008 at 4:54 PM, anthony giardullo <anthonyj(a)gmail.com> wrote: > Scott McMurray wrote: >> On Fri, Aug 15, 2008 at 12:57, Scott Finley <sfinley_at_[hidden]> wrote: >>> bfs_test.cpp: In function 'int main()': >>> >>> bfs_test.cpp:13: error: 'class >>> boost::filesystem::basic_directory_entry<boost::filesystem::basic_path<std::basic_string<char, >>> std::char_traits<char>, std::allocator<char> >, >>> boost::filesystem::path_traits> >' has no member named 'leaf' >>> >>> My mistake, or did something change/break? >>> >> >> It looks like you now need >> itr->path().leaf() >> or >> boost::implicit_cast<bfs::path const &>(*itr).leaf() >> ( http://beta.boost.org/doc/libs/1_36_0/libs/filesystem/doc/reference.html#Cl… >> ) >> >> As I recall, this change was part of optimization efforts to make >> iterating faster by caching the file_status. >> >> Also, it was decided that some of the names for path decomposition >> functions were confusing, particularly branch and leaf, so they were >> changed. The new path decomposition table suggests that you should >> now be using filename() instead of leaf(): >> http://beta.boost.org/doc/libs/1_36_0/libs/filesystem/doc/reference.html#Pa… >> (But it's likely backwards-compatible, so you shouldn't need to.) > > Yes, that's the theory. But it looks like 1.36.0 inadvertently renamed > basic_directory_entry::leaf without providing backward compatibility. > > I'll try to post a hotfix for that. Since basic_directory_entry::leaf is > deprecated anyhow, my initial reaction is that it should have been left > alone. > > --Beman > I would also appreciate a hotfix for this as this change is breaking a project that I am hoping to release soon. Is there an alternative solution I can use that will work in both 1.36 and earlier versions? I tried both filename() and path().leaf, but these do not work in 1.33. Thanks, Anthony

2 1

[GIL] Does not pass test on mingw g++ 4.3
by John C. Femiani 26 Aug '08

26 Aug '08

I used a fresh version of GIL from SVN trunk, and when I say make all from within the boost/libs/gil folder I get a slew of errors including: image.cpp:302: instantiated from here c:\mingw\bin\../lib/gcc/mingw32/4.3.0/include/c++/bits/boost_concept_check.h:223: error: conversion from 'boost::detail::iterator_category_with_traversal<std::input_iterator_tag, boost::random_access_traversal_tag>' to non-scalar type 'std::forward_iterator_tag' requested I have some patches to GIL, but I can not seem to do regression testing since the unmodified version of GIL itself does not seem pass its tests. Am I doing something wrong? I looked at the Jamfile.v2 and it did not seem to add much to what the makefile was doing, but just in case I ran bjam from within the GIL folder and got the same results -- image.cpp filed to compile. I thought the error might involve the iterator library because of the error message I got, but Boost.Iterator seemed to pass all of its tests on my system. Also, the GIL makefile has a line 'CXX=/usr/local/gcc430/bin/g++', so it looks like this version of the compiler has been tested before... --John

1 0

(no subject)
by Timothy Barrett 26 Aug '08

26 Aug '08

I have been using the boost libraries for a number of years on non linux platforms. Recently began developing on redhat EL4 and am running into an issue not seen with my library on other OS platforms. I'm getting an exception in assign_to_own. I have read the previous note on this issue in the forum but do not believe the make file for my application or for the library application are the cause of the exception. Any pointers on how to track down and resolve this issue? Tim Barrett

2 4

using libbost_test_exec_monitor with Visual Studio 2008?
by Ákos Maróy 26 Aug '08

26 Aug '08

Hi, I'm having troulbe using libboost_text_exec_monitor with Visual Studio 2008. Basically a very simple test application, that initializes a test suite in an init_unit_test_suite() function fails to link on Visual Studio, when trying to link against both libboost_unit_test_framework-vc90-mt-gd-1_35.lib and libboost_test_exec_monitor-vc90-mt-gd-1_35.lib . I'm egetting an error of LNK2019 saying that int __cdecl test_main(int, char ** const) is an unresolved external symbol. The error, strangely enough, is reported from libboost_test_exec_monitor-vc90-mt-fd-1_35.lib(test_main.obj), being a reference from the function test_main_caller::operator(). The same code compiles & links fine using gcc on Linux, linking to the same libraries (of course without the -vc90-mt-fd-1_35 suffix). The code is simply: #include <boost/test/unit_test.hpp> using namespace boost::unit_test; void test() { BOOST_CHECK(true); } test_suite * init_unit_test_suite(int, char **) { test_suite *ts = BOOST_TEST_SUITE("Test"); ts->add(BOOST_TEST_CASE(&test)); framework::master_test_suite().add(ts); return 0; } Shall I add something to the above code? shouldn't the text_exec_monitor library contain test_main? Akos

2 2

std::find_if with bind and iequals to create predicate
by Peter Barker 26 Aug '08

26 Aug '08

Hello, I've a container of these objects: class Search { public: ... const std::string& getName() const {return name_;} }; I want to find a search in a case insensitive manner and this works: iter = std::find_if(searches.begin(),searches.end(), boost::bind(&boost::iequals<std::string,std::string>,boost::bind(&Search::getName,_1),searchName,std::locale())); However, I would've preferred to not have to specify the template parameters to iequals, and not have to specify the locale argument to iequals (it's optional). Is there a cleaner way to do this? Regards, Jim

1 0

std::is_tr1_tuple<T> - how?
by Markus Werle 26 Aug '08

26 Aug '08

Hi! While migrating from boost::tuple to std::tr1::tuple (pain!) I have to rewrite a lot of compile-time-magic code. Among other things I am in need of a type info template <typename T> struct is_tr1_tuple; Any implementation hint or proposal welcome. best regards, Markus

3 6

Problem with boost 1.36, vs2005 and serialization library -> linker problem
by Hansi 26 Aug '08

26 Aug '08

Hello, in my application I wrote the serialization like in the pimpl example. With boost 1.34.1 this has worked. But now I have a few strange linker errors. e.g Error 1413 error LNK2001: unresolved external symbol "protected: __thiscall boost::archive::detail::archive_pointer_oserializer<class boost::archive::detail::polymorphic_oarchive_route<class boost::archive::binary_oarchive_impl<class boost::archive::binary_oarchive,char,struct std::char_traits<char> > > >::archive_pointer_oserializer<class boost::archive::detail::polymorphic_oarchive_route<class boost::archive::binary_oarchive_impl<class boost::archive::binary_oarchive,char,struct std::char_traits<char> > > >(class boost::serialization::extended_type_info const &)" (??0?$archive_pointer_oserializer@V?$polymorphic_oarchive_route@V?$binary_oarchive_impl@Vbinary_oarchive@archive@boost@@DU?$char_traits@D@std@@@archive@boost@@@detail@archive@boost@@@detail@archive@boost@@$$FIAE@ABVextended_type_info@serialization@3@@Z) main.obj and Error 1414 error LNK2001: unresolved external symbol "protected: __thiscall boost::archive::detail::archive_pointer_iserializer<class boost::archive::polymorphic_binary_iarchive>::~archive_pointer_iserializer<class boost::archive::polymorphic_binary_iarchive>(void)" (??1?$archive_pointer_iserializer@Vpolymorphic_binary_iarchive@archive@boost@@@detail@archive@boost@@$$FIAE@XZ) main.obj The problem has something to do with BOOST_CLASS_EXPORT(A) function. If I don't include it there is not the problem with the linker error. In the cpp file I have at the moment the instantion template void A::serialize(boost::archive::polymorphic_iarchive & ar,const unsigned int file_version); template void A::serialize(boost::archive::polymorphic_oarchive & ar,const unsigned int file_version); and template void A::serialize(boost::archive::binary_iarchive & ar,const unsigned int file_version); template void A::serialize(boost::archive::binary_oarchive & ar,const unsigned int file_version); What I have to do to solve that? Best regards Hansjörg

3 3

Format local time in W3C format (YYYY-MM-DDThh:mm:ss.sTZD) without providing the time zone myself?
by Paul Bormans 26 Aug '08

26 Aug '08

I'm having trouble in formatting the local time in W3C format. The format is described at http://www.w3.org/TR/NOTE-datetime as follows: YYYY-MM-DDThh:mm:ss.sTZD (eg 1997-07-16T19:20:30.45+01:00) My first attempt was to use a ptime as follows: boost::posix_time::ptime t(boost::posix_time::microsec_clock::local_time()); boost::posix_time::time_facet* output_facet = new boost::posix_time::time_facet(); ss.imbue(std::locale(std::locale::classic(), output_facet)); output_facet->format("%Y-%m-%dT%H:%M:%S%f%Q"); ss << t; But when using a ptime in combination with a time_facet, the time zone (%Q) is ignored (also mentioned in the documentation http://www.boost.org/doc/libs/1_34_1/doc/html/date_time/date_time_io.html) So i thought to switch to the local time library which gives me the following code: typedef boost::date_time::time_facet<boost::local_time::local_date_time, char> ldt_facet; ldt_facet* output_facet = new ldt_facet("%Y-%m-%dT%H:%M:%S%f%Q"); ss.imbue(std::locale(std::locale::classic(), output_facet)); boost::local_time::time_zone_ptr zone; <=========== boost::local_time::local_date_time ldt1(boost::posix_time::microsec_clock::local_time(), zone); ss << ldt1; Problem here is that i have to provide the timezone where the software is running, which (offcourse) i do not know and actually would like the boost library to do this for me ;) After searching the web and reading the unittests i have found some clue's how i could make it work by first getting the difference between localtime and universial time and then format the timezone part myself. This will be much trouble for something that (i believe) should be in the library already. So my question is: How can i format the local time in W3C format (without providing the time zone myself)? Environment info: - Windows XP - VC2005 - Boost 1_34_1 Thanks for any hints & tips, Paul _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/

1 0

[program-options] Problem with paths that have spaces
by Allen Cronce 26 Aug '08

26 Aug '08

Hi all, I'm trying to use program-options to specify a filesystem path object under Boost 1.35.0. The following code works fine when the path specified on the command line has no spaces in it. But when it does have spaces, I get an error. --snip-- #include <iostream> #include <sstream> #include <boost/program_options.hpp> #include <boost/filesystem/path.hpp> namespace po = boost::program_options; namespace fs = boost::filesystem; using namespace std; int main (int argc, char * argv[]) { int result = 0; try { fs::path inPath; po::options_description desc("Allowed options"); desc.add_options() ("in,i", po::value<fs::path>(&inPath), "Input file or directory <pathname>") ; po::variables_map vm; po::store(po::parse_command_line(argc, argv, desc), vm); po::notify(vm); if (0 ==vm.count("in")) { throw invalid_argument("What? No input path specifed!"); } cout << "Input path = \"" << inPath.file_string() << "\"\n"; } catch(const boost::filesystem::filesystem_error& e) { cerr << "\nFilessytem error: " << e.what() << endl; result = 3; } catch(const std::exception& e) { cerr << "\nException error: " << e.what() << endl; result = 2; } catch(...) { cerr << "\nError: Caught an unknown exception!" << endl; result = 1; } return result; } --snip-- If I call the above with the following parameters, it works: --in "my_file" Input path = "my_file" But if I call it with a path that has a space in it, I get an error: --in "my file" Exception error: in option 'in': invalid option value 'my file' I know that I can work around the problem by loading the path into an interim string first, then use the string to initialize a path object. But unless I'm missing something, I would have thought that iteroperability between boost.program-options and boost.filesystem should be more seamless. It seems very common to want to load options into path objects. Best, -- Allen Cronce

2 1

link error trying to open file libboost_date-time-vc71-mt-gd-1_36.lib
by Vincent Buonassisi 26 Aug '08

26 Aug '08

hi, i create the multi-threaded and dll versions of the thread and date_time libraries on windows platform using vc++ 7.1: boost_date_time-vc71-mt-gd-1_36.dll/.lib boost_thread-vc71-mt-gd-1_36.dll/.lib when i link my application it wants the static version of the date_time library. is it possible to change boost so that it uses the dll instead of the static version? thanks, vince

2 1

Re: [Boost-users] [Thread] Beginner question regarding threadgroups
by Mike Marchywka 25 Aug '08

25 Aug '08

Hi, I am afraid that once these papers are published, we no longer have the copyright to it (I know it's weird but that is how academic papers work ...). So they are all on IEEE servers or some other journals. The only thing I could find is a link to some slides for a talk at the NUSOD conference but it is short on details. I have not used cite seer in nearly a decade since the advent of Google Scholar ... We do use geometric optics for this with a few exceptions (i.e. anti-reflection coatings used to increase extraction efficiency use a multi-layer plane wave transfer matrix method to get an effective reflection coefficient). But this is getting a bit off-topic so back on to the code ... The only reason I thought the list would tolerate this was general interest in optimization problems and literature sources. Everyone seems to want to use threads and ignore algorithmic issues that may provide ideas for library code. Certainly not boost specific but could develop into something. Considering a few real examples may be interesting by itself. Sent from my Verizon Wireless BlackBerry

1 0

Re: [Boost-users] [Thread] Beginner question regarding thread groups
by Michel Lestrade 25 Aug '08

25 Aug '08

Hi, I am afraid that once these papers are published, we no longer have the copyright to it (I know it's weird but that is how academic papers work ...). So they are all on IEEE servers or some other journals. The only thing I could find is a link to some slides for a talk at the NUSOD conference but it is short on details. I have not used cite seer in nearly a decade since the advent of Google Scholar ... We do use geometric optics for this with a few exceptions (i.e. anti-reflection coatings used to increase extraction efficiency use a multi-layer plane wave transfer matrix method to get an effective reflection coefficient). But this is getting a bit off-topic so back on to the code ... > Are any of the intro papers on google available as free, full text? That is my reason for going > to citeseer ( as it turns out they even seem to allow text format download ) > AFAIK all of the papers are available for free(IEEE is about the worst in terms of > open access, including for searching with scripts or using computer methods,LOL). > > Are these all homogeneous materials with a few planar discontinuities ? > Offhand it seems you could do a lot analytically with classical geometrical > optics. ================================================================================= Thanks for the code sample. It is not what I envisioned originally but I was not aware that threads were that expensive to create/destroy; this is fancier but makes more sense. I guess this will be my starting point, along with most of the Fortran code that I will be keeping. > > Actually, the while loop lives inside the threads, since creating and > destroying threads is relatively expensive, and you don't need to. > > >> > Did I understand you correctly ? It's a good idea and, from the little I >> > know on the subject, it sounds a lot like a "thread pool". >> > >> > > Yes, it's exactly a thread pool. > > Thanks again, Michel Lestrade Crosslight Software

1 0

Xpressive: UTF-8 and diacritics
by Allan Odgaard 25 Aug '08

25 Aug '08

It looks like the traits aspect of Xpressive is geared toward characters, so I assume that Xpressive is not directly usable with UTF-8 encoded text, am I correct? It might work by having the character type be a 32 bit integer and then use iterator adapters which expose the sequence as ucs-4 code points (after all, the sequence is “encoded”), but that leads me to the next question: diacritics. For example something like é in decomposed unicode is two code points (e followed by a combining ´ mark), so even when the sequence is iterated as ucs-4 code points, a regexp of “.” will match just the e, not the actual (rendered) character. Since I was unable to find any discussion of this while searching for Xpressive, I am curious to hear if any thoughts have gone into these issues.

2 1

[Spirit] C++ Parser library
by Florian Goujeon 25 Aug '08

25 Aug '08

Dear boosters, I post in the mailing list for the first time, so I don't know very much about what could have been said concerning any potential Spirit powered C++ parser. I recently started to develop a C++ parsing library. In the future, I would like to develop an UML tool with round-trip engineering capabilities. Such a library would be the cornerstone of that project. Before I started to develop it, I obviously looked for an similar project in the Internet. I've found this page : http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl?BoostSpiritC… which was out of date… until recently. Since BoostSpiritCXXParser seems to start again, I though it could be a good idea to talk about my work with the contributors of this project. First of all, here is a place where you could find a snapshot of my source code: https://git.berlios.de/cgi-bin/gitweb.cgi?p=cppparser;a=summary (Note that the most recent revisions are located on my computer, thanks to GIT). The Spirit grammar is located in src/cppparser/grammar.h file. I'd be pleased to have some comments about it. Regards. Florian

1 0

Re: [Boost-users] [Thread] Beginner question regarding thread
by Michel Lestrade 25 Aug '08

25 Aug '08

Hi, So if I understand your suggestion, I should create a container holding rays yet to be processed and then run a while loop until this queue/container is empty. Within that loop, I create a smaller number of threads by popping rays from the queue and let the threads run until they finish before going on to the next iteration. Did I understand you correctly ? It's a good idea and, from the little I know on the subject, it sounds a lot like a "thread pool". Guess I misunderstood the concept behind thread_group since I thought it was doing that already ... There's even a hardware_concurrency() function to determine the number of cores/cpus which I thought was intended to control how many threads were to be running at once. Regards, Michel Lestrade Crosslight Software > ------------------------------ > > Message: 7 > Date: Sun, 24 Aug 2008 21:56:49 -0400 > From: "Scott McMurray" <me22.ca+boost(a)gmail.com> > Subject: Re: [Boost-users] [Thread] Beginner question regarding thread > groups > To: boost-users(a)lists.boost.org > Message-ID: > <fa28b9250808241856l384711bfw658deaa79ae77d96(a)mail.gmail.com> > Content-Type: text/plain; charset=UTF-8 > > On Sun, Aug 24, 2008 at 18:58, Michel Lestrade > <michel.lestrade(a)crosslight.com> wrote: > >> > >> > I am considering rewriting part of our ray tracing code to use the Boost >> > thread library. As many of you might know, ray tracing is a task that lends >> > itself well to a parallel approach since each ray is independent from the >> > other. However, the code right now is in Fortran which doesn't support task >> > level parallelism easily ... The thread_group class seems ideally suited to >> > my needs (one thread == one ray) but there is at least one problem I >> > envision that I would like to solve before embarking on this fairly >> > time-consuming rewrite. >> > >> > > It's worth pointing out that, as I recall, you don't really want more > continually-active threads than, say, twice your number of cores. It > would be nice to run one thread per ray, but the switching and OS > management overhead will kill you if you attempt to run thousands of > threads at once, which I assume you'd need with 1:1. (Erlang, for > example, which is conceptually based on many, many communicating > processes, uses its own implementation, rather than OS threads of > processes.) > > I think you'd be much better off creating a "ray queue" of some sort, > then using a thread_group with cores+1 threads that all continually > poll the ray queue for which one to work on. > > So that's not exactly an answer to the question you asked, but I think > that's the usual wisdom with regards to threads. > > HTH, > ~ Scott > > > ------------------------------

2 1

[System] compile error
by Paul Heil 25 Aug '08

25 Aug '08

I'm trying to compile Boost 1.36.0 with Microsoft Visual Studio 9.0 for Windows Mobile 5. Unfortunately, I'm getting a compile error on the system library (below). Can anybody suggest what I may be doing wrong? The command line I'm using to compile boost is: C:\boost\boost_1_36_0>bjam --without-python --without-mpi msvc-WM5/variant=debug,release/threading=multi/*stdlib*=stlport/link=shared,static/runtime-link=static stage The output from the -d2 flag, as it relates to the System library is: file bin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi\error_code.obj.rsp "libs\system\src\error_code.cpp" -Fo"bin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\ *stdlib*-stlport\threading-multi\error_code.obj" -TP /Z7 /Od /Ob0 /W3 /GR /MTd /Zc:forScope /Zc:wchar_t /wd4675 /EHs -D_CRT_SECURE_NO_WARNINGS -D_WIN32_WCE=0x501 -DUNDER_CE=0x501 -DWIN32_WCE_PSPC -D_MSC_VER=1500 -DWINCE -DARM -D_ARM_ -DARMV4I -D_LITTLE_ENDIAN -DUNICODE -D_UNICODE -DPOCKETPC2003_UI_MODEL -c -DBOOST_ALL_NO_LIB=1 -DBOOST_SYSTEM_STATIC_LINK=1 -D_STLP_DEBUG=1 -D_STLP_DEBUG_UNINITIALIZED=1 "-I." "-IC:STLPortSTLport-5.1.5stlport" compile-c-c++ bin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi\error_code.obj cl /Zm800 -nologo @"bin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi\error_code.obj.rsp" error_code.cpp Unknown EVC++ compiler version - please run the configure tests and report the results libs\system\src\error_code.cpp(83) : error C2039: '*strerror*' : is not a member of '*stlpd_std*' libs\system\src\error_code.cpp(83) : error C3861: '*strerror*': identifier not found ...failed compile-c-c++ bin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi\error_code.obj... ...skipped <pbin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi>libboost_system-vc-mt-sgdp-1_36.lib for lack of <pbin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi>error_code.obj... ...skipped <pstage\lib>libboost_system-vc-mt-sgdp-1_36.lib for lack of <pbin.v2\libs\system\build\msvc-WM5\debug\link-static\runtime-link-static\* stdlib*-stlport\threading-multi>libboost_system-vc-mt-sgdp-1_36.lib... For reference, my user-config.jam file: using msvc : WM5 : "cl.exe" : <compileflags>-D_CRT_SECURE_NO_WARNINGS <compileflags>-D_WIN32_WCE=0x501 <compileflags>-DUNDER_CE=0x501 <compileflags>-DWIN32_WCE_PSPC <compileflags>-D_MSC_VER=1500 <compileflags>-DWINCE <compileflags>-DARM <compileflags>-D_ARM_ <compileflags>-DARMV4I <compileflags>-D_LITTLE_ENDIAN <compileflags>-DUNICODE <compileflags>-D_UNICODE <compileflags>-DPOCKETPC2003_UI_MODEL <linkflags>/subsystem:windowsce,5.01 <linkflags>/machine:arm <linkflags>/NODEFAULTLIB:oldnames.lib <linkflags>/STACK:262144,4096 <linkflags>coredll.lib <linkflags>corelibc.lib <linkflags>ole32.lib <linkflags>oleaut32.lib <linkflags>uuid.lib <linkflags>commctrl.lib <architecture>arm <instruction-set>armv4 <interface>wince <assembler>"armasm.exe" ; using stlport : : C:\STLPort\STLport-5.1.5\stlport C:\STLPort\STLport-5.1.5\lib ; Thanks, PaulH

1 0

Re: [Boost-users] [Thread] Beginner question regarding thread groups
by Michel Lestrade 25 Aug '08

25 Aug '08

Hi, Actually, I probably was not clear before: only the refracted ray will be a new ray on the queue: the reflected ray continues on its merry way until it decays. That is probably the same thing as what you are suggesting. The reason I was vague before is that I have two options: follow the reflected ray and spawn new refracted rays (as above) or follow the refracted ray ans spawn the reflected rays. I expect solution #1 is better in terms of locality but #2 might result in a smaller queue since internal reflection coefficients are small (i.e. you often wind up dropping the reflected ray since it no longer carries any significant power). Right now, the code is using #1 but I was considering switching to see what effects it might have... > It may be possible to do fairly well on locality by on a split, rather > than putting both new rays into the queue and doing a new fetch from > the queue, continuing on with the ray most similar to the incident > ray, only adding the other ray(s?) to the queue, so the frame of > reference only jumps when a ray dies. (Not unlike half-tailing > quicksort.) > =========================== We actually almost always develop in Fortran except for the GUI parts. That comes from our company background in physics and the large amount of existing code we have ... But yes, the code is old and mostly written in fixed-form F77. I am at least trying to modernize some parts to use free-form F90/F95 syntax :) I actually have not taken a look recently at these papers since I always worked directly on the code. Hopefully, there is something of interest in there. > Thanks, I always ask people to post links to their work since it makes things more interesting > and I'll look as soon as I can. I thought the code may be a bit old given the language Regards, Michel Lestrade Crosslight Software

2 1

Synchronization support for boost objects
by Zach Turner 25 Aug '08

25 Aug '08

I am wondering if there are any plans, either in the near or distant future, to ever add synchronization support for arbitrary objects. The interface could be simple, every class takes one additional template argument, the class name of a locking object which defaults to some class that provides stub methods so that by default all classes are unsychronized. Then, if I want my object to become synchronized, I can pass it whatever type of lock I deem appropriate. Of course any object can be synchronized currently by making a class that exposes the same interface and acquires a lock before forwarding to the internal boost implementation, but this is inefficient, as the internal details of the object's implementation are not always such that an entire method needs to be locked. Perhaps only a single line of code inside a long method needs to be locked. Has this ever been considered in the past?

2 1

[Thread] Beginner question regarding thread groups
by Michel Lestrade 25 Aug '08

25 Aug '08

Hi, I am considering rewriting part of our ray tracing code to use the Boost thread library. As many of you might know, ray tracing is a task that lends itself well to a parallel approach since each ray is independent from the other. However, the code right now is in Fortran which doesn't support task level parallelism easily ... The thread_group class seems ideally suited to my needs (one thread == one ray) but there is at least one problem I envision that I would like to solve before embarking on this fairly time-consuming rewrite. Just to give you some background, what we are doing is not ray tracing in the visualization/render sense: rather, it is a scientific application used to determine the light emission pattern of certain devices without waveguides (such as LEDs). So keep in mind that not all the conventional ray tracing approaches will work for me. I also come from a physics/engineering background so I am not up to date on all the newer object-oriented techniques and concepts. Now to the heart of the matter: in our approach, each of the initial rays is split into a reflected/refracted ray every time a material interface is reached which generates a secondary ray. Each secondary ray can generate other secondary rays when they hit an interface and so on ... This goes on until the power in each ray has decayed sufficiently (the material is usually absorbing), all rays exit the device or a maximum number of rays has been reached. That means that I do not know ahead of time how many rays (or threads) that I will have. This brings me to my question: can I add new threads to thread group while it is running ? All of the code samples for thread_group seem to call add_thread() on a fixed number of objects before calling join_all(). So it seems that the execution of the thread does not start at the moment it is added to the group but at the moment join_all() is called. In my case, I would have a join_all() first and then periodically add new threads: so this is basically a dynamic queue with new rays being pushed at the end. How can I start these threads automatically if add_thread does not start the execution ? Should I call join() on the new threads before adding them to the thread group instead of using join_all() ? Should I give up on thread_group and store my thread objects in some other container ? Any alternative solutions or a code sample with a dynamically changing thread_group would be appreciated. I am hoping someone has faced a similar problem in another context ... Regards, Michel Lestrade Crosslight Software P.S. Since I am a new user of Boost, I will go ahead and use the latest version (1.36.0). I already built the binaries for my tool set (VS 2008).

3 3

Re: [Boost-users] [Thread] Beginner question regarding thread groups
by Michel Lestrade 25 Aug '08

25 Aug '08

Hi, I actually looked at that first as well as another Intel Fortran solution called TBB (Thread Building Blocks). It seems to me that these approaches are made for instruction-level parallelism (i.e. processing a long loop with parallel_do or different but independent sections of code). I have a ton of other Fortran code that would benefit from this but ray tracing is not one of them. >From what I read so far, what I am looking for is task-level parallelism, not instruction-level, which is why I am looking into threads. > If you want some ideas for reference, I did a quick check on the intel > site and they have some papers and apparently even an openMP Fortran > compiler if that helps you, > [ obviously their comments are specific to their products but quite generally useful > esp if you are looking for ideas ] > http://cache-www.intel.com/cd/00/00/21/92/219292_hyperthreading_extract.pdf > > ( http://www.google.com/search?hl=en&q=site%3Aintel.com+openmp+performance+op… In our approach, light emission is isotropic (physically, it corresponds to spontaneous emission in a semiconductor) and we want to know what the extraction efficiency (how much light actually gets out) and what the output distribution is going to look like. This is affected by die shaping, external packaging, refractive index of cladding layers, position and reflectivity of metal contacts, TE/TM emission, etc... So the difference between nearby rays is critical, unlike a visualization/CG render situation. Just FYI but here is a quick reference to a few papers by my predecessor who wrote this Fortran code in the first place: http://scholar.google.ca/scholar?hl=en&lr=&q=author%3AShmatov+ray+tracing&b… We already have locality to a certain extent since each ray is bouncing around in its own "box" with a fixed number of facets and only interacts with certain predefined neighboring "boxes". So I am not too worried about cache at this point. I may have to change my mind later on though if this turns out to be a bottleneck. > While you say that rays are independent, if you do classical physical optics, > nearby rays tend to have similar trajectories etc. Rather than let an ignorant > but fair thread scheduler decide what piece of memory to access next, if you > are cache aware, you could even consider something like sorting the rays > to get the best locality and making them dependent with a transform > scheme that recognizes they are similar if nearby etc. Regards, Michel Lestrade Crosslight Software

2 1

Problem with <boost/test/unit_test.hpp> in precompiled header.
by Neutrino 25 Aug '08

25 Aug '08

Hi. I'd like a little guidance regarding the use of <boost/test/unit_test.hpp> please. I'm using mingw--gcc-3.4.5 with CodeBlocks 8.02. I'm dynamically linking to boost_unit_test_framework-mgw34-d-1_36.dll via boost_unit_test_framework-mgw34-d-1_36.lib. I have BOOST_TEST_DYN_LINK defined and I've worked out that I need to have #define BOOST_TEST_MODULE <whatever> defined in one source file prior to #include <boost/test/unit_test.hpp>. That's all good so far and the following example compiles and runs. ------------------------ // Main.cpp #define BOOST_TEST_MODULE Main #include <boost/test/unit_test.hpp> ------------------------ // Test1.cpp #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_CASE(test1) { BOOST_CHECK(true); } ------------------------ // Test2.cpp #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_CASE(test2) { BOOST_CHECK(true); } ------------------------ My problem arises when I try to include <boost/test/unit_test.hpp> in a precompiled header. ------------------------ // Pch.h #ifndef PCH_H_INCLUDED #define PCH_H_INCLUDED // Boost headers. #include <boost/test/unit_test.hpp> #endif // PCH_H_INCLUDED ------------------------ // Main.cpp #define BOOST_TEST_MODULE Main #include <boost/test/unit_test.hpp> ------------------------ // Test1.cpp #include "Pch.h" BOOST_AUTO_TEST_CASE(test1) { BOOST_CHECK(true); } ------------------------ // Test2.cpp #include "Pch.h" BOOST_AUTO_TEST_CASE(test2) { BOOST_CHECK(true); } ------------------------ The compiler is exiting with no errors but no output binary is produced. The build log reports: Info: resolving vtable for boost::unit_test::unit_test_log_tby linking to __imp___ZTVN5boost9unit_test15unit_test_log_tE (auto-import) obj\Debug\Mnf\Mnf.EventTest.o: In function `ZN5boost6detail17sp_counted_impl_pINS_9unit_test9ut_detail16callback0_impl_tINS3_6unusedEPFvvEEEED0Ev': F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x0): multiple definition of `boost::unit_test::(anonymous namespace)::unit_test_log' obj\Debug\Mnf\Mnf.BoostFunctionBindTest.o:F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x0): first defined here obj\Debug\Mnf\Mnf.EventTest.o: In function `ZN5boost6detail17sp_counted_impl_pINS_9unit_test9ut_detail16callback0_impl_tINS3_6unusedEPFvvEEEED0Ev': F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x4): multiple definition of `boost::test_tools::(anonymous namespace)::dummy_cond' obj\Debug\Mnf\Mnf.BoostFunctionBindTest.o:F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x4): first defined here obj\Debug\Pch.o: In function `ZN5boost9unit_test13test_observer12test_abortedEv': F:/My Documents/Development/Libraries/boost_1_36_0/boost/test/test_observer.hpp:(.bss+0x0): multiple definition of `boost::unit_test::(anonymous namespace)::unit_test_log' obj\Debug\Mnf\Mnf.BoostFunctionBindTest.o:F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x0): first defined here obj\Debug\Pch.o: In function `ZN5boost9unit_test13test_observer12test_abortedEv': F:/My Documents/Development/Libraries/boost_1_36_0/boost/test/test_observer.hpp:(.bss+0x4): multiple definition of `boost::test_tools::(anonymous namespace)::dummy_cond' obj\Debug\Mnf\Mnf.BoostFunctionBindTest.o:F:/My Documents/Development/Libraries/boost_1_36_0/boost/detail/sp_counted_base_gcc_x86.hpp:(.bss+0x4): first defined here I do not understand why the header file <boost/test/unit_test.hpp> objects to being included via a precompiled header file in this way. The following simple case also works fine. ------------------------ // Pch.h #ifndef PCH_H_INCLUDED #define PCH_H_INCLUDED // Boost headers. #define BOOST_TEST_MODULE Main #include <boost/test/unit_test.hpp> #endif // PCH_H_INCLUDED ------------------------ // Main.cpp #include "Pch.h" BOOST_AUTO_TEST_CASE(test1) { BOOST_CHECK(true); } ------------------------ I'll start having a look at the Boost source and see if I can work out what I'm doing wrong, but if anyone can see a problem with what I'm trying to do here I'd be grateful if they could point it out for me. Thanks Neutrino. -- View this message in context: http://www.nabble.com/Problem-with-%3Cboost-test-unit_test.hpp%3E-in-precom… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Detecting memfun staticness
by joaquin＠tid.es 25 Aug '08

25 Aug '08

Hello, although this is not strictly a Boost-related question I wonder whether some of the utilities in Boost (enable_if, type_traits) can help me solve the problem: Say I've got the following template function template<typename T> void call_f(int x) { T t; t.f(x); } whose purpose is to invoke t.f(x) on an object of type T: struct foo { void f(int){} }; struct bar { static void f(int){}; }; ... call_f<foo>(5); call_f<bar>(5); And now I'd like to optimize call_f's implementation by avoiding the creation of the T object when it is not necessary (as in bar, whose memfun f is static). The following will do for simple cases like foo and bar above: template<typename T,typename R,typename A> void call_f_impl(int x,R(*)(A)) { T::f(x); } template<typename T,typename Q> void call_f_impl(int x,Q) { T t; t.f(x); } template<typename T> void call_f(int x) { call_f_impl<T>(x,&T::f); } but fails when f is a memfun template: struct baz { template<typename T> void f(T){} }; ... call_f<baz>(5); For instance, Comeau says: "ComeauTest.c", line 17: error: no instance of overloaded function "call_f_impl" matches the argument list The argument types that you used are: (int, <unknown-type>) call_f_impl<T>(x,&T::f); ^ detected during instantiation of "void call_f<T>(int) [with T=baz]" at line 42 1 error detected in the compilation of "ComeauTest.c". Any idea about how to solve this in a more comprehensive way? Thank you in advance, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

3 2

[spirit] string_to_list_of_tuples (tr1 version)
by Markus Werle 25 Aug '08

25 Aug '08

> [gmane anti-top-poster-complaint-header] Hi! [OK, Joel, it's not as trivial as I thought. Show me the way, master] Below is my first toy implementation of string_to_list_of_tuples for tr1::tuple. Unfortunately this code only parses the text, but does not store the data. Any idea how to add the correct semantic actions in a pure and elegant way? Markus ---snip--- #define BOOST_SPIRIT_DEBUG_OUT std::cerr #define BOOST_SPIRIT_DEBUG // #include "Enforce.h" #include <iostream> #include <iterator> #include <boost/tr1/tuple.hpp> #include <boost/mpl/assert.hpp> #include <boost/mpl/at.hpp> #include <boost/spirit/include/classic_spirit.hpp> #include <list> #include <string> namespace sp = boost::spirit::classic; namespace fusion = boost::fusion; namespace mpl = boost::mpl; namespace { struct parser_traits { template <typename T, int Dummy = 0> struct parser; #define PARSER_TRAITS(TYPE, PARSER_T) \ template <int Dummy> struct parser<TYPE, Dummy> \ { \ typedef PARSER_T type; \ \ static inline \ PARSER_T p() \ { \ PARSER_T result; \ return result; \ } \ } PARSER_TRAITS(int, sp::int_parser<int>); PARSER_TRAITS(unsigned int, sp::uint_parser<unsigned>); typedef sp::real_parser <double, sp::ureal_parser_policies<double> > real_parser_t; PARSER_TRAITS(float, real_parser_t); PARSER_TRAITS(double, real_parser_t); template <int Dummy> struct parser<std::string, Dummy> { typedef sp::contiguous <sp::confix_parser <sp::strlit<const char*>, sp::kleene_star <sp::anychar_parser>, sp::strlit<const char*>, sp::unary_parser_category, sp::non_nested, sp::non_lexeme> > parser_t; static inline parser_t p() { using namespace phoenix; parser_t result = sp::lexeme_d[sp::confix_p("\"" , (*sp::anychar_p) , "\"")]; return result; } static inline sp::strlit<char const *> comma() { sp::strlit<char const *> str_p(","); return str_p; } }; }; template <typename Tuple, long I> struct rule_generator_helper { typedef typename std::tr1::tuple_element<I, Tuple>::type element_type; typedef typename parser_traits::parser<element_type> pt_type; typedef typename rule_generator_helper<Tuple, (I-1)>::type element_1_t; typedef typename sp::sequence<sp::sequence<element_1_t, sp::chlit<char> >, typename pt_type::type> type; static type apply() { return rule_generator_helper<Tuple, (I-1)>::apply() >> sp::ch_p(',') >> pt_type::p() ; } }; template <typename Tuple> struct rule_generator_helper<Tuple, 0L> { typedef typename std::tr1::tuple_element<0, Tuple>::type element_type; typedef typename parser_traits::parser<element_type> pt_type; typedef typename pt_type::type type; static type apply() { return pt_type::p(); } }; template <typename Tuple> struct rule_generator { typedef std::tr1::tuple_size<Tuple> tuple_size; BOOST_MPL_ASSERT_RELATION(tuple_size::value, >, 0); typedef rule_generator_helper<Tuple, (tuple_size::value - 1)> helper_t; static inline typename helper_t::type apply() { return helper_t::apply(); } }; template <typename Tuple> struct string_to_list_of_tuples { static std::list<Tuple> apply(std::string const & s) { std::list<Tuple> result; typedef sp::rule<sp::phrase_scanner_t> rule_t; rule_t rule_tuple_rep = '(' >> rule_generator<Tuple>::apply() >> ')'; rule_t rule_tuple_list = *(rule_tuple_rep) >> sp::end_p // for trailing whitespace ; BOOST_SPIRIT_DEBUG_RULE(rule_tuple_rep); BOOST_SPIRIT_DEBUG_RULE(rule_tuple_list); sp::parse_info<> info; try { info = sp::parse(s.c_str(), rule_tuple_list, sp::space_p); } catch (std::exception const & e) { std::cerr << e.what() << std::endl; } // ENFORCE(info.full) // ("Failed to generate a list of tuples from ") // ("string representation\n'") // (s)("'\n"); return result; } }; } // namespace int main() { try { { typedef std::tr1::tuple<int, int> tuple_t; std::list<tuple_t> test = string_to_list_of_tuples<tuple_t>::apply(" ( 1, 2 ) ( 3, 4) (5,6)"); std::copy(test.begin(), test.end(), std::ostream_iterator<tuple_t>(std::cout, "\n")); } { typedef std::tr1::tuple<unsigned int, int, double> tuple_t; std::list<tuple_t> test = string_to_list_of_tuples<tuple_t>::apply (" ( 1, -2 , 5.123) ( 3, 4,7.9)(5,6,8.6789)"); std::copy(test.begin(), test.end(), std::ostream_iterator<tuple_t>(std::cout, "\n")); } } catch (std::exception const & e) { std::cerr << e.what() << std::endl; } }

1 0

[mpl|fusion|tr1]mpl::at_c<std::tr1::tuple<...>, I>::type fails
by Markus Werle 25 Aug '08

25 Aug '08

Hi! I am using boost_1_36_0 manually patched with the files mentioned in http://thread.gmane.org/gmane.comp.lib.boost.user/39093/focus=39106 I thought that std::tr1::tuple was a valid mpl sequence, but this code here fails. Is this by design? ---snip--- #include <boost/tr1/tuple.hpp> #include <boost/mpl/assert.hpp> #include <boost/mpl/at.hpp> template <typename Tuple, long I> struct show_problem { typedef typename boost::mpl::at_c<Tuple, I>::type type; }; int main() { typedef std::tr1::tuple<int, int> tuple_t; typedef boost::mpl::at_c<tuple_t, 1L>::type Ouch; }

2 2

[tr1] maximum number of allowed tuple elements?
by Markus Werle 25 Aug '08

25 Aug '08

Hi! Maybe I overlooked this portion of the docs, so could you please give me a hint, whether and how one can set the maximum number of tuple elements. regards, Markus

2 3

[Test] newbie linking problems
by Thijs Koerselman 25 Aug '08

25 Aug '08

Hi, I'm trying to integrate the unit test framework in a library I'm developing but I'm really stuck at something. I suspect its something simple, but I've tried everything I could think of and can't seem to figure it out. My library is loaded as a 3rd party plugin in a host application. On windows it's a dll on OSX its a loadable carbon bundle. I'm currently only developing on OSX. When the bundle is loaded, its main function registers the methods that make up the public interface. So I trigger a method from the host which starts the unit testing. It's something like this #define BOOST_TEST_NO_MAIN #undef nil // because of Carbon nil definition #include <boost/test/included/unit_test.hpp> using namespace boost::unit_test; /* UnitTest class declaration and other member functions */ void UnitTest::run() { BOOST_GLOBAL_FIXTURE( MyConfig ); // setting log output stream etc. unit_test_main( &init_mylib_test_suite, 0, 0 ); } test_suite* init_mylib_test_suite( int argc, char* argv[] ) { std::string header = "Testing MyLib "; header += MYLIB_VERSION_STRING; std::string rule = std::string(header.length(),'='); BOOST_MESSAGE(rule); BOOST_MESSAGE(header); BOOST_MESSAGE(rule); return 0; } Then I'm trying to organize it so that I have different unit test suits for each logical module within the library. I'm separating the test suits into each its own compilation unit (object file output), because I don't want to recompile all the tests if I change something in one of them. So one testing module would have a source file with something like: #define BOOST_TEST_MODULE Example #define BOOST_TEST_NO_MAIN #undef nil // because of Carbon nil definition #include <boost/test/included/unit_test.hpp> using namespace boost::unit_test; BOOST_AUTO_TEST_SUITE( "MyModuleSuite" ); BOOST_AUTO_TEST_CASE( test_case1 ) { BOOST_CHECK( true ); } BOOST_AUTO_TEST_CASE( test_case2 ) { BOOST_CHECK( false ); } BOOST_AUTO_TEST_SUITE_END() With this approach I keep getting linker errors about a duplicate symbol between these 2 object files that result from it : "duplicate symbol boost::unit_test::traverse_test_tree(boost::unit_test::test_case const&, boost::unit_test::test_tree_visitor&)in ......" I've tried linking the framework as a static and a dynamic library. I've also tried including the framework by using the "BOOST_TEST_NO_LIB" flag. All without luck. Am I approaching this the wrong way? Am I missing something obvious? I would really appreciate some help with this. Tips on how to best organize the unit testing in my library are also very welcome. Note that I'm testing the internals of my library, not just the public API, so linking against the library from a simple command line app won't work. The library is also dependant on the host API. In order for those parts to work/initialize properly I'm forced to launch my plugin inside the host at all time. Cheers, Thijs

2 2

Building iostreams with zlib and bzip2
by Peter McGill 25 Aug '08

25 Aug '08

I've read all the documentation on installing/building boost and iostreams. I have built and installed zlib and libbzip2 separately. I've set CPATH and LIBRARY_PATH in the environment to point to them. I'm using mingw on windows xp from the command prompt with boost 1.36. I've tried setting the variables described in the iostreams docs, but nothing happens. I've set the variables in the environment, in user-config.jam with path-constant. Nothing works, bjam always builds boost without zlib and bzip2 support, no errors. It doesn't even attempt to build support for zlib and bzip2. Please help, how do I build boost iostreams with zlib and bzip2 support? Peter McGill

2 1

[Boost.Test] Using debuggers
by Florian Winter 25 Aug '08

25 Aug '08

When a debug version of a unit test application using Boost.Test crashes on Windows, no popup appears that asks for attaching the debugger. Instead, the process prints a stack dump to standard output and then exits normally. Attaching the debugger before the process crashes does not help. Although this is helpful when running tests automatically with no user interaction, it makes it very hard to manually debug unit tests that have crashed. According to the Boost.Test documentation, it is possible to pass framework parameters to Boost. In particular, there is a "catch system errors" parameter, which looks like it should be able to turn off this behavior (please correct me if I am wrong). http://www.boost.org/doc/libs/1_35_0/libs/test/doc/components/utf/parameter… http://www.boost.org/doc/libs/1_35_0/libs/test/doc/components/utf/parameter… We tried the following (on Windows) and then ran a test case which caused an access violation: set BOOST_TEST_CATCH_SYSTEM_ERRORS=no echo %BOOST_TEST_CATCH_SYSTEM_ERRORS The process still exited normally after printing a stack dump. We also tried using a command-line option: <TEST_APPLICATION> --catch_system_errors=no Same result as above. Is this intentional? Is there any way to prevent Boost.Test from catching system errors such as access violations (on Windows) and signals (on Linux and similar platforms) so that conventional debuggers can be used? Best regards, Florian

3 3

Re: [Boost-users] a little Graph guidance (looking for hints)
by trashing＠gmx.net 25 Aug '08

25 Aug '08

> Carsten Hoyer wrote: > > Hi > > Just turned to Boost - What a fantastic library :) > > > > I have a question about the construction of a Graph. I can't figure out how > > to use boost to represent a graph where each Vertex has several outs. F.ex > > Vertex 1 has out 1, 2 and 3. Vertex A has in a, b and c. I can then make > > edges going from 1-a, 1-b, 2-b etc. Like F.ex > > http://messe.harmony-central.com/Musikmesse05/Content/Native_Instruments/PR… > > Reaktor > > > > Any hints? > > Thank you > > Carsten > > > add_edge(node1, node2, g); > add_edge(node1, node3, g); > add_edge(node1, node4, g); I think Raindog needs vertices with different property types, i.e. vertices should contain data of different type. Am I right? -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/wasistshortview.php?mc=sv_ext_mf@gmx

1 0

Is the boost fusion documentation outdated?
by Peng Yu 25 Aug '08

25 Aug '08

Hi, The following header, <boost/fusion/include/at_c.hpp>, is mentioned in http://spirit.sourceforge.net/dl_more/fusion_v2/libs/fusion/doc/html/fusion…. But I don't this header in the boost installation on my machine. Is the documentation outdated or boost on my machine outdated? Thanks, Peng

3 2

Passing an argument to threads (boost::thread)
by Jean-Sebastien Stoezel 25 Aug '08

25 Aug '08

Hello, How to pass arguments to a thread? From the documentation I thought I could do this: mp_ProcessorListenerThread = new boost::thread(ProcessorListener, this); However I get the following error: 'boost::thread::thread' : no overloaded function takes 2 arguments Thanks, Jean

6 7

[serialization] Uninformative exception thrown for binary_archive
by Sebastian.Karlsson＠mmorpgs.org 25 Aug '08

25 Aug '08

Hi, I'm getting an exception thrown in basic_binary_iprimitive.hpp at line 154 ( load_binary ). I don't have a clue what the error really is, but from debugging I can see it happen when the member _IGcount and _IPcount of m_sb reaches 0. Is it perhaps the buffer that runs out of memory? I'm simply using binary_[o/i]archive as a drop in replacement for the text variants for serialization. I'm using 1.36 and MSVC 8.0.

2 1

How to do accumulation on boost::fusion::vector
by Peng Yu 24 Aug '08

24 Aug '08

Hi, I want to do accumulation on boost::fusion::vector. Please refer to the comment in the code. Can somebody let me know how to do it? Thanks, Peng #include <boost/fusion/sequence.hpp> #include <boost/fusion/include/at.hpp> #include <boost/fusion/container.hpp> #include <boost/fusion/iterator.hpp> #include <boost/typeof/typeof.hpp> #include <iostream> struct A { A(int value) : _value(value) { } int _value; }; struct B { B(int value) : _value(value) { } int _value; }; struct C { C(int value) : _value(value) { } int _value; }; template <typename T> int get_it(const T &t) { return t._value; } int main() { BOOST_AUTO(v, boost::fusion::make_vector(A(0), B(1), C(3), B(8), C(7))); // I want to make the following code more generic, the fusion vector length could be anything. // It sum over all the vector elements. int sum = get_it(boost::fusion::at_c<0>(v)) + get_it(boost::fusion::at_c<1>(v)) + get_it(boost::fusion::at_c<2>(v)) + get_it(boost::fusion::at_c<3>(v)) + get_it(boost::fusion::at_c<4>(v)); std::cout << sum << std::endl; }

2 1

STL/fusion/MPL/graph/.. cheat-sheets?
by Zeljko Vrba 23 Aug '08

23 Aug '08

Maybe somebody has already asked this question.. Does there exist a cheat sheet / quick reference that would give a compressed overview of different concept requirements? Something like http://www.hoomanb.com/cs/QuickRef/CppQuickRef.pdf except that it would document concepts and requirements? For example, http://www.boost.org/doc/libs/1_36_0/libs/fusion/doc/html/fusion/sequence/c… in a bit more compressed format, grouped together with other concept tables.

2 1

mpl::switch_
by joel falcou 23 Aug '08

23 Aug '08

Is it me or is mpl::swich_ both undocumented and non-working ? I managed to get this : template<typename Body, typename T> struct switch_ { typedef typename find_if<Body,apply1<first<_1>,T> >::type iter_; typedef typename deref<iter_>::type pair_; typedef typename second<pair_>::type f_; typedef typename apply1< f_, T>::type type; }; working properly. Seems the current swich_ use lambda<> where it shouldn't. Is this a known bug or is there something bound to replace switch_ in the near future ? SHould I fill a trac ticket about this ?

2 4

[function] creating function object for member function call delegation
by Olaf Peter 23 Aug '08

23 Aug '08

Hi, the following code doesn't compile: ---8<--- #include <iostream> #include <boost/function.hpp> #include <boost/bind.hpp> using namespace std; using namespace boost; struct foo { function<void ( const string& )> fn; }; struct bar { foo& m_f; bar(foo& f) : m_f( f ) { m_f.fn = bind( bar::write, this); // L20 } void write( const string& s ) { cout << s; } }; int main() { foo f; bar b( f ); f.fn( "Hello\n" ); } --->8--- class foo (object f) gets called with the string argument and shall delegate the call to the write function of bar. Writing bind( bar::write, this, _1) doesn't make things better. Here the error message: $ LANG=en g++ callback.cpp -o callback callback.cpp: In constructor 'bar::bar(foo&)': callback.cpp:20: error: no matching function for call to 'bind(<unresolved overloaded function type>, bar* const)' /usr/include/boost/bind.hpp:1383: note: candidates are: boost::_bi::bind_t<boost::_bi::unspecified, F, typename boost::_bi::list_av_1<A1>::type> boost::bind(F, A1) [with F = void (bar::*)(const std::string&), A1 = bar*] /usr/include/boost/bind.hpp:1628: note: boost::_bi::bind_t<typename boost::_bi::dm_result<M T::*, A1>::type, boost::_mfi::dm<M, T>, typename boost::_bi::list_av_1<A1>::type> boost::bind(M T::*, A1) [with A1 = bar*, M = void ()(const std::string&), T = bar] Thanks for help, Olaf

3 3

[Config] Reusing results of platform/compiler detection
by joel falcou 23 Aug '08

23 Aug '08

I've looked at the inner implementation of the config components but I didn't find any way to get a simple preprocessor indicating which compiler/platform have been detected. Basiaclly, my goal is to use boost comprehensive detection features to select my own platform/compiler code selection through #ifdef. I tried using Preprocessor magic over the BOOST_PLATFORM string but not to avail. Is there a way to do such a thing or can we request for a #define BOOST_DETECTED_XXX_PLATFORM/COMPILER definition ?

1 0

boost::serialization of a template class
by Takeo 23 Aug '08

23 Aug '08

Hello, I have a base class: class base {...}; and a derived class: class derived : public base {...}; and a template class: template <class T> class ElementList { vector<T*>* m_Elements; ... friend class boost::serialization::access; template<class Archive> inline void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(m_Elements); } } Now i create a class: class myclass : public ElementList<base> { friend class boost::serialization::access; template<class Archive> inline void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_BASE_OBJECT_NVP(ElementList); } } I want to use boost::serialization on myclass I create somwhere in my program several "derived" elements (which are serialized at this point) After, I put them in the m_Elements vector of myclass (hérited from ElementList<base>). So, the "derived" are stored as "base" elements and the problem is that they are serialized as "base" elements and not "derived" elements. Have you got an idea of how I can solve this issue? Thanks by Advance Emmanuel -- View this message in context: http://www.nabble.com/boost%3A%3Aserialization-of-a-template-class-tp191093… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

How to build specified type of library with bjam?
by Bill David 23 Aug '08

23 Aug '08

1.I have checked the Jamroot.jam file, it seems we can specify a build-type of minimal or complete. e.g. bjam --toolset=msvc-9.0 --with-regex --build-type=complete stage But it can not meet my requirement. For example, I just want to build debug release <threading>multi <link>static. How could I specify it? 2.it seems we can specify link/runtime-link options, what's the difference between link static and runtime-link static? Does libboost_regex-vc90-mt-gd-1_35.lib means it's a multi-threading debug version of static link edition of regex library? But how about libboost_date_time-vc90-mt-sgd-1_35.lib? What does the -sgd mean? What's the naming convention?

2 1

Does Program Options need pre-building?
by Scott Morgan 23 Aug '08

23 Aug '08

The Program Options lib is listed in the documentation as header only, but I'm getting error messages when attempting to build against it. In particular it's complaining about not having 'libboost_program_options-vc71-sgd-1_36.lib' Scott

3 3

Which library for saving numerical output?
by e r 23 Aug '08

23 Aug '08

Hi all, A while ago I thought I had seen a library whose purpose is to save numerical output (say for plotting graphs using a separate software). Unfortunately I didnt bookmark it and after browsing a few times the boost and vault contents can't find it. Anyone knows what I'm talking about? Thanks!.

1 0

dereference problems with iterator_facade
by e r 23 Aug '08

23 Aug '08

hello, i have a build error when i try to use an iterator derived from iterator_facade<Iter> with Iter an iterator over a const Range. It's the const that causes the error. See the *.hpp below and an example. any other suggestion to improve the *hpp also appreciated. thanks! /////////////////////////////////////////////////////////////////////////////// // boost::skip_one_step_iterator.hpp // // Copyright 2008 Erwann Rogard. Distributed under the Boost // Software License, Version 1.0. (See accompanying file // LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) #ifndef BOOST_ITERATOR_SKIP_ONE_STEP_ITERATOR_HPP_ER200808 #define BOOST_ITERATOR_SKIP_ONE_STEP_ITERATOR_HPP_ER200808 #include <iterator> #include <stdexcept> #include <boost/iterator/iterator_facade.hpp> #include <boost/iterator/iterator_traits.hpp> #include <boost/range.hpp> #include <boost/assert.hpp> #include <boost/type_traits.hpp> namespace boost{ namespace{ template<typename BaseIter,typename Derived> struct super{ typedef iterator_facade< Derived, typename iterator_value<BaseIter>::type, typename iterator_category<BaseIter>::type> type; }; } template<typename BaseIter> class skip_one_step_iterator : public super<BaseIter,skip_one_step_iterator<BaseIter> >::type { typedef iterator_range<BaseIter> iterator_range_type; typedef typename super<BaseIter,skip_one_step_iterator>::type super_type; public: typedef typename super_type::difference_type difference_type; typedef typename super_type::reference reference; skip_one_step_iterator(){} explicit skip_one_step_iterator( BaseIter i, BaseIter skip ) :base_iter(i),skip_(skip){ if(base_is_skip()){ throw std::runtime_error( "skip_one_step_iterator: base_is_skip at construction" ); } } private: friend class boost::iterator_core_access; void increment() { advance((difference_type)(1)); } difference_type distance_to(skip_one_step_iterator const& other) const{ difference_type d = std::distance(this->base_iter,other.base_iter); difference_type dist_this = std::distance(this->base_iter,skip_); difference_type dist_other = std::distance(other.base_iter,skip_); bool same_sign = ((dist_this<0) && (dist_other<0)) || ((dist_this>0) && (dist_other>0)); if(!same_sign){ if(d<(difference_type)(0)){ d+=(difference_type)(1); }else{ d-=(difference_type)(1); } } return d; } bool equal(skip_one_step_iterator const& other) const { return this->base_iter == other.base_iter; } reference dereference() const { return *base_iter; } void advance(difference_type n) { std::advance(this->base_iter,1); if(base_is_skip()){ std::advance(this->base_iter,1); } } bool base_is_skip()const{ return ((this->base_iter)==(this->skip_)); } BaseIter base_iter; BaseIter skip_; }; template<typename Iter> skip_one_step_iterator<Iter> make_skip_one_step_begin_iterator(Iter b,Iter skip){ Iter new_b = b; if(new_b == skip){ std::advance(new_b,1); } return skip_one_step_iterator<Iter>(new_b,skip); }; template<typename Iter> skip_one_step_iterator<Iter> make_skip_one_step_end_iterator(Iter e,Iter skip){ // constructor below will throw if skip == e return skip_one_step_iterator<Iter>(e,skip); }; } #endif #include <iostream> #include <vector> #include <iterator> #include <boost/range.hpp> #include <boost/assign/std/vector.hpp> #include <boost/iterator/skip_one_step_iterator.hpp> #include <libs/iterator/example/skip_one_step_iterator.h> template<typename R> void func(const R& range,std::size_t i){ //void func(R& range,std::size_t i)//works fine //typedef R range_type;//works fine typedef const R range_type; typedef typename boost::range_iterator<range_type>::type iter_type; typedef boost::skip_one_step_iterator<iter_type> skipr_iter_type; iter_type b = boost::begin(range); iter_type e = boost::end(range); iter_type skip = b; std::advance(skip,i); skipr_iter_type sb = boost::make_skip_one_step_begin_iterator(b,skip); skipr_iter_type se = boost::make_skip_one_step_end_iterator(e,skip); copy( sb, se, std::ostream_iterator<double>(std::cout<<" "," ") ); std::cout << std::endl; }; void example_skip_one_step_iterator(){ using namespace boost; typedef std::vector<double> vector_type; typedef vector_type::size_type size_type; typedef range_iterator<vector_type>::type iter_type; typedef skip_one_step_iterator<iter_type> skipr_iter_type; vector_type vec; { using namespace boost::assign; vec += 0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0; } //invalid initialization of reference of type ‘double&’ from expression of type ‘const double’| //this error disappears if range_type replaced by R in definition of func for(size_type i = 1; i < size(vec); i++){ std::cout << "i=" << i << std::endl; func(vec,i); } };

2 2

serialization of bitsets
by McKay, Steve F. 23 Aug '08

23 Aug '08

I submitted code to serialize std::bitsets in 2006 (Boost version 1.33.1). The code enabled tight packing (1 true/false saved as 1 bit) not otherwise available in Boost.Serialization. Robert seemed to be busy at the time (isn't he always :) ), so my code did not get into Boost.Serialization. I have updated to Boost version 1.35.0, so I needed to update my add-on code. I am prepared to submit my revised code to serialize std::bitsets, if anyone is interested. Steve F. McKay Social Security Administration Room 760 Altmeyer 6401 Security Boulevard Baltimore, Maryland 21235 (410)965-3014

2 1

Re: [Boost-users] boost webclient
by Jim Bell 22 Aug '08

22 Aug '08

lakers fan wrote: > Hello, Does anyone know whether there is a class for webclient that > sends data to or receive data from a URL??. You may want to look into curl. http://curl.haxx.se

1 0

Re: [Boost-users] [function] [stlport] typeinfo in stlport 5.1.5
by Gulevich Vasili Victorovich 22 Aug '08

22 Aug '08

On Sat, Aug 23, 2008 at 1:46 AM, Peter Dimov <pdimov(a)pdimov.com> wrote: >> Is there an appropriate way to tell BOOST or stlport where type_info is? > The correct macro to test is BOOST_NO_STD_TYPEINFO: > I don't know whether it's defined properly on your configuration; can you > check? Modification: ---------function_base.hpp--------- #ifdef BOOST_NO_STD_TYPEINFO #warning BOOST_NO_STD_TYPEINFO defined #endif #ifdef _STLP_INTERNAL_TYPEINFO #warning _STLP_INTERNAL_TYPEINFO defined #endif ---------function_base.hpp--------- gives: function_base.hpp:48:2: warning: #warning _STLP_INTERNAL_TYPEINFO defined meaning BOOST_NO_STD_TYPEINFO is undefined there. _STLP_INTERNAL_TYPEINFO is used in file stlport/typeinfo and means that stlport don't override system (mingw's) typeinfo. Mingw defines typeinfo inside std namespace. Replacing BOOST_NO_EXCEPTION_STD_NAMESPACE check in function_base.hpp with BOOST_NO_STD_TYPEINFO one does the trick. BTW, the only files in boost distribution that contain BOOST_NO_STD_TYPEINFO substring are boost/detail/sp_typeinfo.hpp, docs and tests. Is that ok? -- Гулевич Василий

1 0

[function] [stlport] typeinfo in stlport 5.1.5
by Gulevich Vasili Victorovich 22 Aug '08

22 Aug '08

Hello. I'm using stlport-5.1.5 + mingw32 and trying to compile thread library from boost-1.36. The following error appears: --------Compiler output-------- ./boost/function/function_base.hpp:In static member function `static void boost::detail::function::reference_manager<F>::get(const boost::detail::function::function_buffer&, boost::detail::function::function_buffer&, boost::detail::function::functor_manager_operation_type)': ./boost/function/function_base.hpp:224: error: expected initializer before '&' token --------Compiler output-------- The context in function_base.hpp: 224: const BOOST_FUNCTION_STD_NS::type_info& check_type = 225: *static_cast<const BOOST_FUNCTION_STD_NS::type_info*>(out_buffer.const_obj_ptr); BOOST_FUNCTION_STD_NS is defined as empty because of BOOST_NO_EXCEPTION_STD_NAMESPACE in the same file: function_base.hpp on line 41. BOOST_NO_EXCEPTION_STD_NAMESPACE is found in: --------boost/config/stdlib/stlport.hpp-------- // If STLport is using it's own namespace, and the real names are in // the global namespace, then we duplicate STLport's using declarations // (by defining BOOST_NO_STDC_NAMESPACE), we do this because STLport doesn't // necessarily import all the names we need into namespace std:: // # if (defined(__STL_IMPORT_VENDOR_CSTD) \ || defined(__STL_USE_OWN_NAMESPACE) \ || defined(_STLP_IMPORT_VENDOR_CSTD) \ || defined(_STLP_USE_OWN_NAMESPACE)) \ && (defined(__STL_VENDOR_GLOBAL_CSTD) || defined (_STLP_VENDOR_GLOBAL_CSTD)) # define BOOST_NO_STDC_NAMESPACE # define BOOST_NO_EXCEPTION_STD_NAMESPACE # endif ------boost/config/stdlib/stlport.hpp------ (_STLP_USE_OWN_NAMESPACE and vendor flags is defined in this case) Replacing function_base.hpp:41 with: ----- #if defined(BOOST_NO_EXCEPTION_STD_NAMESPACE) && !defined(_STLP_USE_OWN_NAMESPACE) ----- allows compilation of thread library. Is there an appropriate way to tell BOOST or stlport where type_info is? -- Sincerely, Basilevs

2 1

Query regarding insert function in boost::unordered_set
by Carvalho, Malcolm 22 Aug '08

22 Aug '08

Hi, I need to use boost::unordered_set for inserting about 100000 elements with a constant time efficiency for the insert function. I checked the time taken for inserting 25,000 elements each time, and found that it is as follows: Time for inserting 1-25000 elements: 250 Time for inserting 25001-50000 elements: 250 Time for inserting 50001-75000 elements: 171 Time for inserting 75001-100000 elements: 297 I am using the std::pair<iterator, bool> boost::insert(const value_type& obj) function. Please refer the attached project for details. Is there any way to make this insertion time constant irrespective of the number of elements already present in the set? Regards, Malcolm Carvalho Malcolm Carvalho Senior Software Engineer, Mumbai Delivery Centre malcolm.carvalho(a)geometricglobal.com Geometric Limited T +91.22.67056539 F +91.22.67056891 M +91.9833143936 www.geometricglobal.com <http://www.geometricglobal.com/>

2 1

Regex as shared library with MinGW
by Jens Steinhauser 22 Aug '08

22 Aug '08

Hi, I was surprised to see that Boost.Regex doesn't compile to a shared library when using MinGW. The following lines in lib/regex/build/Jamfile.v2 prevent it: <toolset>gcc-mw:<link>static <toolset>gcc-mingw:<link>static <toolset>gcc-cygwin:<link>static After commenting them out, the library builds as a Dll. I searched for a reason for these 3 lines, but couldn't find anything. Does anyone on this list know why they are there? The program I'm porting to windows works fine with the Dlls build with MinGW. Thanks.

2 1

boost webclient
by lakers fan 22 Aug '08

22 Aug '08

Hello, Does anyone know whether there is a class for webclient that sends data to or receive data from a URL??. Thanks, Lakers. _________________________________________________________________ Get thousands of games on your PC, your mobile phone, and the web with Windows®. http://clk.atdmt.com/MRT/go/108588800/direct/01/

2 1

EqualityComparable of unordered containers
by Daniel Krügler 22 Aug '08

22 Aug '08

Hello, I would like express my respect and congratulations for the provision of boost 1.36 (which had an impressive short release cycle)! One of the very promising new contributions is Boost.Unordered and I I would like to make a comment on the provision of operator==/operator!= as explained in http://www.boost.org/doc/libs/1_36_0/doc/html/unordered/rationale.html#unor… I totally agree that there exist valid and reasonable use-cases for EqualityComparable for unordered containers. In fact many unordered containers (usually named hash containers) exists in one or the other language or user-defined library which provide some form of "unordered" equality comparison, which reflects the unordered nature of the container itself and thus matches a typical user expectation. Nevertheless there exists an unfortunate side-effect of the current implementation of operator== and operator!=, which should be mentioned in the documentation IMO. To keep a long story short: Under some given conditions, EqualityComparable of Boost.Unordered containers will violate the very basic "symmetry" property of EqualityComparable. This can happen, if the equality predicate key_eq() is a runtime-predicate and if the actual comparison is done between two different runtime-comparators. Note that associative containers - even though they also support this kind of runtime/dynamic comparison - are *not* sensitive to this unexpected behaviour. The simple reason is that the implementation uses an asymmetric algorithm to realize the concept. The program provided at tghe end of this posting demonstrates this kind of violation of the usual symmetric property of EqualityComparable. The output should be: s1 == s2: 1 s2 == s1: 0 I hope this comment is not understood as some form of destructive criticism. AFAIK there does not exist any general valid algorithm which can realize a pure *symmetric* unordered equality comparison with an average linear complexitity. My point is, that I strongly propose to mention this unexpected property and/or to declare code as causing undefined behaviour if it attempts to use EqualityComparable of unordered containers if the predicate is dynamic and different between the compared containers. Greetings from Bremen, Daniel #include <cctype> #include <iostream> #include <ostream> #include <boost/unordered_set.hpp> typedef boost::unordered_set<char, std::size_t(*)(char), bool (*)(char, char)> MySet; std::size_t hash_char(char c) { return static_cast<std::size_t>(c); } bool eq_char(char c1, char c2) { return c1 == c2; } std::size_t hash_char_ic(char c) { return static_cast<std::size_t>(std::tolower(c)); } bool eq_char_ic(char c1, char c2) { return std::tolower(c1) == std::tolower(c2); } int main() { MySet s1(1, hash_char, eq_char); s1.insert('A'); MySet s2(1, hash_char_ic, eq_char_ic); s2.insert('a'); bool t1 = s1 == s2; std::cout << "s1 == s2: " << t1 << std::endl; bool t2 = s2 == s1; std::cout << "s2 == s1: " << t2 << std::endl; std::cin.get(); }

4 13

how to make fusion::pair less-than-comparable?
by Zeljko Vrba 22 Aug '08

22 Aug '08

Related to my previous question: fusion::map is a sequence and instances of it could be directly compared with operator < , thus vectors of instances also sorted if fusion::pair supported relational operators other than == and != (which it doesn't). So, two questions: 1. What is the recommended way of defining operator< for particular pair instances (e.g. for pair<net, int>) ? 2. An explanation of why doesn't fusion::pair define at least operator< would be welcome. Thanks. (Yes, I know, I can pass an explicit predicate to std::sort.. )

2 4

Re: [Boost-users] Boost serialization..
by Sangeeta Singh 22 Aug '08

22 Aug '08

Hi, Thanks for the pointer. Is the string serialization for 64 bit and 32 bit library different as the string being serialized using 64bit version can not be read using 32 bit version. Thanks, Sangeeta Message: 1 Date: Thu, 21 Aug 2008 13:10:26 -0800 From: "Robert Ramey" <ramey(a)rrsd.com> Subject: Re: [Boost-users] Boost serialization.. To: boost-users(a)lists.boost.org Message-ID: <g8ki3p$iv7$2(a)ger.gmane.org> Content-Type: text/plain; charset="iso-8859-1" THe library includes a portable_binary archive as a demo. There are also other versions of this in boost downloads. Robert Ramey "Sangeeta Singh" <singhs(a)yahoo-inc.com> wrote in message news:228C81146E16054AA04E582D4B8C34CD96C888@SNV-EXVS08.ds.corp.yahoo.com ... Hi, I have following requirement Application doing the object serialization (binary stream with char trait) uses 64 bit boost serialization version Application deserializing the object uses 32 bit boost serialization version Both are incompatible. Is there a boost serialization library available that works for both 64 and 32 bit m/c or is there any other way that I can achieve this. Thanks, Sangeeta

2 1

Serialization of derived classes with boost 1.34.1
by Hansi 22 Aug '08

22 Aug '08

Hello, I have a problem to serialize a derived class. In some classes I use a shared_ptr<A> which acts as a interface for some other classes. Til now I had the implementation in the header and it worked fine with the BOOST_CLASS_EXPORT(B) declaration after at the end of the header file where B was defined. For performance reason (compile time) I have moved the serialization code from the header file (there is only the declaration) to the .cpp file (now I use also only the polymorphic_binary_archive). There I have also the declaration for the polymorphic_binary_archive template void B::serialize(boost::archive::polymorphic_iarchive& ar, const unsigned int file_version); template void B::serialize(boost::archive::polymorphic_oarchive& ar, const unsigned int file_version); After linking I have now some linker errors. First for the class B which is derived from A and for the class A. All linker errors use the binary_archive. If I declare the "BOOST_CLASS_EXPORT(B)" at the end in the .cpp file I get only an linker error for class A. How I can solve this? Best regards Hansjörg

3 3

Re: [Boost-users] Fusion: some concrete questions about maps and vectors
by Zeljko Vrba 22 Aug '08

22 Aug '08

On Thu, Aug 21, 2008 at 09:39:01PM -0400, Daryle Walker wrote: > > Are the two vectors supposed to represent the same relation? If so, > could Boost's BiMap help? > No, they represent the two halves of a symmetric relation.

1 0

Fusion: some concrete questions about maps and vectors
by Zeljko Vrba 22 Aug '08

22 Aug '08

I'm operating on two vectors of pairs of numbers, currently represented by std::pair<unsigned, unsigned>. The first vector is a list of (net,pin) pairs, while the second vector is a list of (pin, net) pairs. In order to raise a bit the understandability / abstraction level of my code, I think that it'd be nice to be able to write at_key<pin>(x) and at_key<net>(x) instead of x.first or using tie(). Looking at fusion manual, it seems that of all the containers, map suits the purpose: struct pin; struct net; First question: are Map1 and Map2 different types? typedef map<pair<pin, unsigned>, pair<net, unsigned> > Map1; typedef map<pair<net, unsigned>, pair<pin, unsigned> > Map2; According to a simple test program that I wrote, which basically boils down to Map m1; Map m2; m1 = m2; they, are different because the compilation fails. Hmm, so far, so good. Second question: some algorithms are common both for both types, e.g. sort, where I want to have lexicographical comparison between the elements. Is it possible to somehow transform Map1 and Map2 so that they both appear as typedef vector<unsigned, unsigned> > Pair with proper lexicographical ordering? OK, there are conversion metafunctions, but I want something more; the following code std::vector<Map1> m; std::sort(m.begin(), m.end()); should process vector elements (Map1s) _as_if_ they were Pairs or ordinary std::pair. I see there are conversion metafunctions defined in documentation, but I don't know how to use them to convert std::vector<Map1> to std::vector<Pair> when giving m as argument to sort() [if this is possible at all]. Sorry for the long post and if it is somewhat vague. === (I could probably avoid using fusion by manually defining two different struct types, each holding pin/net members and defining operator< for each, but I want to finally get my hands dirty with fusion on a concrete, not too complex problem that I'm facing right now.)

3 2

Boost-users] [python]How to pass C++ object reference to python code?
by Neil Fang 22 Aug '08

22 Aug '08

I'm trying embedding python using boost::python. Just as the same way as the boost::python quick start "embedding.cpp". I have a C++ base class, with a virtual function do something on another C++ object: struct MyClass { void printThis() { std::cout << "0x" << this << std::endl; } }; // An abstract base class class Base : public boost::noncopyable { public: virtual ~Base() {}; virtual void printMyClass(MyClass& myObj) = 0; }; Than, I want to write some Python derived class code: "from embedded_hello import * \n" "class PythonDerived(Base): \n" " def printMyClass(self, myObj): \n" " myObj.printThis() \n" But I found that the "myObj" passed to python code is another obj, not the object created in C++ code. How can use the object created in C++ object rather than a copyed object? Thanks! The full code list below, which is modified form boost::python example. // Copyright Stefan Seefeld 2005. // Distributed under the Boost Software License, Version 1.0. (See // accompanying file LICENSE_1_0.txt or copy at // http://www.boost.org/LICENSE_1_0.txt) #include <boost/python.hpp> #include <boost/detail/lightweight_test.hpp> #include <iostream> namespace python = boost::python; struct MyClass { void printThis() { std::cout << "0x" << this << std::endl; } }; // An abstract base class class Base : public boost::noncopyable { public: virtual ~Base() {}; virtual std::string hello() = 0; virtual void printMyClass(MyClass& myObj) = 0; }; // C++ derived class class CppDerived : public Base { public: virtual ~CppDerived() {} virtual std::string hello() { return "Hello from C++!";} virtual void printMyClass(MyClass& myObj) { myObj.printThis(); } }; // Familiar Boost.Python wrapper class for Base struct BaseWrap : Base, python::wrapper<Base> { virtual void printMyClass(MyClass& myObj) { this->get_override("printMyClass")(myObj); } virtual std::string hello() { #if BOOST_WORKAROUND(BOOST_MSVC, <= 1300) // workaround for VC++ 6.x or 7.0, see // http://boost.org/libs/python/doc/tutorial/doc/html/python/exposing.html#pyt… return python::call<std::string>(this->get_override("hello").ptr()); #else return this->get_override("hello")(); #endif } }; // Pack the Base class wrapper into a module BOOST_PYTHON_MODULE(embedded_hello) { python::class_<BaseWrap, boost::noncopyable> base("Base"); python::class_<MyClass>("MyClass") .def("printThis", &MyClass::printThis) ; } void exec_test() { std::cout << "registering extension module embedded_hello..." << std::endl; // Register the module with the interpreter if (PyImport_AppendInittab("embedded_hello", initembedded_hello) == -1) throw std::runtime_error("Failed to add embedded_hello to the interpreter's " "builtin modules"); std::cout << "defining Python class derived from Base..." << std::endl; // Retrieve the main module python::object main = python::import("__main__"); // Retrieve the main module's namespace python::object global(main.attr("__dict__")); // Define the derived class in Python. python::object result = python::exec( "from embedded_hello import * \n" "class PythonDerived(Base): \n" " def hello(self): \n" " return 'Hello from Python!' \n" " def printMyClass(self, myObj): \n" " myObj.printThis() \n" , global, global); python::object PythonDerived = global["PythonDerived"]; MyClass myObj; // Creating and using instances of the C++ class is as easy as always. CppDerived cpp; BOOST_TEST(cpp.hello() == "Hello from C++!"); cpp.printMyClass(myObj); std::cout << "testing derived class from C++..." << std::endl; // But now creating and using instances of the Python class is almost // as easy! python::object py_base = PythonDerived(); Base& py = python::extract<Base&>(py_base) BOOST_EXTRACT_WORKAROUND; // Make sure the right 'hello' method is called. BOOST_TEST(py.hello() == "Hello from Python!"); py.printMyClass(myObj); std::cout << "success!" << std::endl; } void exec_file_test(std::string const &script) { std::cout << "running file " << script << "..." << std::endl; // Run a python script in an empty environment. python::dict global; python::object result = python::exec_file(script.c_str(), global, global); // Extract an object the script stored in the global dictionary. BOOST_TEST(python::extract<int>(global["number"]) == 42); std::cout << "success!" << std::endl; } void exec_test_error() { std::cout << "intentionally causing a python exception..." << std::endl; // Execute a statement that raises a python exception. python::dict global; python::object result = python::exec("print unknown \n", global, global); std::cout << "Oops! This statement should be skipped due to an exception" << std::endl; } int main(int argc, char **argv) { //BOOST_TEST(argc == 2); //std::string script = argv[1]; // Initialize the interpreter Py_Initialize(); bool error_expected = false; if ( python::handle_exception(exec_test) ) { if (PyErr_Occurred()) { if (!error_expected) BOOST_ERROR("Python Error detected"); PyErr_Print(); } else { BOOST_ERROR("A C++ exception was thrown for which " "there was no exception translator registered."); } } std::cin.get(); // Boost.Python doesn't support Py_Finalize yet, so don't call it! return boost::report_errors(); }

2 1

boost.python - undefined symbol error
by antonyc 22 Aug '08

22 Aug '08

I've just started this morning with boost::python and have been struggling all day to get the basic http://www.boost.org/doc/libs/1_35_0/libs/python/doc/tutorial/doc/html/pyth… 'exposing classes' doc/tutorial: to link correctly under gcc 4.1.2. I would greatly appreciate any help as I'm sure it's probably something simple and I'm pretty much at a loss at the moment. So the source file compiles and the shared object appears to link ok however when I try and import the module, python reports the following missing symbol: ImportError: /test/BoostPython/hello.so: undefined symbol: _ZN5boost6python9converter8registry6insertEPFPvP7_objectEPFvS5_PNS1_30rvalue_from_python_stage1_dataEENS0_9type_infoE de-mangled courtesy of c++filt: boost::python::converter::registry::insert(void* (*)(_object*), void (*)(_object*, boost::python::converter::rvalue_from_python_stage1_data*), boost::python::type_info) Indeed a search through the boost.python headers yields no such function prototype. Nor does searching the symbols defined in libboost_python-gcc41-mt.so. The closest match I can find in libboost_python-gcc41-mt.so is: boost::python::converter::registry::insert(void* (*)(_object*), void (*)(_object*, boost::python::converter::rvalue_from_python_stage1_data*), boost::python::type_info, _typeobject const* (*)()) which appears to match the following function in python/converter/registry.hpp: // Insert an rvalue from_python converter BOOST_PYTHON_DECL void insert( convertible_function , constructor_function , type_info , PyTypeObject const* (*expected_pytype)() = 0 ); Beyond that I'm pretty much stuck. Thanks in advance to anyone who can help me out! antony /*------------------- hello.cpp -------------------*/ #include <string> struct World { void set(std::string msg) { this->msg = msg; } std::string greet() { return msg; } std::string msg; }; #include <boost/python.hpp> using namespace boost::python; BOOST_PYTHON_MODULE(hello) { class_<World>("World") .def("greet", &World::greet) .def("set", &World::set) ; } /*-----------------------------------------------------------*/ g++ -c -g -Wall -fPIC -pipe -DBOOST_PYTHON_MAX_ARITY=20 -I. -I/usr/local/include -I/usr/local/include/python2.5 hello.cpp g++ -shared -o hello.so hello.o -L/usr/local/lib -lpython2.5 -lboost_python-gcc41-mt -- View this message in context: http://www.nabble.com/boost.python---undefined-symbol-error-tp19054023p1905… Sent from the Boost - Users mailing list archive at Nabble.com.

4 3

re: Wrapping a library
by Alan Baljeu 21 Aug '08

21 Aug '08

>> I have this library (one class with 700 methods) I want to wrap all the calls, so I can log the calls and do other stuff. Using emacs macros and stuff, I >>can create a proxy class which has the same signature, but each method calls a C macro which calls the original class method, and does other stuff. > A class or a namespace? A document class. Yes it's huge. >> >> I'm thinking there must be a better way. I'm thinking BOOST_PP library could abstract the above macros into something concise. However, the >> library is intimidating. There are lots of features, and I'm not clear on how I should get started. >For ad hoc code generation, I usually end up with awk and sed. I just did this wrapping >thing with a java class definition ( no actual source code) and could pretty much use lynx to render the javadoc html, extract >method signatures, and use awk to generate methods with sed and finally >vi to touch stuff up. The generated class contained an instance of the original class that >was simply called in each generated method after my custion code,if any, executed. >I wasn't just interested in wrapping as I also made a fake version for >unsupported operations. That does seem to be a good approach. Thanks for the idea. __________________________________________________________________ Yahoo! Canada Toolbar: Search from anywhere on the web, and bookmark your favourite sites. Download it now at http://ca.toolbar.yahoo.com.

1 0

Boost serialization..
by Sangeeta Singh 21 Aug '08

21 Aug '08

Hi, I have following requirement Application doing the object serialization (binary stream with char trait) uses 64 bit boost serialization version Application deserializing the object uses 32 bit boost serialization version Both are incompatible. Is there a boost serialization library available that works for both 64 and 32 bit m/c or is there any other way that I can achieve this. Thanks, Sangeeta

2 1

Decoding hex stream of bytes to binary form
by Mateusz Loskot 21 Aug '08

21 Aug '08

Hi, I've a small function (see below) to convert hex encoded stream of bytes to binary stream. My question is, could anyone suggest equivalent function but implemented with help of Boost libraries? I don't have any particular goal or requirement like better performance, simpler implementation, etc. I'm just curious what Boost components I can use to achieve this kind of tasks. I've scanned the archives but there seem to be not much on encoding/decoding hex string, just: http://lists.boost.org/boost-users/2007/11/32269.php // hex-to-bin function #include <sstream> #include <string> #include <vector> typedef std::vector<unsigned char> ewkb_t; // hexstr [in] - string with hex encoded bytes // bytes [out] - output in binary form void hex_to_bin(std::string const& hexstr, ewkb_t bytes) { bytes.clear(); for(std::string::size_type i = 0; i < hexstr.size() / 2; ++i) { std::istringstream iss(hexstr.substr(i * 2, 2)); unsigned int n; iss >> std::hex >> n; bytes.push_back(static_cast<unsigned char>(n)); } } Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org

1 0

Wrapping a library
by Alan Baljeu 21 Aug '08

21 Aug '08

I have this library (one class with 700 methods) I want to wrap all the calls, so I can log the calls and do other stuff. Using emacs macros and stuff, I can create a proxy class which has the same signature, but each method calls a C macro which calls the original class method, and does other stuff. For example: void logfunc(char *func) const { OutputDebugStringA(func); } void logarg(int a) // overload for each type of parameter { CString x; x.Format(L"%d", a); OutputDebugString(x); } ULONG GetID(const CKSEntity &pEntity) const { return proxy1(GetID, pEntity); } // repeat for all 700 methods. // Repeat these macros for every count of arg between 0 and 24. #define proxy10(Func, Arg1,Arg2,Arg3,Arg4,Arg5,Arg6,Arg7,Arg8,Arg9,Arg10) \ ( log10(#Func, Arg1,Arg2,Arg3,Arg4,Arg5,Arg6,Arg7,Arg8,Arg9,Arg10), \ thePart.Func( Arg1,Arg2,Arg3,Arg4,Arg5,Arg6,Arg7,Arg8,Arg9,Arg10)) #define log10(Func, Arg1,Arg2,Arg3,Arg4,Arg5,Arg6,Arg7,Arg8,Arg9,Arg10) \ log9(Func, Arg1,Arg2,Arg3,Arg4,Arg5,Arg6,Arg7,Arg8,Arg9), logarg(Arg10) #define log0(Func ) \ logfunc(Func) I'm thinking there must be a better way. I'm thinking BOOST_PP library could abstract the above macros into something concise. However, the library is intimidating. There are lots of features, and I'm not clear on how I should get started. __________________________________________________________________ Looking for the perfect gift? Give the gift of Flickr! http://www.flickr.com/gift/

3 2

[asio] wacky behavior in async_write/async_read
by n0vicd_c0der 21 Aug '08

21 Aug '08

Hi, I am trying to code a simple POP3S client. I am using the boost::asio sockets and SSL wrapper. Somehow, the output from the async_read_util is directed to the input of the async_write. Not sure if the issue is the memory or the order of calls. Any idea? The code snippet is below, but here is the output of my program first (I replaced the IP with x's) bash# ./main 209.85.199.111 995 foo.bar(a)gmail.com pass_in_plain_text Got RES (66) bytes +OK Gpop ready for requests from xxx.xxx.xxx.xxx c20pf1156691rvf.3 Sent REQ: APOP foo.bar(a)gmail.com pass_in_plain_text Got RES (36) bytes -ERR bad command c20pf1156691rvf.3 Sent REQ: STAT Read failed: End of file Exception: End of file and here is the code snippet #include <cstdio> #include <cstdlib> #include <unistd.h> #include <iostream> #include <boost/bind.hpp> #include <boost/asio.hpp> #include <boost/asio/ssl.hpp> #include <boost/asio/detail/throw_error.hpp> enum { max_length = 8192 }; class client { public: client(boost::asio::io_service& io_service, boost::asio::ssl::context& context, boost::asio::ip::tcp::endpoint& endpoint) : socket_(io_service, context) { //connect to the pop3 server socket_.lowest_layer().async_connect(endpoint, boost::bind(&client::handle_connect, this, boost::asio::placeholders::error)); } ~client() { socket_.lowest_layer().close(); } void handle_connect(const boost::system::error_code& error) { if (!error) { //do the SSL handshake socket_.async_handshake(boost::asio::ssl::stream_base::client, boost::bind(&client::handle_handshake, this, boost::asio::placeholders::error)); } else { std::cerr << "Connection failed: " << error << "\n"; boost::asio::detail::throw_error(error); } } void handle_handshake(const boost::system::error_code& error) { if (error) { std::cerr << "SSL Handshaking failed: " << error << "\n"; boost::asio::detail::throw_error(error); } } void write(std::string buffer) { str_req=buffer; boost::asio::async_write(socket_, boost::asio::buffer(str_req), boost::bind(&client::handle_write, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred)); } void handle_write(const boost::system::error_code& error, size_t bytes) { if (!error) { std::cout << "Sent REQ: " << str_req <<std::endl; try{ boost::asio::async_read_until(socket_, buff_, "\r\n", boost::bind(&client::handle_read_until, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred)); }catch(const boost::system::system_error &err){ std::cerr << " code: " << err.code().value() << " :" << err.code().message() << std::endl; } } else { std::cerr << "Write failed: " << error.message() << "\n"; boost::asio::detail::throw_error(error); } } void read() { try{ boost::asio::async_read_until(socket_, buff_, "\r\n", boost::bind(&client::handle_read_until, this, boost::asio::placeholders::error, boost::asio::placeholders::bytes_transferred)); }catch(const boost::system::system_error &err){ std::cerr << " code: " << err.code().value() << " :" << err.code().message() << std::endl; } } void handle_read_until(const boost::system::error_code& error, size_t bytes_transferred) { if (!error) { std::cout << "Got RES (" << bytes_transferred<<") bytes\n"; std::cout << &buff_; std::cout << "\n"; } else { std::cout << "Read failed: " << error.message() << "\n"; boost::asio::detail::throw_error(error); } } private: boost::asio::ssl::stream<boost::asio::ip::tcp::socket> socket_; boost::asio::streambuf buff_; //for receiving std::string str_req; //for sending }; int main(int argc, char* argv[]) { try { if (argc != 5) { std::cerr << "Usage: " << argv[0] << " <host> <port> <user> <pass>\n"; return 1; } boost::asio::io_service io_service; boost::asio::ssl::context ctx(io_service, boost::asio::ssl::context::sslv23); ctx.set_verify_mode(boost::asio::ssl::context::verify_none); boost::asio::ip::tcp::endpoint endpoint( boost::asio::ip::address::from_string(argv[1]), atoi(argv[2])); client c(io_service, ctx, endpoint); io_service.run(); //should execute till handshake completes //read the pop3 greeting io_service.reset(); io_service.post(boost::bind(&client::read,&c)); io_service.run(); //send the authentication command std::string str_cmd="APOP " + std::string(argv[3]) + " " + std::string(argv[4]) +"\r\n"; io_service.reset(); io_service.post(boost::bind(&client::write,&c,str_cmd)); io_service.run(); //get how many messages are in the inbox str_cmd="STAT\r\n"; io_service.reset(); io_service.post(boost::bind(&client::write,&c,str_cmd)); io_service.run(); //quit the pop3 session str_cmd="QUIT\r\n"; io_service.reset(); io_service.post(boost::bind(&client::write,&c,str_cmd)); io_service.run(); } catch (std::exception& e) { std::cerr << "Exception: " << e.what() << "\n"; } return 0; } -- View this message in context: http://www.nabble.com/-asio--wacky-behavior-in-async_write-async_read-tp190… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

[shared_ptr] Implicit cast shared_ptr<derived> to shared_ptr<base>
by Ruediger Berlich 21 Aug '08

21 Aug '08

Hi there, after making what I then thought was a bad mistake, I discovered that the code below actually works (gcc 4.3.1, OpenSUSE 11 64bit, Boost 1.36). I.e., there is an implicit cast from shared_ptr<derived> to shared_ptr<base> to make it fit the function argument, although both represent different types. Is this intended behaviour (i.e. portable) ? I did not find a corresponding description in the docu. Thanks and Best Regards, Ruediger /**************************************************************************/ #include <iostream> #include <boost/shared_ptr.hpp> class base{ public: virtual ~base(){} virtual void print() const { std::cout << "Hello world from base" << std::endl; } }; class derived :public base { public: void print() const { std::cout << "Hello world from derived" << std::endl; } }; void executeBasePrint(const boost::shared_ptr<base>& base_ptr){ base_ptr->print(); } main(){ boost::shared_ptr<derived> derived_ptr(new derived()); // Results in "Hello world from derived" executeBasePrint(derived_ptr); }

2 1

[filesystem] Bugs in path::append ?
by Sachin Garg 21 Aug '08

21 Aug '08

I have following 3 lines of code, is this usage correct? fs::wpath input_path=somepath; fs::wpath output_path=anotherpath; output_path.append(input_path.begin()+7,input_path.end()); The 3rd line gives following error (using Visual Studio 9.0 and boost 1.35.0) C:\Program Files\boost\boost_1_35_0\boost/filesystem/path.hpp(1058) : error C2664: 'boost::filesystem::detail::is_separator' : cannot convert parameter 1 from 'const std::basic_string<_Elem,_Traits,_Ax>' to 'wchar_t' Looking at the code, it does looks like a problem with path.hpp as is_separator expects a wchar_t parameter but append gives it a wstring (unless there some conversion that should have happened but isn't happening. I temporarily changed *first at line 1058 to first->at(0) and this fixed the error.). And there are other errors too like path.hpp(1068) : error C2677: binary '&&' : no global operator found which takes type 'const std::basic_string<_Elem,_Traits,_Ax>' (or there is no acceptable conversion) This too seems valid as path.hpp:1068 'seems' to assume that *first would be same as first->empty() Is this a known issue fixed in 1.36.0? Seems like some presumed conversions are not happening in path.hpp. Or am I making some super dumb mistake in my usage of append()? Sachin Garg

2 3

Sample code for boost::thread - Timed Lockable mutex ?
by Stéphane Vandenbogaerde 21 Aug '08

21 Aug '08

Hello, I want to use timed lockable mutex in boost::thread, but the usage is not so obvious for a beginner. Would there be anywhere a sample code with this kind og mutex ? I have a code which works with a boost::mutex in a consumer/producer queue (code below). The purpose is to have a maximum waiting delay for waitForNotEmpty(). Thanks. ------------------------------------------------------------------------- template <typename T> class MTBornedQueue { public: .... bool waitForNotEmpty() { if (!m_container.empty()) return true; boost::mutex::scoped_lock lock(m_accessMutex); m_notEmpty.wait(lock, boost::bind(&MTBornedQueue<value_type>::isNotEmpty, this)); return true; } ... private: bool isNotEmpty() const { return m_unread > 0; }; size_type m_unread; container_type m_container; boost::mutex m_accessMutex; boost::condition m_notEmpty; }

1 0

boost::graph issues.
by Raindog 21 Aug '08

21 Aug '08

Hello, This will be the 2nd time I've tried to use boost::graph in a project, the first time I ended up scrapping the effort and using a different graph implementation because of performance reasons, but I was also considering scrapping it for the reasons I have listed here. My newest project will be using much smaller graphs. First, I *want* to use bgl, the problem I have is that it is by far the most complex and difficult to use library I have ever seen. I equate it to something harder than COM using only C. The majority of my difficulties with BGL are related to the extensive use of template magic that does two things for me: makes code impossible to follow and secondly, it makes debugging even more difficult. For example, someone tell me from looking at this, what *values* will actually be passed to astar_search: astar_search (g, s, h, choose_param(get_param(params, graph_visitor), make_astar_visitor(null_visitor())), choose_param(get_param(params, vertex_predecessor), p_map), cost, distance, weight, index_map, color, choose_param(get_param(params, distance_compare_t()), std::less<C>()), choose_param(get_param(params, distance_combine_t()), closed_plus<C>()), choose_param(get_param(params, distance_inf_t()), std::numeric_limits<C>::max BOOST_PREVENT_MACRO_SUBSTITUTION ()), choose_param(get_param(params, distance_zero_t()), C())); } I would be hard pressed to believe that anyone can answer that in a short amount of time even assuming you have memorized all of the template parameters of g and s and even h. On with more issues. One of the biggest issues I have is that the the property_map concept is exceptionally vague and exceedingly difficult to work with. For example, I have created a graph, it's concise typedef: typedef boost::adjacency_list < boost::vecS, boost::vecS, boost::directedS, boost::property<boost::vertex_name_t, Ogre::Vector3>, boost::property<boost::edge_weight3_t, edge_cost> > GraphT; Because of the use of free functions, it's hard to quickly distinguish "get" calls on a property_map from "get" calls on a graph. Additionally, the property_map installation and retrieval to and from a BGL graph is also hard to follow. One might see the following code in the examples directory of bgl: typename property_map<Graph, edge_mycapacity_t>::const_type capacity = get(edge_mycapacity, G); typename property_map<Graph, edge_myflow_t>::const_type flow = get(edge_myflow, G); Where "edge_mycapacity" is an enum; and how exactly enums of the same value can be used to look up different property_maps from the graph is a gaping hole left in the documentation and I still have no idea how it works. Another example is the following: get(vertex_color_t(), g, vertex(2,g)) Here we seem to be left to our imaginations as to how this works. "get" in this case is obviously operating on a graph in this call, but why are we no longer using an enum and instead constructing a temporary "vertex_color_t". Looking at the definition of "get", we see the following: template<typename OutEdgeListS, typename VertexListS, typename DirectedS, typename VertexProperty, typename EdgeProperty, typename GraphProperty, typename EdgeListS, typename T, typename Bundle, typename Key> inline T get(T Bundle::* p, adjacency_list<OutEdgeListS, VertexListS, DirectedS, VertexProperty, EdgeProperty, GraphProperty, EdgeListS> const & g, const Key& key) { return get(get(p, g), key); } Hmm, looks like we'll need to look further. I personally am not knowledgeable enough in template magic to make sense of the majority of the BGL source, for me there are way too many things to keep track of in order to be able to completely follow the code. I can't follow very well 100 character variable declarations. So, for some real questions and to end the rant. 1. What language mechanism is allowing property_maps to be stored and retrieved from an instance of a graph? How does "BOOST_DEF_PROPERTY(edge, index)" work exactly? 2. What *is* the required interface for a property_map? get,put, operator[] of course, but the BGL properties appear to function nothing like the example given in the documentation. 3. Why is using get(p,g) better than using myPropMap.get_from(myGraph) and myGraph.get_property_map(prop_map_type) and myGraph.get_edge(edge_descriptor). What benefit does the chosen method provide? 4. The edge property example contains the following: typedef property<edge_mycapacity_t, int> Cap; typedef property<edge_myflow_t, int, Cap> Flow; typedef adjacency_list<vecS, vecS, bidirectionalS, no_property, Flow> Graph; But accesses the Cap property_map as if it is directly associated with the graph, not a nested property of Flow: typename property_map<Graph, edge_mycapacity_t>::const_type capacity = get(edge_mycapacity, G); Can someone explain how that works in better detail? Thanks, Josh PS. This is the 4rd time i've attempted to send this message as I've not seen it yet appear on the list. The 1st 2 attempts were rejected as I had not yet completed subscription process.

5 9

AIX 5.3 xlC error on ptr_list.hpp
by Gary Fu 21 Aug '08

21 Aug '08

Hi, The simple test program a.cpp below has no problem to compile on linux g++, but failed on AIX xlC compiler. Does any one know if the problem is related to the xlC compiler or the ptr_list.hpp file ? Thanks, Gary 4:43pm 19 Ggfu@sds:~/tmp> xlC_r -I/home_dir/Ggfu/BOOST/boost_1_36_0 -c a.cpp "/home_dir/Ggfu/BOOST/boost_1_36_0/boost/ptr_container/ptr_list.hpp", line 44.9: 1540-0400 (S) "boost::ptr_list::release" has a conflicting declaration. "/home_dir/Ggfu/BOOST/boost_1_36_0/boost/ptr_container/ptr_list.hpp", line 44.9: 1540-0425 (I) "release" is defined on line 44 of "/home_dir/Ggfu/BOOST/boost_1_36_0/boost/ptr_container/ptr_list.hpp". ----------------- a.cpp #include <boost/ptr_container/ptr_list.hpp> boost::ptr_list< std::string > y; int main() { return 0; }

3 2

How to find a sorting permutation
by Zeljko Vrba 21 Aug '08

21 Aug '08

I have a vector of element V that needs to be sorted. I want to also produce a permutation vector P which reflects the element's position in the original vector. P is initialized with identity permutation (P[i] == i). 1st possibility: code my own sorting algorithm, and swap elements of P along with swapping elements of V. This I want to avoid, if at all possible. 2nd possibility: copy elements of V into new vector V' holding pair<T, size_t> and sort that. (pair holds the original element as well as its starting position in the vector, i.e. .second is again initialized to identity permutation.) Drawback: extra space and time spent on copying to V'. 3rd possibility: sort elements of P with less_than doing lookups into V to do its comparison. Drawback: while P holds the correct result, V is left unsorted; I have to do another pass through V and swap elements according to the permutation recorded in P, which is not trivial to do in-place. 4th possibility: I'm not sure whether it'd work: have a separate compilation unit that will provide function sort_vector_of_T, which will take references to V, T and also provide a specialization of std::swap for T which will then also swap elements of P). (Separate compilation unit is needed because I don't want to *globally* redefine std::swap for T) Is there some non-obvious, efficient and elegant way of achieving what I want?

3 3

Implementation question about boost::circular_buffer and allocator
by Peter Barker 21 Aug '08

21 Aug '08

Hello, I've been going through the Boost libraries and encountered boost::circular_buffer. My question is not with the operation of Boost.Circular Buffer as-such, but about how the instantiation of it works without specifying the allocator. The introductory example in the documentation(http://www.boost.org/doc/libs/1_35_0/libs/circular_buffer/doc… contains the following line: // Create a circular buffer with a capacity for 3 integers. boost::circular_buffer<int> cb(3); But looking at the header file base.hpp in the circular_buffer directory, we have the following: template <class T, class Alloc> class circular_buffer { ...... }; The Alloc template parameter is not a default one, so I can't understand how we can instantiate cb in the example? Apologies if this is not strictly boost-user related. Regards, Jim

2 2

multi_array, resize and default constructable
by gast128 21 Aug '08

21 Aug '08

Dear all, from the documentation I read that if a multi_array is reshaped new elements are default constructed: "Resizing an Array The boost::multi_array class provides an element-preserving resize operation. ...Any new elements in the array are default constructed..." It is thus not possible providing a default fill element?

2 2

weak_ptr for intrusive_ptr
by Alan M. Carroll 21 Aug '08

21 Aug '08

I greatly prefer intrusive_ptr to shared_ptr, but one drawback of that is there is no equivalent of weak_ptr. I have started work on a class to do that and my biggest question, honestly, was what to name it. The basic logic turned out to be fairly simple. The target class declares a member that is an instance of the weak pointer. Other classes that want a weak pointer to the target instance also declare a member weak pointer, which must be initialized with a copy of the target instance member. The member provides a method to generate an intrusive_ptr to the target, which is either valid or nil depending on whether the target has been destructed. When the target and all weak pointers are destructed, everything gets cleaned up. The weak pointers don't affect the target reference count. No changes to Boost.intrusive_ptr are required. The one improvement I see would be to merge it into the reference counter utility class discussed earlier (optionally, of course - perhaps via a subclass of the reference counter). Yeah, I think I'll do that. My questions: 1) What would be a good name for such a class? I am using "tell_tale" for now, in analogy to tell tale lights on devices that go out when the device fails. Obviously "weak_ptr" is out. "weak_intrusive_ptr" seems a bit verbose, but may that's OK. 2) Is there any interest in Boostifying the class, or am I the only one who would use it? It's a matter of how much I indulge my idiosyncrasies.

4 8

Re: [Boost-users] Multi-index container (Modifying non-unique key in the container)
by JOAQUIN M. LOPEZ MUÑOZ 20 Aug '08

20 Aug '08

________________________________ De: boost-users-bounces(a)lists.boost.org [boost-users-bounces(a)lists.boost.org] En nombre de Ramesh [rramesh1(a)gmail.com] Enviado el: martes, 19 de agosto de 2008 20:16 Para: boost-users(a)lists.boost.org Asunto: Re: [Boost-users] Multi-index container (Modifying non-unique key in the container) On Wed, Aug 13, 2008 at 10:20 AM, JOAQUIN M. LOPEZ MUÑOZ <joaquin(a)tid.es<mailto:joaquin@tid.es>> wrote: ________________________________ De: boost-users-bounces(a)lists.boost.org<mailto:boost-users-bounces@lists.boost.org> [boost-users-bounces(a)lists.boost.org<mailto:boost-users-bounces@lists.boost.org>] En nombre de Ramesh [rramesh1(a)gmail.com<mailto:rramesh1@gmail.com>] Enviado el: martes, 12 de agosto de 2008 20:40 [...] Hmm... I'm not referring to your myclass wrapper, but rather to the fact that you're storing DataMaps objects into the container rather than A*s directly. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo > Sorry if I got your question wrong, I am storing only pointers to objects > inside the map and each instance of myclass would hold one such > container to hold pointers to objects of class A along with keys. Yes, but DataMap seems to *replicate* the keys that are already present in A. Cannot you just have the following? typedef multi_index_container< A*, indexed_by< ordered_unique< tag <Key1>, BOOST_MULTI_INDEX_MEMBER(A,std::string,Key1)>, ordered_non_unique< tag<Key2>,BOOST_MULTI_INDEX_MEMBER(A,long,Key2)>, ordered_unique< tag<Key3>,BOOST_MULTI_INDEX_MEMBER(A,long,Key3)> > > tDMap; Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

[Exception] 1.36.0: mutable Patch
by Peter Klotz 20 Aug '08

20 Aug '08

gcc 3.2 complains heavily about keyword mutable not being at the beginning of declarations. The attached patch fixes that. The C++ standard does not enforce mutable placement however since Boost is already full of compiler specific workarounds I would be happy if this patch could be applied. Regards, Peter. Peter Klotz Software Engineer Phone: +43 (0) 50 8648-4100 Fax: +43 (0) 50 8648-4111 E-Mail: peter.klotz(a)ith-icoserve.com ---------------------------------------------------------------- ITH icoserve technology for healthcare GmbH A-6020 Innsbruck, Innrain 98 www.ith-icoserve.com Rechtsform: Gesellschaft mit beschränkter Haftung Firmensitz: 6020 Innsbruck, Innrain 98 Firmenbuchnummer: FN 174117f Firmenbuchgericht: Innsbruck DVR: 0983039

2 3