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

unit test framework XML output - CDATA sections?
by Maróy Ákos 23 Nov '09

23 Nov '09

Hi, I have a suggestion for the XML output generated by the boost unit test framework: currently, if tests fail, the message detailing the failure of the test is put in verbatim as a test element into the XML output. This can be problematic, if the output contains XML-like structures, like text that looks like an XML element, or invalid characters. thus, I'd suggest enclosing such output in a CDATA section, that would ensure that the generated XML file is indeed always well formed XML. see more here: http://www.w3.org/TR/REC-xml/#sec-cdata-sect - but basically this would mean printing the failures like this: <![CDATA[ reason of the failure... ]]> Akos

2 2

[test] utf now supports log/report sinks
by Gennadiy Rozental 23 Nov '09

23 Nov '09

Hi, Based on users' requests I've added 2 new runtime parameters: '--log_sink=' or '-k ' or BOOST_TEST_LOG_SINK: Allows to set the log sink. Valid values are 'stdout' (default), 'stderr', and anything else which is treated as file name. '--report_sink=' or '-e ' or BOOST_TEST_REPORT_SINK: Allows to set the report sink. Valid values are 'stderr' (default), 'stdout', and anything else which is treated as file name. Gennadiy

1 0

[Spirit.Qi] How do I get Rule Name/Id through Semantic Action / Phoenix variable
by HT4N - 23 Nov '09

23 Nov '09

I want to create a generic Parse Tree from a simple expression grammar (see below code snippet). Using the Mini_XML as the reference, I was able to create a Parse Tree close to what I want. If I run my simple program and parse an expressions "[foo=1 bar=2]", for example, it will printout the following Parse Tree: [foo=1 bar=2] ------------------------- Parsing succeeded <nonterminal> { <nonterminal> { <value_str>foo</value_str> <value_int>1</value_int> } <nonterminal> { <value_str>bar</value_str> <value_int>2</value_int> } } The next thing I want to achieve is to have the Parse Tree looks very similar to the one produced in Spirit.Classic. In Spirit.Classic for each node it has RuleId for which we associate it with a RuleName so that when we are iterating the node of the Parse Tree we can printout the RuleName. In other words, instead I want the above Parse Tree to look like the following : <Option> <Expression> <ParameterName>foo</ParameterName> <ParameterValue>1</ParameterValue> </Expression> <Expression> <ParameterName>bar</ParameterName> <ParameterValue>2</ParameterValue> </Expression> </Option> In other words, what I need to know is how I can get the Rule Name/Id through semantic actions/phoenix variable. Any pointer is appreciated. Thanks again. HT Code Snippet: namespace client { struct ptnode; typedef boost::variant< nil, unsigned int, string, boost::recursive_wrapper<ptnode> > ptnode_var; /////////////////////////////////////////////////////////////////////////// // Our AST /////////////////////////////////////////////////////////////////////////// struct ptnode { ptnode() : data(nil()) { } template<typename T> ptnode(const T& data) : data(data) { } ptnode_var data; vector<ptnode_var> children; }; } // We need to tell fusion about our mini_xml struct // to make it a first-class fusion citizen //[tutorial_xml1_adapt_structures BOOST_FUSION_ADAPT_STRUCT( client::ptnode, (client::ptnode_var, data) (std::vector<client::ptnode_var>, children) ) //] namespace client { template <typename Iterator> struct ExpressionGrammar : qi::grammar<Iterator, ptnode(), ascii::space_type> { ExpressionGrammar() : ExpressionGrammar::base_type(Option) { using qi::_val; using qi::_1; using qi::char_; using qi::alnum; using qi::alpha; using qi::lexeme; using qi::raw; using qi::uint_; using phoenix::at_c; using phoenix::push_back; Option = char_('[') >> *Expression [push_back(at_c<1>(_val), _1)] >> char_(']'); Expression = ParameterName [push_back(at_c<1>(_val), _1)] >> (char_('=')>> ParameterValue [push_back(at_c<1>(_val), _1)]); ParameterName = lexeme[+(alnum) [_val += _1]]; ParameterValue = uint_ [_val = _1]; Option.name("Option"); Expression.name("Expression"); ParameterName.name("ParameterName"); ParameterValue.name("ParameterValue"); } qi::rule<Iterator, ptnode(), ascii::space_type> Option; qi::rule<Iterator, ptnode(), ascii::space_type> Expression; qi::rule<Iterator, string(), ascii::space_type> ParameterName; qi::rule<Iterator, int(), ascii::space_type> ParameterValue; }; /////////////////////////////////////////////////////////////////////////// // Print out the mini xml tree /////////////////////////////////////////////////////////////////////////// int const tabsize = 4; void tab(int indent) { for (int i = 0; i < indent; ++i) { printf_s(" "); } } struct ptnode_printer { ptnode_printer(int indent = 0) : indent(indent) { } void operator()(ptnode const& node) const; int indent; }; struct ptnode_var_printer : boost::static_visitor<> { ptnode_var_printer(int indent = 0) : indent(indent) { } void operator()(const ptnode& var) const { ptnode_printer(indent+tabsize)(var); } void operator()(const nil&) const { // Print node type ... tab(indent); printf_s("<nonterminal>\n"); } void operator()(const string& val) const { tab(indent+tabsize); printf_s("<value_str>%s</value_str>\n", val.c_str()); } void operator()(const int& val) const { tab(indent+tabsize); printf_s("<value_int>%d</value_int>\n", val); } int indent; }; void ptnode_printer::operator()(const ptnode& node) const { boost::apply_visitor(ptnode_var_printer(indent), node.data); tab(indent); printf_s("{\n"); BOOST_FOREACH(const ptnode_var& child, node.children) { boost::apply_visitor(ptnode_var_printer(indent), child); } tab(indent); printf_s("}\n"); } } _________________________________________________________________ Windows 7: I wanted simpler, now it's simpler. I'm a rock star. http://www.microsoft.com/Windows/windows-7/default.aspx?h=myidea?ocid=PID24…

2 3

[test] Malloc exceptions when using test framework on 64bit OS X 10.6.1
by Vrai Stacey 23 Nov '09

23 Nov '09

Hi, I've been making my first forays in to using Boost on the new OS X release and have encountered some strange behaviour on 64bit machines. Testing with Boost 1.37.0 and 1.40.0 I've found that everything compiles as expected, but attempting to run any unit tests results in an error of the form: *** No errors detected test_exceptions(92517) malloc: *** error for object 0x3000100200460: pointer being freed was not allocated *** set a breakpoint in malloc_error_break to debug Running the test through gdb gives the following stack trace: Program received signal SIGABRT, Aborted. 0x00007fff83f98ff6 in __kill () (gdb) bt #0 0x00007fff83f98ff6 in __kill () #1 0x00007fff8403a072 in abort () #2 0x00007fff83f51095 in free () #3 0x000000010003668a in boost::unit_test::framework_impl::clear (this=0x1000698a0) at framework.ipp:133 #4 0x0000000100036788 in boost::unit_test::framework_impl::~framework_impl (this=0x1000698a0) at framework.ipp:122 #5 0x00000001000386f2 in __tcf_1 () at framework.ipp:222 #6 0x00007fff83f5d274 in __cxa_finalize () #7 0x00007fff83f5d18c in exit () #8 0x0000000100006e93 in start () at unit_test_parameters.ipp:76 This is using the vanilla GCC build that comes with the operating system (see below), default compile flags and strictly single threaded test applications. Using built-in specs. Target: i686-apple-darwin10 Configured with: /var/tmp/gcc/gcc-5646~6/src/configure --disable-checking --enable-werror --prefix=/usr --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with-slibdir=/usr/lib --build=i686-apple-darwin10 --with-gxx-include-dir=/include/c++/4.2.1 --program-prefix=i686-apple-darwin10- --host=x86_64-apple-darwin10 --target=i686-apple-darwin10 Thread model: posix gcc version 4.2.1 (Apple Inc. build 5646) Some Googling suggests it might be problem with a missing virtual destructor but irritatingly, none of the threads I've found reach any kind of conclusion. I was hoping the collective genius of the list would be able to inform me whether: 1. This is a known bug; I'm happy to put my debugging hat on and dive in, but not if it's already being investigated by people more qualified to do so. 2. There are any work arounds for the problem. 3. There's no problem at all; I'm being a colossal idiot and have overlooked something obvious. Thanks, vrai.

9 28

Re: [Boost-users] obtaining the day of the week as a string
by Igor R 22 Nov '09

22 Nov '09

> Is there a way of obtaining the day ofthe week as a string in the local > language using boost.datetime? > > If yes what is the best way of doing this? > Look at this example: http://www.boost.org/doc/libs/1_41_0/doc/html/date_time/examples.html#date_…

1 0

obtaining the day of the week as a string
by Sean Farrow 22 Nov '09

22 Nov '09

Hi: Is there a way of obtaining the day ofthe week as a string in the local language using boost.datetime? If yes what is the best way of doing this? Regards Sean.

1 0

boost::interprocess container of pointers
by Mateusz Ziobro 22 Nov '09

22 Nov '09

Hi, I have a problem with container of pointers in shared memory using boost::interprocess, object created by construct method is mapped to local memory address, so I can not just put it into shared memory container because the address would be invalid in other process, probably I have to remap or use some kind of smart pointer and than put into container. My question is Does someone know how to do that? Please see following piece of code. #include <iostream> #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <boost/interprocess/containers/set.hpp> class TestData { int value_; public: TestData(int value): value_(value) { } int get_value() const { return value_; } }; int main(int argc) { using namespace boost::interprocess; managed_shared_memory mgr(open_or_create, "shm_test", 1024*1024); TestData* t1 = mgr.construct<TestData>("test_data")(99); std::cout << t1->get_value() << std::endl; //TestData* is address in local memory //we can not keep those addresses in shared memory set, would be invalid in other process //typedef allocator<const TestData*, managed_shared_memory::segment_manager> allocator; //typedef set<const TestData*, std::less<TestData*>, allocator> TestDataSet; //TestDataSet* s1 = mgr.construct<TestDataSet>("test_data_set")(std::less<TestData*>(), // mgr.get_segment_manager()); }

2 1

[Asio] serial port example doesn't work on Mac: Operation not supported, bad file descriptor
by philipp_muenzel＠web.de 22 Nov '09

22 Nov '09

Hello, I'm trying to communicate with the serial port using the example code "minicom" that was posted here: http://lists.boost.org/boost-users/att-41140/minicom.cpp On Linux, the example works flawlessly and lets me communicate with both "native" serial port hardware and USB-to-serial converted hardware with FTDI chips, which can be adressed via /dev/ttyUSB. On a Power Mac with OS Tiger, using boost 1.40 and the gcc 4.0 compiler, everytime I start the example I get the following errors: $ ./minicom.app/Contents/MacOS/minicom 57600 /dev/cu.usbserial-4B11 Error: Operation not supported Press Enter to exit Error: Bad file descriptor Press Enter to exit However, I can talk to the serial-via-usb hardware using some native socket functions, so the file descriptor itself can't be the problem. How can I get the minicom example to work on Mac also? Regards, Philipp ___________________________________________________________ Preisknaller: WEB.DE DSL Flatrate für nur 16,99 Euro/mtl.! http://produkte.web.de/go/02/

1 0

[Review] GGL Review ends today, November 22nd
by Hartmut Kaiser 22 Nov '09

22 Nov '09

Hi all, The review of the Generic Geometry Library officially ends today, November 22nd. Please send in your reviews if you have not done so yet. If you need more time for your review, please get in contact directly with me to work out some individual schedule. Regards Hartmut GGL Review Manager ------------------- Meet me at BoostCon http://boostcon.com

1 0

[Spirit.Qi] Compile Error for MiniXml/Employee example
by HT4N - 22 Nov '09

22 Nov '09

As I am considering to use Spirit.Qi for my project, I am ramping up myself through the examples given. I was able to compile the calcX examples. However, I have issues compiling Mini_XML* and Employee.cpp examples. I got the following errors: errors in directory c:\src\scratch\spirit.qi c:\src\boost141\boost\proto\extends.hpp(537) : error C2220: warning treated as error - no 'object' file generated warnings in directory c:\src\scratch\spirit.qi c:\src\boost141\boost\proto\literal.hpp(40) : warning see reference to class template instantiation 'boost::proto ::exprns_::extends<Expr,Derived,Domain>' being compiled c:\src\scratch\spirit.qi\spiritqimain.cpp(81) : warning see reference to class template instantiation 'boost::proto::utilit y::literal<T,Domain>' being compiled c:\src\scratch\spirit.qi\spiritqimain.cpp(70) : warning while compiling class template member function 'client::employee_pa rser<Iterator>::employee_parser(void)' c:\src\scratch\spirit.qi\spiritqimain.cpp(116) : warning see reference to class template instantiation 'client::employee_pa rser<Iterator>' being compiled c:\src\boost141\boost\proto\extends.hpp(537) : warning C4512: 'boost::proto::exprns_::extends<Expr,Derived,Domain >' : assignment operator could not be generated c:\src\boost141\boost\proto\literal.hpp(76) : warning C4512: 'boost::proto::utility::literal<T,Domain>' : assignm ent operator could not be generated Any help appreciated! Thanks HT _________________________________________________________________ Windows 7: It works the way you want. Learn more. http://www.microsoft.com/Windows/windows-7/default.aspx?ocid=PID24727::T:WL…

2 1

using boost::thread with Sun Studio C++ gives error
by Maróy Ákos 21 Nov '09

21 Nov '09

Hi, I'm trying to compile a code which uses the boost thread library, but I get the following error: CC -library=stlport4 -features=tmplife -features=tmplrefstatic -mt -c -o t t.cpp "/opt/boost/include/boost/thread/pthread/mutex.hpp", line 142: Error: The function "pthread_mutex_timedlock" must have a prototype. "/opt/boost/include/boost/thread/pthread/mutex.hpp", line 142: Error: The function "pthread_mutex_timedlock" must have a prototype. "/opt/boost/include/boost/thread/detail/thread.hpp", line 344: Error: boost::thread::thread(boost::thread&) is not accessible from boost::move(boost::detail::thread_move_t<boost::thread>). "/opt/boost/include/boost/thread/pthread/recursive_mutex.hpp", line 179: Error: The function "pthread_mutex_timedlock" must have a prototype. "/opt/boost/include/boost/thread/pthread/recursive_mutex.hpp", line 179: Error: The function "pthread_mutex_timedlock" must have a prototype. 5 Error(s) detected. make: *** [t] Error 5 the sample source code in question here just a single include: #include <boost/thread.hpp> This is with boost 1.40.0. I wonder what I'm doing wrong.. Akos

1 0

smart_ptr subclass as template function parameter issues with Sun Studio
by Maróy Ákos 21 Nov '09

21 Nov '09

Hi, I have the following problem, when using the Sun Studio C++ compiler and boost::smart_ptr. Consider the following code: class A {}; class B : public A {}; template<typename T> void fn(shared_ptr<A> a, T t) { } void foo() { shared_ptr<A> a; shared_ptr<B> b; fn(a, 1); fn(b, 1); } this compiles fine with a number of compilers, including gcc, icc and MS Visual Studio. But when I try to compile it using the Sun Studio compiler, I get: Error: Could not find a match for fn<T>(boost::shared_ptr<B>, int) needed in foo(). it only works if I explicitly cast the derived class' shared_ptr to that of the base class, that is, instead of: fn(b, 1); I need to use: fn((shared_ptr<A>) b, 1); which seems rather inconvenient. is the fn(b, 1); style really bad? is there a way to make the Sun compiler accept this form? Akos

1 0

BOOST-UBLAS: how to implement A.B.A^t operation.
by Arman Khalatyan 21 Nov '09

21 Nov '09

Hello, I would like to calculate for B=NxN matrix and A=N vector following operation: RES= A (dot)* *B (dot) A^T using namespace boost::numeric::ublas; matrix<double> B (2, 2); matrix<double> A(1, 2); // this works when I do step by step matrix<double> res=prod(A,B); res=prod(res,trans(B)); // but this does not compile res=prod(prod(A,B),trans(A)) // it gives error error C2027: use of undefined type 'boost::STATIC_ASSERTION_FAILURE<x>' So what is the difference with previous? Thanks Arman.

1 0

[type_traits] typedef problem on Microware/Radisys OS9
by Doug Crawford 21 Nov '09

21 Nov '09

I am trying to port parts of boost to the Microware/Radisys OS9 platform. This is really an old embedded systems C++ compiler that has not been updated since about 1997. Anyways I can get shared_ptr to work and convert between convertible types. But when I try to compile anything out of type_traits library the compiler can't handle anything like: typedef typename se_t::template type<T> helper; (line 156 in is_enum.hpp) My compiler error is "expected an identifier" at ::template. I am guessing that I am out of luck, but is there any way to rewrite or simplify these kind of constructs?

4 3

Linker not pulling in boost_system properly
by Jason Sewall 20 Nov '09

20 Nov '09

I'm having trouble with linking to an installation of Boost. I just downloaded boost_1_41_0.tar.bz2, untarred it, and did: $ ./bootstrap.sh --prefix=/home/sewall/src/boost-1.41.0 $ ./bjam --prefix=/home/sewall/src/boost-1.41.0 install Now I've got a project (using autotools+libtool) of my own making that uses a number of Boost libraries. After pointing the configure script to the right spot for the Boost libraries, I wind up with errors like the following: /bin/sh ../libtool --tag=CXX --mode=link g++ -I../sim/include -I/home/sewall/src/boost-1.41.0/include -I/usr/include/libxml2 -fopenmp -O0 -fno-inline -ggdb -Wall -Wextra -Wno-unused-parameter -Wno-enum-compare -o xml-test xml-test.o ../sim/src/libhwm_net.la -lrt libtool: link: g++ -I../sim/include -I/home/sewall/src/boost-1.41.0/include -I/usr/include/libxml2 -fopenmp -O0 -fno-inline -ggdb -Wall -Wextra -Wno-unused-parameter -Wno-enum-compare -o xml-test xml-test.o ../sim/src/.libs/libhwm_net.a -L/home/sewall/src/boost-1.41.0/lib -lxml2 -lz -lboost_filesystem -lboost_regex -lrt /usr/bin/ld: warning: libboost_system.so.1.41.0, needed by /home/sewall/src/boost-1.41.0/lib/libboost_filesystem.so, not found (try using -rpath or -rpath-link) /home/sewall/src/boost-1.41.0/lib/libboost_filesystem.so: undefined reference to `boost::system::get_system_category()' /home/sewall/src/boost-1.41.0/lib/libboost_filesystem.so: undefined reference to `boost::system::get_generic_category()' I don't claim to understand libtool terribly well, but I think that the xml-test target is pulling in the dependencies of libhwm_net.la, which in turn pulls in boost_filesystem and boost_regex. It seems that it boost_filesystem is then asking for libboost_system. Everything is acting as it should up to this point, but then the linker seems to be at a loss to where libboost_system can be found. If I manually add -lboost_system to the above command line, the target builds fine. Of course I can just have my configure script general -lboost_filesystem and -lboost_system whenever boost_filesystem is needed, but I suspect there's a better way to do it. Is there something I should have added to my Boost build to have it know to look for boost_system in /home/sewall/src/boost-1.41.0? Thanks, Jason P.S. I have added /home/sewall/src/boost-1.41.0 to my LD_LIBRARY_PATH, if that helps.

2 2

[Spirit] poor performance
by HT4N - 20 Nov '09

20 Nov '09

I just shocked to find out that my parser with Spirit-classic is 4x slower than my original ad-hoc parser. I do generate parsetree with pt_parse(). So does my adhoc parser but my adhoc one generates its own custom parse tree. My question are: 1. Does spirit classic with generation of parse tree is known to be slow? 2. Does anyone have ever done perf study and compare Spirit.Classic with Spirit.Qi (assuming to generate custom parse tree in Spirit.Qi)? 3. If the bottleneck is in the Spirit.Classic parse tree generation, has anyone ever done a custom parse tree generation in Spirit.Classic? For now it seems I am hitting a roadblock to use Spirit as performance is my main issue. Your insight is highly appreciated. Thanks HT _________________________________________________________________ Bing brings you maps, menus, and reviews organized in one place. http://www.bing.com/search?q=restaurants&form=MFESRP&publ=WLHMTAG&crea=TEXT…

3 2

Boost.Fusion 1.41 issue ?
by Surya Kiran Gullapalli 20 Nov '09

20 Nov '09

Hello, he following test program fails to compile with Boost-1.41 on msvc8.0 SP1. It was working fine with Boost-1.40. Am I missing something here ? Thanks, Surya #include <iostream> #include <boost/preprocessor.hpp> #include <boost/fusion/container.hpp> using namespace std ; using namespace boost ; #define MAX_ARGS 5 template <int id, BOOST_PP_ENUM_PARAMS(MAX_ARGS, typename T)> void func (const boost::fusion::vector<BOOST_PP_ENUM_PARAMS(MAX_ARGS, T)> p_args) { cout << "Multiple Arguments: id " << id << endl ; } template<int id> void func (void) { cout << "No arguments: id " << id << endl ; func <id>(boost::fusion::make_vector()) ; } int main (void) { func<10> (boost::fusion::make_vector(1, 2)) ; func<6> () ; return 0 ; }

2 3

multi-index container random access index ordered
by Vladimir Voronkov 20 Nov '09

20 Nov '09

Hi, I defined a container like this: typedef boost::multi_index_container< std::pair<int, int>, boost::multi_index::indexed_by< boost::multi_index::random_access<>, boost::multi_index::ordered_unique< boost::multi_index::member<std::pair<int, int>, int, &std::pair<int, int>::first> > > > IdContainer; typedef IdContainer::nth_index<0>::type IdsArr; typedef IdContainer::nth_index<1>::type IdsSet; and then I insert a data into it: a_set.insert(std::make_pair(1, 0)); a_set.insert(std::make_pair(3, 0)); a_set.insert(std::make_pair(2, 0)); a_set.insert(std::make_pair(5, 0)); a_set.insert(std::make_pair(4, 0)); in this case elements will be “pushed back” in terms of random access index, i. e. for (size_t i = 0; i < a_arr.size(); i++) { std::cout << a_arr[i].first << std::endl; } will return 1 3 2 5 4 I would like to have them inserted ordered way, and as long as we have ordered index I suppose it’s possible. How can I do that? I suspect I could do like this: a_arr.insert(a_set.find(1), std::make_pair(1, 0)); but it errors about incompatible types of iterators for what a_set.find returns and what a_arr.insert accepts. Or… may be there are some kind of definition of random access index with respect to another index which is ordered??? I don’t know. Your help is appreciated.

2 1

Re: [Boost-users] [Filesystem] API enhancement request
by Alan M. Carroll 20 Nov '09

20 Nov '09

I would like to have an extension added to Boost.Filesystem, which I have been hand patching along the way. For version 1.41.0, in boost/filesystem/path.hpp, at line 274, add public: typedef boost::BOOST_FILESYSTEM_NAMESPACE::basic_path<String, Traits> path_type; What this does is provides access to the path type for a path iterator from just the iterator. I find it very handy when writing generic path handling code. Otherwise you have to hard code for a particular type of path, or pass the type along even though the iterator depends on that type.

1 0

libboost_filesystem-vc90-xxx-1_39.lib requirement in boost 41?
by Archie14 20 Nov '09

20 Nov '09

My project (windows, vs2008, debug) was linking successfully under boost-39. When I switched to boost 41 version the linker started to complain about missing lib (fatal error LNK1104: cannot open file 'libboost_filesystem-vc90- mt-gd-1_39.lib'). My linker properties does not specify the library explicitly, just the path to boost libraries. The path had been changed to boost 41 (I have $BOOST system var that both compiler and linker properties refer to) and there are corresponding libraries under the same relative path in boost 41 (libboost_filesystem-vc90-mt-gd-1_41.lib). Specifying new library explicitly in the linker properties did not help. I renamed old "39" directory and both compiler and linker are referring to boost via system variable. I double checked that this system variable points to the new "41" directory. Compiling completes sucessfully, but linker fails. How can I correct the problem?

3 2

different behavior for boost::filesystem with VC9 in debug and release
by frederic cervenansky 20 Nov '09

20 Nov '09

Hello, Sorry to ask this question but I investigate and I didn't find any previous message. I try to use boost::filename::is_directory( path) in my code developed on visual studio 2008. In debug mode, I have an error on this function and specificaly at file_status result( detail::status_api( ph.external_file_string(), ec ) ); in operations.hpp. (Detail status crashes, it seems). In release mode, no problem everything is fine, my path is valid and my program is running.... And yes, my path is valid. I have boost 1.40, the compilation option for boost and my program are the same. The linkage is valid (I checked that my compiled libraries are used). So anyone has a clue for this bug? Because I'd like to understand and stay in debug mode to track bugs. Thanks by advance

1 0

Newbie looking for example of try...catch for Date_Time
by Ramon F Herrera 20 Nov '09

20 Nov '09

I have only cursory experience with exemption catching. I already looked in the documentation but didn't find an example. My program keeps on crashing when a bad month or date is used. TIA, -RFH

1 0

posix_time_zone::calc_zone probably buggy
by Sorin Manolache 20 Nov '09

20 Nov '09

Hello, The Posix specification says that the offset in the TZ variable gives the time to add to the local time in order to get UTC. Thus, western hemisphere zones have positive offsets and eastern hemisphere zones have negative offsets. time_zone_base::base_utc_offset gives the offset of the timezone from UTC. Thus, western hemisphere zones have negative base_utc_offset_ and eastern hemisphere zones have positive base_utc_offset_. Posix specifies that the offset specified in TZ must be added to the localtime in order to obtain UTC. base_utc_offset_ must be added to UTC in order to obtain local times. Thus, base_utc_offset_ must be minus the TZ offset. So I think a minus should be added in front of date_time::str_from_delimited_time_duration<time_duration_type,char_type> in the calc_zone method of date_time/local_time/posix_time_zone.hpp. Regards, Sorin

1 0

[function] Using boost::function when rtti is not available
by David Genest 20 Nov '09

20 Nov '09

Hi, I am in a gaming build environment that has traditionally turned of rtti because of space and portability issues. But, I would like to use the very useful boost::function. On MSVC boost::function works, but fails on g++ for the ps3. I have read on the lists that it is possible to implement a rtti-less version, and that this is even wanted in the future. See: http://article.gmane.org/gmane.comp.lib.boost.devel/192378 And http://article.gmane.org/gmane.comp.lib.boost.devel/191945 I know of a lot of people that would like this dependency on rtti removed. My question to the list is: did anyone have a try at an implementation and could share it ? I have toyed with boost::function and sp_typeinfo.hpp and tried to replace occurrences of type_info in function_base.hpp, but I don't have enough understanding to finish the work and my early attempts have failed. Thank you, David.

1 0

boost pointer serialisation broken on IBM AIX
by Avi Bahra 20 Nov '09

20 Nov '09

Serialisation of pointers is ok linux, hpux, but has the following exception: archive_exception::unregistered_class on IBM AIX v10.1 using boost 1.39 Trivial case can be made to work. However as the number of files grows, pointer serialisation start/stops working. There seems to undefined behaviour. There is a _known_ problem with serialisation via base pointer, however this problem is exhibited with normal pointer serialisation. It is extremely difficult to create small enough test case which shows the problem. Sometimes a test case will work and at other times a seemly innocuous change (like adding/removing more hierarchy) will show the problem again. There does not appear to be any rhyme or reason why, Could it be a compiler issue ? template<class T> struct polymorphic { static void save( Archive &ar, T & t ){ ........ const boost::serialization::extended_type_info * true_type = i.get_derived_extended_type_info(t); // note:if this exception is thrown, be sure that derived pointer // is either registered or exported. if(NULL == true_type){ boost::serialization::throw_exception( archive_exception(archive_exception::unregistered_class) <<<<<<<<<<<<<<<<<<<<<<<< ); } Has any one else come across this weird behaviour, in fact I would like to here from anyone who has got pointer serialization working consistently on AIX. Any help greatly appreciated Ta, Avi

1 0