August 2007 - Boost-users - lists.boost.org

[thread] Never enter critical section [NC]
by frederic.mayot＠sgcib.com 30 Aug '07

30 Aug '07

Hi, I'm having a problem with mutexes. In one thread T1 I have: while(!end) { /* TTT */ lock lk(mutex); dosomethingquick(); } In another thread T2 I have: void f() { lock lk(mutex); doanotherthingquick(); } I'm sure dosomethingquick returns but T2 can never acquires the mutex. If I had a timer on line /* TTT */ in T1 (sleeps for a millisecond), everything works fine. Can anyone understand such a behavior? Am I doing something wrong here? Thanks, ____________________________ Frederic MAYOT Société Générale - Equity IT 1221 Avenue of the Americas New York, NY 10020 +1 (212) 278-5558 +1 (917) 673-1339 ************************************************************************* This message and any attachments (the "message") are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited. E-mails are susceptible to alteration. Neither SOCIETE GENERALE nor any of its subsidiaries or affiliates shall be liable for the message if altered, changed or falsified. *************************************************************************

5 8

Boost.CMake's boost_add_executable
by ope 30 Aug '07

30 Aug '07

Hi, I've read http://svn.boost.org/trac/boost/wiki/CMakeAddExecutable and try to use it - and it doesn't work correct. This macros isn't used in the boost cmake build (in real there isn't a wave example as the web page promised)! Anyway, I've got the linker error: /usr/bin/ld: cannot find -lboost_program_options-static It should look for libboost_program_options-gcc41-mt-1_34! The part responsible is at tools/build/CMake/BoostCore.cmake: else (LIB MATCHES ".*-.*") # The user has given the name of just the library target, # e.g., "boost_filesystem". We add on the appropriate variant # name(s). list(APPEND THIS_EXE_ACTUAL_DEPENDS "${LIB}${VARIANT_TARGET_NAME}") list(APPEND THIS_EXE_RELEASE_ACTUAL_DEPENDS "${LIB}${RELEASE_VARIANT_TARGET_NAME}") list(APPEND THIS_EXE_DEBUG_ACTUAL_DEPENDS "${LIB}${DEBUG_VARIANT_TARGET_NAME}") endif (LIB MATCHES ".*-.*") using: boost_add_executable(... DEPENDS boost_program_options-gcc41-mt-1_34 ... ) works! Did I something wrong or is there an error inside boost's BoostConfig.cmake? Thanks, Olaf

3 5

boost::multi_index iterators and std::distance
by Jeffrey Holle 30 Aug '07

30 Aug '07

In experimenting with boost::multi_index, I find that std::distance does not work properly multi_index iterators. My multi_index as a single ordered index for now, and I've tried: C cs; Index& index = cs.get<0>(); ... populate cs Index::iterator iter=index.find(7); cout << "distance is \"" << distance(cs.begin(),iter) << "\"" << endl; cout << "distance is \"" << distance(index.begin(),iter) << "\"" endl; Both outputs print that the distance is 0, and should be 7. Note both output lines compile and I've also tried using operator- and operator+ with the iterators and found that they don't exist (as expected). Is there a work around for this?

2 2

[BGL] add_vertex(vp, graph) with mix of internal & bundled properties
by Sara Ting 30 Aug '07

30 Aug '07

Hello, I need to use the overload of add_vertex that takes the vertex property object as a parameter. I am using a mix of internal and bundled properties because I had to add the vertex_index property but did not want to clutter my bundled property with an index field. How do I instantiate this mixed property object to call add_vertex(vp, graph)? For reference, my graph is defined like this: adjacency_list<vecS, listS, indirectedS, boost::property<boost::vertex_index_t, int, MyNodeProperties>, MyEdgeProperties> I basically want to do this call: add_vertex(myNodeProp, graph) but I am getting a compilation error that it is not a boost::property<....> type. How do I wrap my bundled property in an object that I can pass to this function? Thanks, Sara

2 1

error LNK2019: unresolved external symbol
by desai3＠cooper.edu 30 Aug '07

30 Aug '07

I am using boost build v1, VC 8. While building my project I get following errors, ***************************************************************** DQEasy.obj : error LNK2019: unresolved external symbol "public: bool __thiscall DQP::putString(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &,char const *)" (?putString@DQP@@QAE_NABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@PBD@Z) referenced in function "public: class DQP * __thiscall DQEasy::getBaseRequest(void)" (?getBaseRequest@DQEasy@@QAEPAVDQP@@XZ) DQEasy.obj : error LNK2019: unresolved external symbol "public: __thiscall DQP::DQP(void)" (??0DQP@@QAE@XZ) referenced in function "public: class DQP * __thiscall DQEasy::getBaseRequest(void)" (?getBaseRequest@DQEasy@@QAEPAVDQP@@XZ) ******************************************************************** There are 63 unresolved symbols altogether. Any Ideas on how can I fix it? I have tried entering bjam "-sBUILD=<runtime-link>static" instead of bjam, still no help. I am copying my jamfile for reference. Thank you. ********************************************************************* # This is the top of our own project tree project-root ; # Include definitions needed for Python modules import python ; extension DQEasy # Declare a Python extension called DQEasy : DQEasy.cpp # source # requirements and dependencies for Boost.Python extensions <template>@boost/libs/python/build/extension : <include>. ; ***********************************************************************

1 0

Program_options : option couples and order
by Patrick Sizun 30 Aug '07

30 Aug '07

Hi, I would like to know if it's possible, with program_options, to handle the order in which several values of the same options are entered. E.g. : when entering prog -f file1.txt -n 2 -f file2.txt -n 3 I need to get the (f,n) couples right, and the use of the -n option should be mandatory for each file. How is it possible to enforce that? or does it all belong to the post-processing of the options ? Regards, Patrick

1 0

Re: [Boost-users] Program_options assertion failure
by Fletcher, Matthew (Information Technology) 29 Aug '07

29 Aug '07

Well I presume you've already tried turning off the unicode option but I would try recompiling with the setting 'Use Multi-Byte Character Set' and also have 'Treat wchar_t as Built-in Type' set to No. The unicode setting doesn't affect the char *'s, but could certainly affect what happens when you create a std::string from one without explicitly setting a size/allocator. I agree with James that a mismatch between these settings is a very likely cause. Matt -- This e-mail is confidential and the information contained in it may be privileged. It should not be read, copied or used by anyone other than the intended recipient. If you have received it in error, please contact the sender immediately by telephoning +44 (0)20 7623 8000 or by return email, and delete the e-mail and do not disclose its contents to any person. We believe, but do not warrant, that this e-mail and any attachments are virus free, but you must take full responsibility for virus checking. Please refer to http://www.dresdnerkleinwort.com/disc/email/ and read our e-mail disclaimer statement and monitoring policy. Dresdner Kleinwort is the trading name of the investment banking division of Dresdner Bank AG, and operates through Dresdner Bank AG, Dresdner Kleinwort Limited, Dresdner Kleinwort Securities Limited and their affiliated or associated companies. Dresdner Bank AG is a company incorporated in Germany with limited liability and registered in England (registered no. FC007638, place of business 30 Gresham Street, London EC2V 7PG), and is authorised by the German Federal Financial Supervisory Authority and by the Financial Services Authority ('FSA') and regulated by the FSA for the conduct of designated business in the UK. Dresdner Kleinwort Limited is a company incorporated in England (registered no. 551334, registered office 30 Gresham Street, London EC2V 7PG), and is authorised and regulated by the FSA. Dresdner Kleinwort Securities Limited is a company incorporated in England (registered no. 1767419, registered office 30 Gresham Street, London EC2V 7PG), and is authorised and regulated by the FSA.

2 4

metaclass must be exactly Boost.Python.class, not a subclass?
by william.r.chesters＠jpmorgan.com 29 Aug '07

29 Aug '07

Hi, I'd like to create a metaclass M which I can use to construct new Python classes C inheriting from a C++ base class B. B's metaclass is of course "Boost.Python.class", so I inherit M from that. It's not quite working for me though, because when I try to pass an instance of one of the C's back to C++, I get a "TypeError: No registered converter was able to extract a C++ pointer to type B from this Python object of type C". It seems that BP is testing for "type(type(obj)) is Boost.Python.class" rather than "isinstance(Boost.Python.class, type(type(obj))". Is this an easy issue to fix or work around? Many thanks, William This communication is for informational purposes only. It is not intended as an offer or solicitation for the purchase or sale of any financial instrument or as an official confirmation of any transaction. All market prices, data and other information are not warranted as to completeness or accuracy and are subject to change without notice. Any comments or statements made herein do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries and affiliates. This transmission may contain information that is privileged, confidential, legally privileged, and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is STRICTLY PROHIBITED. Although this transmission and any attachments are believed to be free of any virus or other defect that might affect any computer system into which it is received and opened, it is the responsibility of the recipient to ensure that it is virus free and no responsibility is accepted by JPMorgan Chase & Co., its subsidiaries and affiliates, as applicable, for any loss or damage arising in any way from its use. If you received this transmission in error, please immediately contact the sender and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. Please refer to http://www.jpmorgan.com/pages/disclosures for disclosures relating to UK legal entities.

1 0

serialization lib, const data, documentation
by Robbie Morrison 29 Aug '07

29 Aug '07

Hello all Regarding the Boost.Serialization library. Archiving to XML requires the use of name-value-pairs for all serialized objects, be they stand-alone or encapsulated. I wanted to confirm the following is correct. In reference to the online and bundled HTML documentation: http://www.boost.org/libs/serialization/doc/serialization.html#const file:///.../boost_1_34_1/libs/serialization/doc/serialization.html#const It would appear that the code block which reads: ar & const_cast<T &>(t); should also mention that, in the case of XML serialization: using boost::serialization::make_nvp; // for convenience ar & make_nvp("t", const_cast<T &>(t)); // cast used on load Moreover, this code could be a candidate for a macro, something like (or would this create its own can of worms?): ar & BOOST_SERIALIZATION_NVP_CONSTDATA(T, t); ar & BOOST_SERIALIZATION_CONSTDATA_NVP(T, t); --- On that note, a new page describing user Boost.Serialization macros and their usage might be helpful. For instance I have created the following list (readers should note that my comments may well be incorrect): BOOST_CLASS_EXPORT_GUID(SubClass, "MySC") // GUID is "globally unique id" BOOST_CLASS_EXPORT(SubClass) // shortcut if class name is acceptable BOOST_CLASS_VERSION(Class, 2) // reset class version, the default is zero BOOST_IS_ABSTRACT(AbstractClass) // only some compilers [which ones ?] BOOST_SERIALIZATION_NVP(data) // shortcut if object name acceptable [1] BOOST_SERIALIZATION_BASE_OBJECT_NVP(Base) // shortcut if object name acceptable [2] BOOST_SERIALIZATION_SPLIT_MEMBER() // 'serialize' to 'save'/'load' BOOST_SERIALIZATION_SHARED_PTR(Class) // depreciated [3] BOOST_CLASS_TRACKING(MyVirtualBaseClass, boost::serialization::track_always) BOOST_SHARED_POINTER_EXPORT(T) // depreciated from Boost version 1.32 BOOST_SHARED_POINTER_EXPORT_GUID(T, K) // depreciated from Boost version 1.32 BOOST_SERIALIZATION_LIBRARY_VERSION // yields a "release" number [4] [1] boost::serialization::make_nvp("data", data); [2] requires #include <boost/serialization/base_object.hpp> [3] more specifically, for types without tracking set, that is using boost::serialization::track_never also under 1.34.1 this generates code only for particular Metrowerks and Borland compilers [4] requires #include <boost/serialization/serialization.hpp> (but doesn't appear to be maintained) I don't mind drafting out such a page. Could it be part of the official web documentation or would I be better to add it to the boost wiki as an informal contribution. many thanks in advance Robbie --- Robbie Morrison PhD student -- policy-oriented energy system simulation Institute for Energy Engineering (IET) Technical University of Berlin (TU-Berlin), Germany [from IMAP client]

1 0

Program_options assertion failure
by Johnathan Bunn 29 Aug '07

29 Aug '07

im having an issue with a project using visual studio pro 2005 if it makes a difference the librarys im using are Qt, pdflib, boost, and some crystal com stuff thorugh activex my problem is that when I build in release my command line args are parsed fine and my application behaves as expected. when I build in debug mode, I get an asertion failure as follows Assertion failed: n == name.size()-2, file libs\program_options\src\options_description.cpp, line 122 the relevent code is as follows // process command line args.. po::options_description desc( "TEST" ); desc.add_options( ) ("help", "produce help message" ); ("hideSplash", "hide splash message" ) ("report", po::value<int>( ), "show requested report and quit" ) ; po::variables_map vm; po::store( po::parse_command_line( argc, argv, desc ), vm ); po::notify( vm ); if( vm.count( "help" ) ){ Log::Info::out( ) << "Help Message requested"; // show help window return true; } there is more blocks the same as the last if block that look for my other messages they are commented out now for testing though I am completely stumped and sought help on the #boost irc channel volodya and others on that channel were stumped as well if anyone has any insight into this issue I thank you in advance for your help :-D -- To follow the Path, Look to the master, Walk with the master, See through the master, Become the master

3 3

Boost Parameter Question.
by Surya Kiran Gullapalli 29 Aug '07

29 Aug '07

Hello all, I'm a newbie to boost parameter library. I had no problem in using BOOST_PARAMETER_MEMBER_FUNCTION as long as the function is non static. what is the macro to use for static member function. I gave BOOST_PARAMTER_MEMBER_FUNCTION ( (static void), ... ) But i was getting warning messages with MSVC 8.0 compiler, with boost-1.34.1 warning C4042: 'unnamed-parameter': has bad storage class. I do not know whether these are harmless or not. Please help. Thanks, Surya

2 1

[boost::format] How to not display 0 integer?
by tomasz2k＠poczta.onet.pl 29 Aug '07

29 Aug '07

Hello group I'm using boost::format heavily to format my output for user. I've run into a problem that I cannot resolve for now. I use this format: int i = 4; cout << format ("(%|1$+2|)") % i; // -> (+4) i = -4; cout << format ("(%|1$+2|)") % i; // -> (-4) That's fine and expected. But my question for the one below: int i = 0; cout << format ("(%|1$+2|)") % i; // -> ( 0) // -> ( ) how? How to avoid printing 0? I want formatter to print values in plus or in minus only. When there's value 0 - display only spaces to width. My problem is that I cannot create wrapper over that integer, I must feed formatter with int for various reasons. Does anyone know how to achive this? Maybe I override some parts of library or parsing method? Maybe some of you added your own formatts to formatter and want to share? Any hint or example would be valuable. Thanks in advance, Tomasz Kalkosiñski

1 0

warning C4180 in VS 2005 for bind.hpp
by Adam Badura 29 Aug '07

29 Aug '07

I have a large project at work. We use Visual Studio 2005 with SP 1 and Boost 1.34.0. When we use boost::bind in algorithms on collections we very often get following warning (at level 3): J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored the detailed message is following (just an example extract from compilation log, sadly I was unable to produce a simple stand-alone example - but if it is required then I will give some more work into finding one): J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1575) : warning C4180: qualifier applied to function type has no meaning; ignored J:\ThirdParty\boost_1_34_0\boost/bind.hpp(1609) : see reference to class template instantiation 'boost::_bi::add_cref<Pm,I>' being compiled with [ Pm=boost::shared_ptr<nle::views::ScrollEventListener> , I=1 ] j:\sources\Common/Helpers/Observable.h(128) : see reference to class template instantiation 'boost::_bi::dm_result<Pm,A1>' being compiled with [ Pm=boost::shared_ptr<nle::views::ScrollEventListener> (__thiscall boost::weak_ptr<nle::views::ScrollEventListener>::* )(void), A1=boost::arg<1> ] j:\sources\Common/Helpers/Observable.h(120) : while compiling class template member function 'void common::helpers::Observable<T>::registerListener(const boost::weak_ptr<T> &) const' with [ T=nle::views::ScrollEventListener ] j:\sources\NLE/Views/ScrollInfo.h(65) : see reference to class template instantiation 'common::helpers::Observable<T>' being compiled with [ T=nle::views::ScrollEventListener ] One of my colleagues found a "solution" to this problem. If the modifier "const" is removed in one of typedef's the no warning are generated. However we would like a better solution since modifying the library doesn't seem a good one. What to do? Adam Badura

3 2

Graph: "extra" edge processing during BFS
by Eric Fowler 29 Aug '07

29 Aug '07

I have an algorithm that just screams breadth first search. After some thrashing around in the graph library I am confident in about 90% of the implementation. However there is one little snag: When I examine a grey vertex, I want to do some basic logic on the vertex and the "candidate" next edge to see if what I want to do with this edge and this vertex is "legitimate", according to my own definition of legitimacy(*). If the edge is good, I use it, perhaps relaxing it, if not, I just pass on it and go on to the next one. I would like to do this without tweaking or rewriting the BFS algo. Can this be done with just visitors and the various template parameters or must I write my own BFS variant (not that hard to do of course)? Eric * My idea of "legitimate" is to check to see if a path to a given vertex already contains a certain vertex. The path is built as I run the algorithm. "Parent" vertices are stored as the algorithm runs so I can follow a chain up from the examined vertex back to the root. So the basic logic is: Vertex u = Q.top(); Q.pop(); vis.examine_vertex(u, g); //this is already in BFS if(has_certain_property(u, g)) add_to_tree(u, g); else do_not_add(u, g); .... etc.

1 0

Different functions based on type of iterator?
by HighCommander4 29 Aug '07

29 Aug '07

Let's say I had a function template <typename Iterator> void function(Iterator first, Iterator last); and I want it to call two different functions based on if Iterator is a random access iterator or just a forward iterator. Essentially, I want two versions of this function: template <typename RandomAccessIterator> void function(RandomAccessIterator first, RandomAccessIterator last); template <typename ForwardIterator> void function(ForwardIterator first, ForwardIterator last); I want a function call to function(a, b) to translate to the first one if a and b are random access iterators, and the second one if they are not. I'm thinking of somehow combining enable_if and concept checking, I'm just not sure how... -- View this message in context: http://www.nabble.com/Different-functions-based-on-type-of-iterator--tf4345… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

Linking to DLLs on Windows with GCC
by John Pye 29 Aug '07

29 Aug '07

Hi all, I am having trouble linking against Boost DLLs. Even though I think I am using the correct compiler commands, I seem to end up with statically linked binaries (everything works fine, but Dependency Walker shows that I have no DLL dependency on Boost). Specifically, I am trying to link against Boost:Serialization. Because I'm using MinGW, my Boost libraries had to be compiled by me, so I used "bjam --toolset=gcc --layout=system". I then ran "bjam install" with the same options, which resulted in (over 1GB of) files in c:\Boost. Next I build my own program. It compiles against the Boost files using GCC parameters "-IC:\Boost\include" and then links against Boost::Serialisation using "-LC:\Boost\lib" -lboost_serialization-mt So, what can I do to set GCC to use dynamic linking to the DLL in this case, instead of the static linking that it appears to be using at the moment? (I am a little unclear about the purpose of the .a files currently present in the c:\Boost\lib directory. Do they refer in some way to the DLLs or are they self-contained static libraries?) Cheers JP

1 0

[ptr_container]Swap/reorder elements in ptr_vector
by Bill Buklis 28 Aug '07

28 Aug '07

What is the best or recommended way to swap or reorder elements in a ptr_vector? Naturally, I simply want to swap pointers and not copy any elements. Using swap with the "base" iterator works (see below), but it feels like I'm using an undocumented feature or something. Is there a preferred method? Or is this the best approach? typedef boost::ptr_vector<int> VEC; VEC v; v.push_back( new int(1) ); v.push_back( new int(2) ); VEC::iterator a = v.begin(); VEC::iterator b = v.begin() + 1; // Swap the two underlying pointers using the base iterator std::swap( *a.base(), *b.base() ); -- Bill --

1 0

Adding content to a vertex in a graph
by Stephen Torri 28 Aug '07

28 Aug '07

I made this property graph according to the documentation: class Component { public: typedef boost::shared_ptr<Component> ptr_t; }; class My_Graph { typedef boost::property< boost::vertex_name_t, Component::ptr_t > VertexProperty_t; typedef boost::adjacency_list<boost::vecS, // OutEdgeList boost::vecS, // VertexList boost::directedS, // Directed VertexProperty_t> // VertexProperties Graph_t; Graph_t m_graph; }; I cannot see from the MutablePropertyGraph page how I am suppose to add the VertexProperty_t to the vertex when its created. The documentation for the MutablePropertyGraph shows a use of "add_vertex" giving a vertex_propety_type with the graph. template <class G> struct MutablePropertyGraphConcept { typedef typename boost::graph_traits<G>::edge_descriptor edge_descriptor; void constraints() { function_requires< MutableGraphConcept<G> >(); v = add_vertex(vp, g); p = add_edge(u, v, ep, g); } G g; std::pair<edge_descriptor, bool> p; typename boost::graph_traits<G>::vertex_descriptor u, v; typename boost::graph_traits<G>::vertex_property_type vp; typename boost::graph_traits<G>::edge_property_type ep; }; My confusion lies in the fact that the add_vertex function requires a vertex_property_type the first parameter. So I am unsure how to create a vertex_property_type from a Component::ptr. Is a vertex_property_type equivalent to a Component::ptr? boost::graph_traits<Graph_t>::vertex_property_type vertex_propert_obj = Component::ptr ( new Component() ); boost::graph_traits<Graph_t>::vertex_descriptor new_vertex = add_vertex ( vertex_property_obj, m_graph ); Stephen

1 0

[BGL] Bundled Properties and setS Edgelist container
by Sam Peterson 28 Aug '07

28 Aug '07

I know that if I use a setS for the edgelist container for an adjacency_list class, each edge of my graph will only be added once as I build it. This functionality is very desirable for me, but I've found I can't seem to use it with the equally desirable Bundled Property Maps. Specifically, this line of code doesn't seem to work: (*g)[i].paper_id = atoi(row[0]); g is a pointer to my graph. The exact compiler error with g++ 4.1.1 is: error: invalid conversion from 'size_t' to 'void*' Is there anyway I can use these two features together? If not, I'm going to be very disappointed. If that's the case I'd advise making a large note of it in the bundled properties documentation as I wouldn't have bothered with bundled properties if I knew they carried this limitation. -- Sam Peterson peabody(a)freeshell.org peabodyenator(a)gmail.com

3 7

Short mailing list outage at 9:00PM EDT
by Doug Gregor 28 Aug '07

28 Aug '07

There will be a brief outage of the Boost mailing lists from 9:00PM-9:30PM EDT today. No mail will be lost. Subversion and the Trac will still be available during this time. - Doug

1 0

[BGL] read_graphviz and dynamic properties
by Sam Peterson 28 Aug '07

28 Aug '07

I have been trying (unsuccessfully for the past hour), to read a graphviz file into an adjacency_list type. The program stops with a SIGABRT because I haven't mapped "node_id" to anything within the dynamic properties argument. However, I really have no desire to map node_id to anything; this is a very large graph and I'd like to avoid the memory consumption. If I try to map node id to the vertex_index_t map I receive a compiler error stating that's a const property map. If I want to map node_id to vertex_name_t map, I have to add that property to my adjacency_list type which will bloat the memory consumption. I just want to build the graph without worry about the properties. Is there anyway this can be done easily? The documentation on read_graphviz and write_graphviz are sorely lacking on this subject in my opinion. -- Sam Peterson peabody(a)freeshell.org peabodyenator(a)gmail.com

2 3

boost undefined reference md5.
by chun ping wang 28 Aug '07

28 Aug '07

while trying to link to lib boost md5 i get these errors I build using make with -l"boost_md5-mgw41-1_34_1. booost md5 was build using gcc 4.1.2 mingw. C:/Boost/include/boost-1_34_1/boost/md5.hpp:33: undefined reference to `boost::md5::context::context()' C:/Boost/include/boost-1_34_1/boost/md5.hpp:33: undefined reference to `boost::md5::context::context()' C:/Boost/include/boost-1_34_1/boost/md5.hpp:57: undefined reference to `boost::md5::process(void const*, unsigned int, bool)' C:/Boost/include/boost-1_34_1/boost/md5.hpp:46: undefined reference to `boost::md5::clear()' C:/Boost/include/boost-1_34_1/boost/md5.hpp:46: undefined reference to `boost::md5::clear()'

1 0

[Boost.Test] User Exception handling
by Ewgenij Sokolovski 28 Aug '07

28 Aug '07

Hello. On the webpage I read: In case if your test cases may throw custom exceptions (see here for list of exceptions translated by default), you could register translator specific for that exception. http://www.boost.org/libs/test/doc/components/utf/index.html So, if I understand it right, if I throw some custom exception in my test program and do not catch it, then it is catched by the boost test framework and reported in the proper way - if the exception can be translated by default. Now I look at the exceptions, which can be translated by default and see: In majority of the cases the monitored function doesn't need to throw the boost::execution_exception to report an error in the Execution Monitor. If you want a custom error message to be included into the execution_exception's error message, use one of the following C++ types as an exception: * C string. * std:string. * any exception class in std::exception hierarchy. http://www.boost.org/libs/test/doc/components/execution_monitor/index.html#… So, I can use a C string for my exception. I have written a small example test case and included throwing of an uncaught exception in it: try{ throw "blub"; } catch(char* ex){ throw; } As I started the testing process I didn't get a "blub"-Exception message, but Boost.Test internal framework error: unknown reason So, what's wrong? Why is "blub" not reported as "blub"? Can anybody explain that? Regards Ewgenij

6 8

Serialization in an array
by Hansi Hansi 27 Aug '07

27 Aug '07

Hello, is it possible to serialize a class in a byte-array(e.g. array stream)? best regards Hansjörg _______________________________________________________________________ Jetzt neu! Schützen Sie Ihren PC mit McAfee und WEB.DE. 3 Monate kostenlos testen. http://www.pc-sicherheit.web.de/startseite/?mc=022220

3 2

boost::thread Destructor being called multiple times
by kaliwanagan 27 Aug '07

27 Aug '07

I was just curious as to why the class destructor are (apparently) being called multiple times. I have tried searching the archives to no avail (which prompted me to register and post to the mailinglist). #include <boost/thread/thread.hpp> struct TestingClass { TestingClass() { std::cout << "constructor called" << std::endl; } ~TestingClass() { std::cout << "destructor called" << std::endl; } void operator()() { std::cout << "operator() called" << std::endl; } } testclass; void testBoostThread() { boost::thread thrd (testclass); thrd.join(); } int main () { testBoostThread(); return 0; } The output: constructor called destructor called destructor called destructor called destructor called operator() called destructor called I would appreciate if someone can enlighten me on this. Will having the destructor called multiple times be a cause of concern? Will it cause some undefined symptoms? Is there a way to completely avoid this situation?

4 3