October 2009 - Boost - lists.preview.boost.org

[thread][thread_specific_ptr]
by Sid Sacek 28 Oct '09

28 Oct '09

In VC++ and GCC a user creates global TLS like this: VC++ __declspec( thread ) int tls_1; GCC __thread int tls_2; But using the boost library boost::thread_specific_ptr< int > tls_3; doesn't provide an initialized global variable as one would expect from thread-local-storage. Using tls_3 crashes the program if it's not pre-initialized to some object. The implication is that every time a new thread is created, it needs to first instantiate all global TLS objects. Am I not understanding the intentions of this class? Is there a proper or intended way of using this class? I couldn't find any example code in the boost archives. Thanks, -Sid

2 2

Can we suppress integer_traits warnings?
by John Maddock 27 Oct '09

27 Oct '09

If you have a program that simply includes boost/integer_traits.hpp and is compiled with g++ -pedantic you get a bunch of warnings: ./boost/integer_traits.hpp:164:57: warning: use of C99 long long integer constant ./boost/integer_traits.hpp:164:73: warning: use of C99 long long integer constant ./boost/integer_traits.hpp:170:62: warning: use of C99 long long integer constant ./boost/integer_traits.hpp:170:62: warning: use of C99 long long integer constant And no matter where I insert __extension__ I can't seem to get gcc to suppress these:-( Any ideas? BTW I'm asking because this came to me as a - somewhat misplaced - bug report: https://svn.boost.org/trac/boost/ticket/1451. Thanks, John.

7 17

[task in sandbox] memory leak?
by Jeffrey Hellrung 27 Oct '09

27 Oct '09

All: A colleague of mine is trying to use the proposed Boost.Task library in the sandbox, but he is observing some memory leaks. See forwarded email below and attached cpp file. Does anyone have any information or suggestions for him? I agreed to query the list for him, but if this becomes too involved, I'll ask him to subscribe and communicate directly. Thanks! - Jeff -------- Original Message -------- Subject: Boost.task memory leak Date: Tue, 27 Oct 2009 15:14:44 -0700 From: Paul Jones <pajones(a)gmail.com> To: Jeffrey Hellrung <jhellrung(a)ucla.edu> References: <7f0a1570910271324o3dcb2d9xb361501267b7a5db(a)mail.gmail.com> This code leaks memory on my system: - gcc version 4.4.1 (Ubuntu 4.4.1-4ubuntu8) - Boost 1.40 - boost.task-0.3.0 When this code is run you can observe the program's allocated memory steadily increasing in the system monitor. The output displays the number of (undestroyed) functor objects when all the pool has completed all of its tasks. This seems to be capped at 64, so this isn't the source of the leak. The maintainer's name is Oliver Kowalke. But I couldn't find his email address. Paul

1 0

asio library
by Peter Foelsche 27 Oct '09

27 Oct '09

Out of curiousity (reading somebody comment here regarding socket calls) I looked into the asio library and found the following example code below. My question is: What is the use of the socket object created in the first line? I think this line must always be followed by some more code. Because of this it does not make sense, that a library exports this functionality. boost::asio::ip::tcp::socket socket(io_service); boost::system::error_code ec; socket.open(boost::asio::ip::tcp::v4(), ec); if (ec) { // An error occurred. }

4 7

Proposal for Library to perform dependency injection
by Adam Young 27 Oct '09

27 Oct '09

Having had a career that has bounced me between a few languages, I have had the opportunity to work with a couple of tools in the Java world that didn't see, to have analogues in C++. One of these is the inversion of control containers, of which Spring is one of the best known. Ken Fowler suggested the term "Dependency Injection" as a way to better describe how inversion of control is performed, and I prefer his term. I've written a fairly simple dependency injection framework in C++ that seems to make proper use of the C++ Language. It is not attempting to do any COM like component model. Rather, it allows the developer the ability to cleanly separate the Object usage from the construction policy. I've posted the code on Sourceforge at https://sourceforge.net/projects/cpp-resolver/. I have not yet taken the time to get it into proper format for a Sandbox project, but will do that next if there seems to be enough interest in the project. I wrote up the philosophy behind the design here: http://adam.younglogic.com/?p=146 The code is currently g++ only, although I've attempted to use only core language constructs. It has no dependencies on other libraries other than stl collections. Any constructive feed back will be most appreciated. Adam

2 2

[conversion] Release v0.4: Replace function template specialization by ADL
by vicente.botet 27 Oct '09

27 Oct '09

Hi, I have just released the version 0.4 of Boost.Conversion , which includes since my last publicly announced version 0.3: Replace function template specialization by ADL A Source class is convertible to a Target class if: * Either: A function with the signature convert_to<Target>(Source const&, boost::dummy::type_tag<To> const&) is available via argument dependent lookup * Or: A template specialization of boost::conversion::convert_to<Target, Source> exists for Target and Source * Or: Target is copy constructible from Source (default implementation) The template function convert_to allows to convert a source type to a target type, using argument dependent lookup to select a specialized convert_to function if available. If no specialized convert_to function is available, boost::conversion::convert_to is used. The generic convert_to function requires that the elements to be converted are assignable and copy constructible. It is implemented using the Target copy construction from a Source or the Source conversion operator Target - this is sometimes unavailable. For standard types, we can not add a specialized convert_to function on the namespace std. The alternative to using argument dependent lookup in this situation is to provide a template specialization of boost::conversion::convert_to for every pair of standard types that requires a specialized convert_to. You can found all the detail for dowloading, documentation, ... in http://svn.boost.org/trac/boost/wiki/LibrariesUnderConstruction#Boost.Conve… I'd appreciate it if people could take a look, test it on their compiler and give me some feedback. Thanks to Edward Diener for showing me that multiple assignments should be taken in account. Thanks also to Jeffrey Hellrung and Robert Stewart for showing me that Boost.Conversion should use also ADL. BTW, I'm looking for a review manager. Anyone around here who would like to volunteer? Best regards, _____________________ Vicente Juan Botet Escribá

1 0

[system][filesystem v3] A possible error_code argument compromise
by Scott McMurray 27 Oct '09

27 Oct '09

2009/10/20 Beman Dawes <bdawes(a)acm.org>: > > One particular concern is that as currently worded, the N2838 proposal > requires throws() return a null reference, which they see as relying > on undefined behavior. > So it sounds like the idea solution is something that looks like a reference to the caller, for shortness of typing, but looks like a pointer to the implementation, to ease checking and avoid the weirdness of null references. struct optional_error_code_ref { optional_error_code_ref() : p_() {} optional_error_code_ref(optional_error_code_ref const &ecr) : p_(ecr.p_) {} optional_error_code_ref(error_code &r) : p_(&r) {} error_code &operator*() const { return *p_; } error_code *operator->() const { return p_; } operator safe_bool() const { return safe_bool(p_); } private: error_code *p_; }; Then the call is done the same was as before: error_code ec; foo(ec); But it's obvious to check, and never deals in null references: void foo(optional_error_code_ref ec = optional_error_code_ref()) { ... if (ec) { ec->value = bar; } else { throw bar_exception(); } ... } It would of course be better spelt qux<error_code&>, but I'm not sure exactly what name that should be. ~ Scott (Hopefully the subject change will get us out of the exceptions vs error codes quagmire.)

2 2

[python] AttributeError
by Okko Willeboordse 27 Oct '09

27 Oct '09

Running Boost 1.40 ( Windows server 2003/XP, Visual Studio 2008) Given; #include<boost/python.hpp> struct Foo{ static int bar; }; int Foo::bar=3; BOOST_PYTHON_MODULE(foo){ boost::python::class_<Foo>("Foo",boost::python::init<>()) .def_readonly("bar",Foo::bar); } With Python 2.6.3 and 2.6.4 I get; >c:\Python26\python.exe -c "import foo" Traceback (most recent call last): File "<string>", line 1, in <module> AttributeError: 'Boost.Python.StaticProperty' object attribute '__doc__' is read-only Also; >python Python 2.6.4 (r264:75708, Oct 26 2009, 08:23:19) [MSC v.1500 32 bit (Intel)] on win32 Type "help", "copyright", "credits" or "license" for more information. >>> import foo Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: 'Boost.Python.StaticProperty' object attribute '__doc__' is read-only >>> import foo >>> dir(foo.Foo) ['__class__', '__delattr__', '__dict__', '__doc__', '__format__', '__getattribute__', '__hash__', '__init __', '__instance_size__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr_ _', '__sizeof__', '__str__', '__subclasshook__', '__weakref__'] So importing foo for the second time succeeds but foo.Foo holds no bar With Python < 2.6.3 everything works as expected. See also http://mail.python.org/pipermail/cplusplus-sig/2009-August/014747.html

1 0

[Graph] problem with strong_components.hpp and variant
by Nil Geisweiller 27 Oct '09

27 Oct '09

Hi, it seems there is a conflict between strong_components.hpp and variant, maybe the method boost::get is wrongly overloaded. The problem occurs with 1.40 but not with 1.37, and it is still occurs with the last svn. I attach a minimal code to reproduce the bug + compile error ###### test.cpp ###### //uncomment the line below to get the bug #include <boost/graph/strong_components.hpp> #include <boost/variant.hpp> typedef boost::variant < int, float > intorfloat; int main() { int i(2); intorfloat iof(i); return boost::get<int>(iof); //problem is here } ########### compile error ########### a-lin@a-lin-desktop:~/OpenCog/test/bin$ make [100%] Building CXX object CMakeFiles/test.dir/test.cpp.o In file included from /usr/include/boost/graph/graph_concepts.hpp:17, from /usr/include/boost/graph/depth_first_search.hpp:18, from /usr/include/boost/graph/strong_components.hpp:17, from /home/a-lin/OpenCog/test/test.cpp:2: /usr/include/boost/property_map/property_map.hpp: In instantiation of ‘boost::property_traits<int>’: /usr/include/boost/graph/two_bit_color_map.hpp:46: instantiated from ‘boost::two_bit_color_map<int>’ /home/a-lin/OpenCog/test/test.cpp:10: instantiated from here /usr/include/boost/property_map/property_map.hpp:31: error: ‘int’ is not a class, struct, or union type /usr/include/boost/property_map/property_map.hpp:32: error: ‘int’ is not a class, struct, or union type /usr/include/boost/property_map/property_map.hpp:33: error: ‘int’ is not a class, struct, or union type /usr/include/boost/property_map/property_map.hpp:34: error: ‘int’ is not a class, struct, or union type make[2]: *** [CMakeFiles/test.dir/test.cpp.o] Error 1 make[1]: *** [CMakeFiles/test.dir/all] Error 2 make: *** [all] Error 2 Is it a bug? Thanks a lot! Nil

1 1

[interprocess] mapped_file with wide filename on Windows
by Cory Nelson 27 Oct '09

27 Oct '09

Could the mapped_file constructor allow a wchar_t filename on Windows? Or some cross-platform way to get Unicode support? -- Cory Nelson http://int64.org

2 2