- Boost-users - lists.boost.org

Can I get filesystem::basic_path<> template to use backslashes ?
by Timothy Madden 30 Jul '10

30 Jul '10

Hello Is there a way for the basic_path observer functions (string(), root_path(), relative_path(), filename()) to use backslashes instead of forward slashes on Windows ? I know Windows also accepts forward slashes, but they are not suitable for display to the user, in a dialog box. Thank you, Timothy Madden

3 3

How to build for the VS version that I have loaded the vcvars for ?
by Timothy Madden 30 Jul '10

30 Jul '10

Hello I have a project that I would like to include boost in, and I would like to build boost as part of my project build process. However on a computer with both Visual Studio 2008 and Visual Studio 2010 installed the .\bjam command will build the libraries for VS 2010, even though my project compiles with VS 2008. It is pretty useless to build the 2010 libraries for a 2008 project, because they have different names and can not be used from one Visual Studio version to another. Is there a way for the boost build procedure to compile libraries for the Visual Studio version that the environment variables currently point to ? I mean if I build boost with make from a VS 2008 project than the environment variables will point to the VS 2008 installation, and if I build boost with make from a VS 2010 project than the environment variables will point to the VS 2010 installation. Can I get bjam to build boost for the currently loaded configuration, and not the latest version that can be found installed ? Thank you, Timothy Madden

2 2

boost::interprocess segfault when calling pop on deque
by Matthew Stump 30 Jul '10

30 Jul '10

I built a datastrucuture which I called BufferedQueue. It's essentially a double buffered queue but instead of two buffers I allow a user specified number of buffers. Each buffer is a struct which contains the managed_mapped_file for the particular buffer and a pointer to a boost::interprocess::deque. Each mapped file contains only 1 deque and the deque is used to store items pushed onto the queue by the user. I automatically handle rotation between the various buffers when the user calls push or pop. Specific problem: All of my functional unit tests pass without issue and I have 100% test coverage. When I run the multi-threaded unit tests I encounter a segfault, but not immediately. I have some tests which will spawn one thread which instantiates the datastrucutre and will push 100K items onto the queue (this will span multiple buffers). I spawn a second thread which also instantiates the datastrucutre and begin to pop items off of the queue. After about 40K items have been poped off of the queue I get a segfault when calling boost::interprocess::deque::pop_front(). It's always on the same line, and occurs when the segment manager is doing it's sanity checks while de-allocating the memory. Sometimes this occurs when it's the last item in the deque which is being poped, sometimes it's not; there is no detectable pattern. For my test I'm also using ints so I'm not pushing something onto the datastrucutre that isn't compatible with boost::interprocess. Any helpful ideas as to what may be the cause of my problem would be greatly appreciated. // Type declarations for the deque typedef boost::interprocess::managed_mapped_file ManagedFile; typedef ManagedFile::segment_manager SegmentManager; typedef typename boost::interprocess::allocator<T, SegmentManager> TypeAllocator; typedef typename boost::interprocess::list<T, TypeAllocator> MMContainer; // instantiation of the deque queue = file.find_or_construct<MMContainer>(getBufferName(name, buffNum).c_str())(TypeAllocator(file.get_segment_manager())); // my pop method template<class T> void BufferedQueue<T>::pop() { checkBufferCount(); { boost::interprocess::sharable_lock<Mutex> master_lock(*m_masterMutex); boost::interprocess::scoped_lock<typename Buffer<T>::Mutex> lock(*getReadBuffer().mutex); assert(getReadBuffer().queue); if (!isBufferEmpty<T>(getReadBuffer())) { assert(!getReadBuffer().queue->empty()); getReadBuffer().queue->pop_front(); return; } } rotate_reader(); boost::interprocess::sharable_lock<Mutex> master_lock(*m_masterMutex); boost::interprocess::scoped_lock<typename Buffer<T>::Mutex> lock(*getReadBuffer().mutex); if (isBufferEmpty<T>(getReadBuffer())) throw Exception("called pop() on an empty buffer"); assert(!getReadBuffer().queue->empty()); getReadBuffer().queue->pop_front(); } /libs2/external/boost/boost/interprocess/mem_algo/rbtree_best_fit.hpp:1244: void boost::interprocess::rbtree_best_fit<MutexFamily, VoidMutex, MemAlignment>::priv_deallocate(void*) [with MutexFamily = boost::interprocess::mutex_family, VoidPointer = boost::interprocess::offset_ptr<void>, unsigned int MemAlignment = 0u]: Assertion `priv_is_allocated_block(block)' failed. #0 0x0094a410 in __kernel_vsyscall () #1 0x02850df0 in raise () from /lib/libc.so.6 #2 0x02852701 in abort () from /lib/libc.so.6 #3 0x0284a26b in __assert_fail () from /lib/libc.so.6 #4 0x0806ca9f in boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>::priv_deallocate (this=0xb5fcd004, addr=0xb5fcd0b8) at /libs2/external/boost/boost/interprocess/mem_algo/rbtree_best_fit.hpp:1244 #5 0x0807b00b in boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>::deallocate (this=0xb5fcd004, addr=0xb5fcd0b8) at /libs2/external/boost/boost/interprocess/mem_algo/rbtree_best_fit.hpp:1233 #6 0x0807b05f in boost::interprocess::segment_manager_base<boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u> >::deallocate (this=0xb5fcd004, addr=0xb5fcd0b8) at /libs2/external/boost/boost/interprocess/segment_manager.hpp:217 #7 0x0807b5fe in boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> >::deallocate (this=0xb5fcd0b0, ptr=...) at /libs2/external/boost/boost/interprocess/allocators/allocator.hpp:152 #8 0x0807b61c in boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> >::deallocate_one (this=0xb5fcd0b0, p=...) at /libs2/external/boost/boost/interprocess/allocators/allocator.hpp:234 #9 0x0807b693 in boost::container::containers_detail::allocator_destroyer<boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > >::priv_deallocate (this=0x43511dc, p=...) at /libs2/external/boost/boost/interprocess/containers/container/detail/destroyers.hpp:133 #10 0x0807b6ca in boost::container::containers_detail::allocator_destroyer<boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > >::operator() (this=0x43511dc, p=...) at /libs2/external/boost/boost/interprocess/containers/container/detail/destroyers.hpp:143 #11 0x0807b786 in boost::intrusive::list_impl<boost::intrusive::listopt<boost::intrusive::detail::base_hook_traits<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::intrusive::list_node_traits<boost::interprocess::offset_ptr<void> >, normal_link, boost::intrusive::default_tag, 1>, unsigned int, true> >::erase_and_dispose<boost::container::containers_detail::allocator_destroyer<boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > > > ( this=0xb5fcd0b4, i=..., disposer=...) at /libs2/external/boost/boost/intrusive/list.hpp:649 #12 0x0807b7ce in boost::intrusive::list_impl<boost::intrusive::listopt<boost::intrusive::detail::base_hook_traits<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::intrusive::list_node_traits<boost::interprocess::offset_ptr<void> >, normal_link, boost::intrusive::default_tag, 1>, unsigned int, true> >::erase_and_dispose<boost::container::containers_detail::allocator_destroyer<boost::interprocess::allocator<boost::container::containers_detail::list_node<int, boost::interprocess::offset_ptr<void> >, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > > > ( this=0xb5fcd0b4, i=..., disposer=...) at /libs2/external/boost/boost/intrusive/list.hpp:656 #13 0x0807b839 in boost::container::list<int, boost::interprocess::allocator<int, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > >::erase (this=0xb5fcd0b0, p=...) at /libs2/external/boost/boost/interprocess/containers/container/list.hpp:930 #14 0x0807b881 in boost::container::list<int, boost::interprocess::allocator<int, boost::interprocess::segment_manager<char, boost::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0u>, boost::interprocess::iset_index> > >::pop_front (this=0xb5fcd0b0) at /libs2/external/boost/boost/interprocess/containers/container/list.hpp:623 #15 0x0808a2e2 in Ovid::BufferedQueue<int>::pop (this=0x43512f0) at /ovid/server/shared/ovidserver-bufferedqueue.cpp:287 #16 0x080557fb in reader_multithread_two_writer_single_reader_throughput () at test-ovidserver-bufferedqueue.cpp:389 #17 0x07653832 in start_thread () from /lib/libpthread.so.0 #18 0x028f9e0e in clone () from /lib/libc.so.6

2 1

[preprocessor] It works, but how can I make it better/more elegant?
by Geoff Hilton 29 Jul '10

29 Jul '10

Yay! I finally got my type map generating code to work! There has got to be a better way though because it feels really kludgey. Opinions? If you have a more elegant way of accomplishing the same thing (without misusing BOOST_PP_APPLY as I think I did, because I only used it to remove the extra pair of parens surrounding the tuple in TEST_PP_get_type, something that, oddly, BOOST_PP_TUPLE_REM_CTOR doesn't succeed at doing (at least not without breaking the current code). Was I correct in my assumption (which is why I handled it the way I did in my below code) that basically anytime a a comma appears, the two surrounding tokens must be treated as a tuple (and thus parenthesized? This is what I've gathered from the online documentation as well as the "C++ Template Metaprogramming" book. The purpose of this code is to create a type map between array types and corresponding value types of 3rd party classes which don't conform at all to STL usage practices, that which I'd like to use, so I'm writing an array wrapper template class with associated iterator template classes using iterator_facade & the MPL. The map generated by this code is used to map between array and value types for type selection purposes. The following (reformatted for readability) code should generate this (also reformatted) preprocessed output, though I've only tested the real code and not this that is mostly copy/pasted: typedef bmpl::map< bmpl::pair<SomeIntArray, SomeInt>, bmpl::pair<SomeDoubleArray, SomeDouble>, bmpl::pair<SomeBoolArray, SomeBool>, bmpl::pair<SomeOtherClassArray, SomeOtherClass> > supported_types_map; --test.hpp-- (omitting #inclusions) namespace blah { namespace bmpl = boost::mpl; //must match size as below tuples in sequence for readability. #define TEST_PP_tuple_size 2 //index of array type inside tuple for readability. #define TEST_PP_array_descriptor 0 //index of value type inside tuple for readability. #define TEST_PP_value_descriptor 1 //sequence of tuples of supported types #define TEST_PP_supported_types \ ((SomeIntArray, SomeInt)) \ ,((SomeDoubleArray, SomeDouble)) \ ,((SomBoolArray, SomeBool)) \ ,((SomeOtherClassArray, SomeOtherClass))\ /**/ //size of TEST_PP_supported_types (BOOST_PP_SEQ_SIZE doesn't work on it) #define TEST_PP_types_size 4 #define TEST_PP_get_type(index, type_descriptor) \ BOOST_PP_TUPLE_ELEM(TEST_PP_types_tuple_size, \ type_descriptor, \ TEST_PP_get_tuple(index) \ ) \ /**/ //gets tuple from TEST_PP_supported_types #define TEST_PP_get_tuple(index) \ BOOST_PP_APPLY( \ BOOST_PP_TUPLE_ELEM( \ TEST_PP_types_size, index, (TEST_PP_supported_types)\ ) \ ) \ /**/ //this returns a tuple containing the first and last part of the mpl::pair #define TEST_PP_bmpl_pair(index) \ (bmpl::pair<TEST_PP_get_type(index, TEST_PP_array_descriptor), \ TEST_PP_get_type(index, TEST_PP_value_descriptor)>) \ /**/ #define TEST_PP_pair_fwder(next_index, index, data) \ BOOST_PP_TUPLE_REM_CTOR(2,TEST_PP_bmpl_pair(index)) \ /**/ typedef bmpl::map< BOOST_PP_ENUM(TEST_PP_types_size, TEST_PP_pair_fwder, ~) > supported_types_map; } //namespace blah Thank you! Geoff

2 5

[units] Basic usage questions
by Bryce L Nordgren 29 Jul '10

29 Jul '10

Q1 ============================================================================= What are the rules about what is allowed where? Or perhaps more appropriately, why are the rules so strange? I've been playing with units in boost 1.43.0 using gcc 4.4.3/mingw on win xp and I think I've deduced what is allowed. For instance, when declaring variables, you can do: quantity<si::length> l(1.0*si::meters) ; quantity<si::meter_base_unit::unit_type> l(1.0*si::meters) ; quantity<us::yard_base_unit::unit_type> l(1.0*si::meters) ; but not quantity<si::length> l(si::meters) ; (missing the scalar in front of the units) or quantity<si::meters> l(1.0*si::meters) ; (si::meters is not quantity-variable-declaration material??!?) Q2 ============================================================================= I spent a long time looking for a way to construct a variable with on-the-fly defined units, in order to store interim quantities which may have strange units when implementing the computation of a long expression. This would also be useful to precompute some terms which occur more than once, but which don't necessarily represent any kind of sensible value that you would return as a final result. I wanted to do something like this: quantity<si::length / si::mass> dummy_var_1 = (some_interim_expression reducing to si::length/si::mass) but I can't figure out how to do it. For the moment, I plan on just using the vast array of predefined units on the derived dimensions. However, it would be nice to be able to divide up the computation in whatever way seems natural given the equation, even if the interim variables don't represent rational physical quantities. Q3 =============================================================================== And I suppose my last question is "What does a unit system buy you?", meaning that since "yards" is not part of a unit system and I can declare a variable of type quantity<us::yard_base_unit::unit_type>, what is the added benefit of constructing a unit system which contains yards? Don't misinterpret this as a string of complaints. :) I am very grateful to not have to write this myself. I'm just having some trouble understanding how to apply the library to my situation. ;) Thanks, Bryce

2 1

[Serialization] Serialize base and derived classes.
by BlueAngel 29 Jul '10

29 Jul '10

Hi lists, I'm working with the Serialization library but I'm having some troubles with it. The tutorials in the website shown me much but still I need help. I'm using XML archives. I have this classes actually (these classes are all builders/prototypes, they serialize the datas and then they should make a new object with those datas): Serializer (an interface, every class which need serialization must be derived from this, however this is only a marker, it will be not serialized) RenderableSerializer : public Serializer - this isn't pure virtual however shouldn't be never serialized directly (since the object made by this class is a pure virtual one) - These will use the tag <renderable> SkyboxSerializer : public RenderableSerializer - these will use the tag <skybox> EntitySerializer : public Serializer - these will use the tag <entity> many other classes will be added in future and they'll all derive form RenderableSerializer. Now the problem is EntitySerializer have a pointer to a RenderableSerializer and when I serialize EntitySerializer the pointer to that class is considered to be a RenderableSerializer and not the derived class. I don't know at priori which class is the base class and this lead to problem. If I serialize a SkyboxSerializer I have the correct tag: a <skybox> which have inside a <renderable> and the datas are correct. If I try to serialize an Entity I had a lot of problem, mostly just make the library crash. As I've seen in the tutorial I should register all the derived classes before pass the pointer to the Archive. I did it registering the SkyboxSerializer classes but it lead to a wrong xml: a correct <entity> tag which have a correct <skybox> with all it's data but also a wrong <renderable> tag which have inside the same data of the tag <skybox> instead of the data of a RenderableSerializer class. Now the question is: is there a correct way to use the library this way? With a pointer to a base class of which the derived is not known? My code is too big and I made many attempt to use the library so I'm not posting it in this message. If you need it I will post it. Thank you for your replies.

2 5

suspend an asio operation
by Lloyd 29 Jul '10

29 Jul '10

Hi, I am using asio for a network server. Based on the client request the server performs the operation. Some operation could take a very long time. so we would like to implement a "feature" to terminate the long process from client side by sending some other command. Is there any mechanism in "ASIO" to do this? Thanks, Lloyd ______________________________________ Scanned and protected by Email scanner

2 1

[wave] problem with "/##/" macro
by Evgeniy Rigkov 29 Jul '10

29 Jul '10

Hello, I found a problem with preprocessing "/##/" macro. For example, I have a next code: #define COMMENT_MACRO /##/ void test() { COMMENT_MACRO comment } This code can pe preprocessed (with help any correct preprocessor) to next: void test() { // comment } But after usage boost::wave utility (from boost 1.43) I have this code: void test() { comment } How you see the symbol "//" is forgotten. How fix it?

1 0

[asio] Newbie problem running tutorial example Daytime.4 - Asynchronous UDP daytime client
by Adolfo Jimenez 29 Jul '10

29 Jul '10

Hi list: Using MSVCPP C++ 9.0 and Boost_1_43_0, and trying to follow the examples in the tutorial. The Daytime.4 example hangs in line size_t len = socket.receive_from(boost::asio::buffer(recv_buf), sender_endpoint); it seem that wait indefinitely for a response. Using this server: char* serverName = "ntp-nist.ldsbc.edu"; // 198.60.73.8 LDSBC, Salt Lake City, Utah Neither with: char* serverName = "time-nw.nist.gov" or any other in the http://tf.nist.gov/tf-cgi/servers.cgi# list. By the way, the Daytime.1 (asynchronous TCP datytime client) example and Timer1 to 5 examples works fine for me. Any suggestion would be appreciated. Thanks A.J.Millan

1 0

[datetime] Stability of ptime string representation
by Rush Manbert 28 Jul '10

28 Jul '10

I have an application that uses the string representation of a ptime object in a way that requires that representation to be stable across many years and across multiple releases of the DateTime library. We started out using the to_iso_extended_string() method to generate the string representation. I am worried that the periodic addition of leap seconds could mess me up. The scenario is rather complicated: Let's say that I go into production tomorrow using Boost 1.42. All is well. The Powers That Be decide that there will be a leap second added at the end of 2010. We keep using 1.42. A year from now I record a time value. If I did it exactly one year from when I'm writing this, the to_iso_extended_string() representation would be: 2011-07-28T10:59:30.532000 In 2012, we change out our Boost libraries for version 1.50, which contains a new version of the DateTime library which I assume knows about the leap second that was added at the end of 2010. Here's what I'm worrying about: If, sometime in 2012 after we switch libraries, I retrieve the ptime value that I recorded in 2011 by calling to_iso_extended_string(), will I get the same string that I was getting in 2011 when I was running on a pre-leap-second-addition version of DateTime, or will it be different by a second? Thanks, Rush

1 0