January 2010 - Boost-users - lists.preview.boost.org

Problems with boost serialization/shared_ptr/container
by Guo, Yanchao 22 Jan '10

22 Jan '10

Hi I am using boost/1.41.0, and the following code give me compilation error when I try to deserialize a shared_ptr. The serialize part it compiled successfully. Can someone advise me if this is a bug in my code or a general issue for boost? Thanks. Yanchao #include <iomanip> #include <iostream> #include <cstddef> // NULL #include <fstream> #include <string> #include <list> #include <cstdio> // remove #include <boost/config.hpp> #if defined(… [View More]

2 2

Q: N00b spinlock
by Moritz Moeller 21 Jan '10

21 Jan '10

I'm new to multi-threaded programming. So I apologize for aksing what is possibly a silly question. Will below code give me the behaviour of what is commonly called a "spinlock"? boost::unique_lock< boost::shared_mutex > guard( mutex, boost::try_to_lock ); if ( !guard.owns_lock() ) { while ( !guard.try_lock() ) { boost::thread::yield(); } } // access data Cheers, Moritz

4 8

boost::thread object initialization. cannot understand.
by rabbitko＠tenbit.pl 21 Jan '10

21 Jan '10

hallo Group Members compilation of following simple program leads to an error: p1.cpp: In function "int main(int, char**)": p1.cpp:18: error: request for member "join" in "f", which is of non-class type "boost::thread(F (*)())" #include <boost/thread.hpp> #include <iostream> struct F { void operator()() { while (true) { std::cout << "hallo" << std::endl; } } }; int main(int argc, char *argv[]) { // boost::thread f = boost::thread(F()); // But … [View More]

4 5

[thread] NULL cleanup routine for thread_specific_pointer
by Jeremiah Willcock 21 Jan '10

21 Jan '10

There was a post in 2008 (http://lists.boost.org/boost-users/2008/05/36589.php) about using a NULL cleanup function for thread_specific_ptr, but no response to the issue. It appears that there is now a special case in the code for thread_specific_ptr to not call NULL cleanup functions; is this official (and thus can be relied upon)? I did not see it mentioned in the documentation. -- Jeremiah Willcock

1 0

A [bimap] getting iterator on bimap second side
by ishay lavi 21 Jan '10

21 Jan '10

2 1

Problem in boost/boost_1_40/libs/serialization/example/demo.cpp
by John Austin 21 Jan '10

21 Jan '10

This is a versioning bug. The first of the following two lines is missing from the example shipped with 1.40: BOOST_CLASS_VERSION(bus_schedule::trip_info, 2) // missing from example // BOOST_CLASS_VERSION(bus_schedule, 2) // not necessary for example to work. When the schedule is recreated, the driver is not copied to the bus_schedule as the field is contained in the nested struct. When I replace the second macro with the first one, the example works.

1 0

Question about the boost::thread_group
by watsonsong 21 Jan '10

21 Jan '10

I want to use thread_group for multi-thread, and maybe using it like this: void thread2(); void thread1(boost::thread_group& tg) { // create an other thread in the thread_group tg.create_thread(thread2); // do some work for(;;) { //... } } void thread2() { // do some work for(;;) { //... } } int main() { boost::thread_group tg; tg.create_thread(boost::bind(thread1, boost::ref(tg))); tg.join_all(); return 0; } I am puzzling about the join_all() function. Whether is it safe to invoke … [View More]

2 1

Re: [Boost-users] fusion struct adapt macro, construct the struct from a fusion vector
by Hicham Mouline 21 Jan '10

21 Jan '10

----- Original Message ----- From: "Hicham Mouline" <hicham(a)mouline.org> To: <boost-users(a)lists.boost.org> Sent: Tuesday, January 19, 2010 4:22 PM Subject: Re: [Boost-users] fusion struct adapt macro,construct the struct from a fusion vector >> What about: >> >> template <typename Seq> >> params(Seq const& seq, >> typename boost::enable_if< >> fusion::traits::is_sequence<Seq> >> >::type* = NULL); &… [View More]

3 8

Getting Original Regular Expression Out From sregex (Boost Xpressive)
by Yan Cheng Cheok 21 Jan '10

21 Jan '10

I have the following code. sregex rex = sregex::compile( "(\\w+) (\\w+)!" ); How I can get "(\w+) (\w+)!" out from rex? Thanks and Regards Yan Cheng CHEOK

2 1

[interprocess] shared map of string to string?
by Anthony Foiani 21 Jan '10

21 Jan '10

I'm using boost.interprocess to build a map from string to string that can be shared between multiple processes, with all read/write access protected by a mutex. I'm having a hard time making it work. Here's a cut-down version of the declaration: template < typename KeyType, typename ValueType, typename StoredKeyType = KeyType, typename StoredValueType = ValueType > class SharedMap : private boost::noncopyable { … [View More]public: SharedMap( const std::string & shmName, const std::size_t shmSize, const std::size_t initBucketCount = 10 ); ~SharedMap(); void setOneValue( const KeyType & key, const ValueType & value ); bool getOneValue( const KeyType & key, ValueType & value ); void clear(); private: typedef std::pair< const StoredKeyType, StoredValueType > StoredPairType; typedef boost::interprocess::managed_shared_memory ShmManagerType; typedef boost::interprocess::allocator< StoredPairType, ShmManagerType::segment_manager > ShmAllocatorType; typedef boost::interprocess::interprocess_mutex ShmMutexType; typedef boost::interprocess::scoped_lock< ShmMutexType > ShmScopedLock; typedef boost::unordered_map< StoredKeyType, StoredValueType, boost::hash< StoredKeyType >, std::equal_to< StoredKeyType >, ShmAllocatorType > ShmMapType; std::string m_sShmName; ShmManagerType m_manager; ShmMutexType * m_pMutex; ShmMapType * m_pMap; }; // end class SharedMap typedef SharedMap< /* KeyType = */ std::string, /* ValueType = */ std::string, /* StoredKeyType = */ boost::interprocess::string, /* StoredValueType = */ boost::interprocess::string > SharedStringMap; And some important bits of the implementation: #define SHARED_MAP( ret_type ) \ template < \ typename KeyType, \ typename ValueType, \ typename StoredKeyType, \ typename StoredValueType \ > \ ret_type \ SharedMap< \ KeyType, \ ValueType, \ StoredKeyType, \ StoredValueType \ > SHARED_MAP()::SharedMap( const std::string & shmName, const std::size_t shmSize, const std::size_t initBucketCount /* = 10 */ ) : m_sShmName( shmName ), m_manager( boost::interprocess::open_or_create, shmName.c_str(), shmSize ), m_pMutex( m_manager.find_or_construct< ShmMutexType >( "mutex" )() ), m_pMap( m_manager.find_or_construct< ShmMapType >( "map" )( initBucketCount, boost::hash< StoredKeyType >(), std::equal_to< StoredKeyType >(), m_manager.get_allocator< StoredPairType >() ) ) { } SHARED_MAP()::~SharedMap() { try { m_pMutex->unlock(); } catch ( ... ) { } } SHARED_MAP( void )::clear() { ShmScopedLock lock( *m_pMutex ); m_pMap->clear(); } It works fine the first time I run the program, but on the second attempt I get a SIGSEGV when I try to clear the map. My test loop is basically: create or find map for each rep clear map insert values done GDB tells me that this error is originating here: Program received signal SIGSEGV, Segmentation fault. __libc_free (mem=0x8143a48) at malloc.c:3709 3709 if (chunk_is_mmapped(p)) /* release mmapped memory. */ (gdb) bt #0 __libc_free (mem=0x8143a48) at malloc.c:3709 #1 0x001efd12 in operator delete (ptr=<value optimized out>) at ../../../../libstdc++-v3/libsupc++/del_op.cc:44 #2 0x08053d49 in __gnu_cxx::new_allocator<char>::deallocate (this=0xb7fa4334, __p=0x8143a48 <Address 0x8143a48 out of bounds>) at /usr/lib/gcc/i686-redhat-linux/4.4.2/../../../../include/c++/4.4.2/ext/new_allocator.h:95 #3 0x0805e677 in boost::interprocess_container::containers_detail::basic_string_base<std::allocator<char> >::deallocate (this=0xb7fa4334, p=0x8143a48 <Address 0x8143a48 out of bounds>, n=24) at /usr/include/boost/interprocess/containers/container/string.hpp:309 #4 0x0805dac2 in boost::interprocess_container::containers_detail::basic_string_base<std::allocator<char> >::deallocate_block (this=0xb7fa4334) at /usr/include/boost/interprocess/containers/container/string.hpp:341 #5 0x0805d167 in boost::interprocess_container::containers_detail::basic_string_base<std::allocator<char> >::~basic_string_base (this=0xb7fa4334, __in_chrg=<value optimized out>) at /usr/include/boost/interprocess/containers/container/string.hpp:119 #6 0x0805bfe7 in {b::ic::string}::~basic_string (this=0xb7fa4334, __in_chrg=<value optimized out>) at /usr/include/boost/interprocess/containers/container/string.hpp:617 #7 0x0805d671 in std::pair<{b::ic::string} const, {b::ic::string} >::~pair (this=0xb7fa4334, __in_chrg=<value optimized out>) at /usr/lib/gcc/i686-redhat-linux/4.4.2/../../../../include/c++/4.4.2/bits/stl_pair.h:68 #8 0x08063636 in boost::unordered_detail::destroy<std::pair<{b::ic::string} const, {b::ic::string} > > (x=0xb7fa4334) at /usr/include/boost/unordered/detail/hash_table.hpp:233 #9 0x08062d55 in boost::unordered_detail::hash_table_data_unique_keys<ip::allocator<std::pair<{b::ic::string} const, {b::ic::string} >, ip::segment_manager<char, ip::rbtree_best_fit<ip::mutex_family, ip::offset_ptr<void>, 0u>, ip::iset_index> > >::allocators::destroy (this=0xb7f9f0d4, ptr=...) at /usr/include/boost/unordered/detail/hash_table_impl.hpp:133 #10 0x080619fc in boost::unordered_detail::hash_table_data_unique_keys<ip::allocator<std::pair<{b::ic::string} const, {b::ic::string} >, ip::segment_manager<char, ip::rbtree_best_fit<ip::mutex_family, ip::offset_ptr<void>, 0u>, ip::iset_index> > >::delete_to_bucket_end (this=0xb7f9f0d4, begin=...) at /usr/include/boost/unordered/detail/hash_table_impl.hpp:832 #11 0x0805fc63 in boost::unordered_detail::hash_table_data_unique_keys<ip::allocator<std::pair<{b::ic::string} const, {b::ic::string} >, ip::segment_manager<char, ip::rbtree_best_fit<ip::mutex_family, ip::offset_ptr<void>, 0u>, ip::iset_index> > >::clear_bucket ( this=0xb7f9f0d4, b=...) at /usr/include/boost/unordered/detail/hash_table_impl.hpp:867 #12 0x0805e2ff in boost::unordered_detail::hash_table_data_unique_keys<ip::allocator<std::pair<{b::ic::string} const, {b::ic::string} >, ip::segment_manager<char, ip::rbtree_best_fit<ip::mutex_family, ip::offset_ptr<void>, 0u>, ip::iset_index> > >::clear ( this=0xb7f9f0d4) at /usr/include/boost/unordered/detail/hash_table_impl.hpp:879 #13 0x0805d918 in boost::unordered_map<{b::ic::string}, {b::ic::string}, boost::hash<{b::ic::string} >, std::equal_to<{b::ic::string} >, ip::allocator<std::pair<{b::ic::string} const, {b::ic::string} >, ip::segment_manager<char, ip::rbtree_best_fit<ip::mutex_family, ip::offset_ptr<void>, 0u>, ip::iset_index> > >::clear (this=0xb7f9f0c4) at /usr/include/boost/unordered/unordered_map.hpp:271 #14 0x0805cdc3 in Tony::SharedMap<{std::string}, {std::string}, {b::ic::string}, {b::ic::string} >::clear (this=0xbffff40c) at SharedMap.cpp:203 #15 0x0804e3e4 in main (argc=1, argv=0xbffff744) at SharedMapTest.cpp:99 I've done the following substitutions: std::basic_string<char, std::char_traits<char>, std::allocator<char> > -> {std::string} boost::interprocess_container::basic_string<char, std::char_traits<char>, std::allocator<char> > -> {b::ic::string} ip:: -> ip:: Interestingly, it looks like the mutex survives program stops and starts just fine: if I try to run my test case a third time, it sits and waits on the mutex (which got locked just before the 'clear' crashed out in the 2nd invocation) Platform is GCC 4.4.2 (fedora rpm 4.4.2-20) with the RPM packages of boost (1.39.0-8). Any hints would be very much appreciated. I can get a full working test case up on the web, but I need to pull out some bits (especially from the test routine) that rely on other parts of my project. Alternately, if anyone knows of a freely-available implementation of this functionality (or, even better, one that stores std::strings directly instead of having to go to/from b::ip::strings, and maybe one that auto-sizes its shared memory pool)... I'd *definitely* like to hear about that. :) Thanks in advance! Regards, Tony [View Less]

2 1