May 2008 - Boost-users - lists.preview.boost.org

[threads] unable to UPX boost 1.35
by Daniel Önnerby 09 May '08

09 May '08

Hi! After upgrading my app from boost 1.34.1 to 1.35 I found that I am unable to pack my exe-files using upx.exe. The error message upx gives is: CantPackException: TLS callbacks are not supported I'm guessing this has anything to do with boost::threads that I'm using? Any way to get this fixed? I mean, it worked using boost 1.34.1? Best regards Daniel

2 2

[Thread] shared_mutex problem
by Mark Westcott 09 May '08

09 May '08

Hi all, I'm experiencing a deadlock in the following code. After an timed out call to boost::shared_mutex::timed_lock(), boost::mutex::state.exclusive_waiting_blocked equals 1, which makes further attempts to shared_lock the mutex block indefinitely. Is this a bug, or are my expectations wrong? Thanks, Mark #include <boost/thread/thread.hpp> #include <boost/thread/shared_mutex.hpp> #include <iostream> using namespace std; boost::shared_mutex mutex; void thread() { boost::system_time timeout=boost::get_system_time() + boost::posix_time::milliseconds(500); if (mutex.timed_lock(timeout)) { mutex.unlock(); } else { cout << "thread1: failed to obtain exclusive lock" << endl; // now mutex.state.exclusive_waiting_blocked == 1; cout << "thread1: attempting to obtain exclusive lock" << endl; mutex.lock_shared(); //never get to here cout << "thread2: obtained shared lock" << endl; mutex.unlock_shared(); } } int main(int argc, char* argv[]) { mutex.lock_shared(); boost::thread my_thread(&thread); my_thread.join(); mutex.unlock_shared(); return 0; }

4 7

memory mapped io and boost
by peter_foelsche＠agilent.com 09 May '08

09 May '08

Dear All, from what it looks like the memory mapped io classes in boost::interprocess cannot deal with readonly files. This is a big problem. I think opening readonly files should be possible but all write operations to the file should fail. Of course mapping a readonly file should be done with the matching flags so that the mapping can be shared and that the matching address space is readonly. Also another drawback is that at least on Windows disk-space is being allocated when growing a file. This is slow and it can be avoided by making the file sparse. On UNIXs this is the default. Thanks Peter

3 4

Tar a single boost library - Graph
by Menasheh Fogel 09 May '08

09 May '08

Hi- I would like to use the Boost.Graph library and not all of the Boost library release. I am using a shared Unix server and have a usage quota. I read the instructions on the bcp tool, but it seems I need to tar the whole release before making my single library extraction. Any suggestions on using tar to only extract the Graph library? Thanks! ________________________________________________________________ Sent via the WebMail system at fogelnet.com

1 0

Post again for help
by Match Point 08 May '08

08 May '08

I installed boost library and include files of 1.34 into /home/qhwang/boost_libs/1.34 several months ago. I tried a helloword program of thread tutorial today by command: g++ -g -I /home/qhwang/boost_libs/1.34/include/ -L /home/qhwang/boost_libs/1.34/lib/ -lboost_thread-gcc34-mt-1_34_1 -o helloworld helloworld.cpp The executable was generated. But after I set LD_LIBRARY_PATH and started helloword, there was a segmentation error and a core file was generated. Program terminated with signal 11, Segmentation fault. #0 0x00002aaaaaab5477 in boost::thread_group::join_all () from /home/qhwang/boost_libs/1.34/lib/libboost_thread-gcc34-mt-1_34_1.so.1.34.1 (gdb) bt #0 0x00002aaaaaab5477 in boost::thread_group::join_all () from /home/qhwang/boost_libs/1.34/lib/libboost_thread-gcc34-mt-1_34_1.so.1.34.1 #1 0x0000003fdb4062f7 in start_thread () from /lib64/libpthread.so.0 #2 0x0000003fda8ce85d in clone () from /lib64/libc.so.6 What could be the reason? I tried to install boost 1.35 into /home/qhwang/boost_libs. But when I compiled the same program g++ -I /home/qhwang/boost_libs/include/ -L /home/qhwang/boost_libs/lib/ -lboost_thread-gcc41-mt -o helloworld helloworld.cpp and g++ -I /home/qhwang/boost_libs/include/ -L /home/qhwang/boost_libs/lib/ -lboost_thread-gcc41-mt-1_35 -o helloworld helloworld.cpp But both of them generated link error: /tmp/cc2ZKV4R.o: In function `main': helloworld.cpp:(.text+0x91): undefined reference to `boost::thread::thread(boost::function0<void, std::allocator<boost::function_base> > const&)' collect2: ld returned 1 exit status What could be reason for this? My compiler and os is Linux version 2.6.18-53.1.6.el5 (brewbuilder(a)hs20-bc1-5.build.redhat.com) (gcc version 4.1.2 20070626 (Red Hat 4.1.2-14)) #1 SMP Wed Jan 16 03:56:15 EST 2008 Thanks. Qihong -- View this message in context: http://www.nabble.com/Post-again-for-help-tp17093647p17093647.html Sent from the Boost - Users mailing list archive at Nabble.com.

2 2

Bug in enable_if?
by Germán Diago Gómez 08 May '08

08 May '08

Hello. I'm trying to compile this code, but it doesn't do the trick and I don't know why. struct print { typedef void result_type; template <class T> void operator()(const T & elem) const { std::cout << elem << std::endl; } }; template <class Tuple, class UnaryF, int i = 0> struct tuple_for_each_helper { private: Tuple & t_; UnaryF f_; public: typedef typename UnaryF::result_type result_type; tuple_for_each_helper(Tuple & t, UnaryF f) : t_(t), f_(f) { } result_type operator()(typename boost::enable_if<is_same<result_of<UnaryF (const Tuple &)>, void> >::type * = 0) { f_(get<i>(t_)); typename boost::mpl::if_c<i == tuple_size<Tuple>::value - 1, tuple_for_each_helper<Tuple, UnaryF, -1>, tuple_for_each_helper<Tuple, UnaryF, i + 1> >::type(t_, f_)(); } }; template <class Tuple, class UnaryF> struct tuple_for_each_helper<Tuple, UnaryF, -1> { typedef typename UnaryF::result_type result_type; tuple_for_each_helper(Tuple & t, UnaryF f) {} result_type operator()() {} }; template <class Tuple, class UnaryF> typename result_of<UnaryF (const Tuple &)>::type tuple_for_each (Tuple & t, UnaryF f, typename boost::enable_if<is_same<result_of<UnaryF (const Tuple &)>, void> >::type * = 0) { tuple_for_each_helper<Tuple, UnaryF>(t, f)(); } int main(int argc, char * argv[]) { tuple<int, float, string, string> t(3, 5.3, "hello", "goodbye"); tuple_for_each(t, print()); } And the error is: tuple_each.cpp: In function ‘int main(int, char**)’: tuple_each.cpp:99: error: no matching function for call to ‘tuple_for_each(std::tr1::tuple<int, float, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::tr1::_NullClass, std::tr1::_NullClass, std::tr1::_NullClass, std::tr1::_NullClass, std::tr1::_NullClass, std::tr1::_NullClass>&, print)’ If I change boost::enable_if by boost::disable_if the code compiles. But if I check typeid(typename print::result_type).name() and typeid(result_of<UnaryF (const Tuple &)>::type).name() they are the same. I'm using g++ 4.2.3 and boost 1.34.1. Thanks in advance.

9 25

[build] alignment issues
by Ovanes Markarian 08 May '08

08 May '08

Hello *, for some reason we have in our project fixed alignment of 4bytes (Win32, VC71, Boost 1.34.0). Unfortunately boost was either compiled with default alignment or boost threads compilation ignored the alignment setting. In this regards I have the following questions: 1. Are boost libs being tested with different alignments especially with fixed ones? Is it safe to set an alignment of 4bytes and maybe later of 8bytes? 2. Is there a special flag (for 1.34.0) which needs to be passed to bjam to set a fixed alignment? Currently these are the only setting which are overridden in user.hpp: #define BOOST_DISABLE_ABI_HEADERS #define BOOST_ALL_DYN_LINK Many thanks, Ovanes

1 0

Re: [Boost-users] [MultiIndex] re-indexing on load
by Mark Westcott 08 May '08

08 May '08

Mark Westcott escribi?: > > Hi Joaquin. > > [...] > > I think, for the time being at least, the performance is ok now (using > > ordered rather than hashed indices). > > Is there any chance that future versions of MI might be able to store > > the hashes/complete ordering whilst serializing, or is that Just Not > > Possible (TM)? > > > I think this would pose problems, as hashes are not guaranteedly > platform portable, so storing > them transparently to the user could trigger subtle bugs. > But maybe for your particular applicatin you can consider caching the > hash values --from your > description, this can not only accelarate serialization, but also the > general lookup ops, given > that your hash algorithm is expensive. This is very neat! Thanks a lot for the suggestion/code. Mark

1 0

Can any body help me figure it out? Thanks.
by Peisheng Wang 08 May '08

08 May '08

g++ boo_t.cpp -L../resource/boost_1_35_0/bin.v2/libs/thread/build/gcc-4.1.2/release/threading-multi/libboost_thread-gcc41-mt-1_35.so.1.35.0 -Lpthread boo_t.cpp: In function ‘void increment_count()’: boo_t.cpp:14: error: reference to ‘mutex’ is ambiguous boo_t.cpp:10: error: candidates are: boost::mutex mutex /usr/include/boost/thread/mutex.hpp:35: error: class boost::mutex code: #include <boost/thread.hpp> #include <boost/thread/mutex.hpp> #include <iostream> using namespace std; using namespace boost; int count = 0; boost::mutex mutex; void increment_count() { boost::mutex::scoped_lock lock(mutex); std::cout << "count = " << ++count << std::endl; } int main(int argc, char* argv[]) { boost::thread_group threads; for (int i = 0; i < 10; ++i) threads.create_thread(&increment_count); threads.join_all(); "boo_t.cpp" 24L, 480C 1,1 Top c -- Peisheng Wang Software Engineer iZENEsoft (Shanghai) Co., Ltd Room 601 Marine Tower, No. 1 Pudong Ave. Tel:86-21-68860698-114 Fax:86-21-68860699 Company Website:<a target="_blank" href="http://www.izenesoft.com/">www.izenesoft.com</a> <a href="../../../../index.php?menuaction=felamimail.uicompose.compose&send_to=Wmhlbi5ZdUBNb3JnYW5TdGFubGV5LmNvbQ==">Zhen.Yu(a)MorganStanley.com</a>

3 2

online change of iostream filter
by Daniel Oberhoff 08 May '08

08 May '08

I was wondering if it is possible to change the filter used to write to or read form a file without closing/reopening the file? The usecase is reading/writing matlab files, which contain zipped blocks within a context of unzipped data. So I would need to be able to switch on compression once such a block starts and switch off afterwards, and it would be nice not to have to ftell/close/create appropriate stream/open/ fseek every time. Best Daniel

1 0