July 2006 - Boost-users - lists.preview.boost.org

[endian] Endian types in union
by Dave Dribin 20 Jul '06

20 Jul '06

Hello, This may be more appropriate for the devel list, but someone just pointed out on this list that an endian library has been submitted for approval. I've *just* needed classes like this (actually I wrote my own similar classes), and decided to see how the Boost library stacked up. It looks quite nice, and I'd be excited to see this included. One suggestion I would make is that the endian classes should not have constructors. As of version 0.5, trying to put them into a union causes a compiler error. Take the struct from the example, and add a union: struct header { big4_t file_code; big4_t file_length; little4_t version; little4_t shape_type; union { big4_t foo; little4_t bar; } baz; }; I get the following errors: endian_example.cpp:37: error: member 'boost::integer::big4_t <unnamed>::header::<anonymous union>::foo' with constructor not allowed in union endian_example.cpp:38: error: member 'boost::integer::little4_t <unnamed>::header::<anonymous union>::bar' with constructor not allowed in union By removing all constructors, the errors go away. I cam across this very issue when developing my classes, and found the answer in this comp.lang.c++.moderated post: <http://groups.google.com/group/comp.lang.c++.moderated/msg/ 103835a7b493723e?hl=en&> As an experiment, I modified boost/integer/endian.hpp to remove the constructors, and add an operator=() to see if this worked, and it did: template <typename T, std::size_t n_bytes> class endian< big, T, n_bytes, unaligned > : cover_operators< endian< big, T, n_bytes >, T > { public: typedef T value_type; operator T() const { return detail::load_big_endian<T, n_bytes>(bytes); } void operator=(T i) { detail::store_big_endian<T, n_bytes> (bytes, i); } private: char bytes[n_bytes]; }; I'd suggest making this change to all classes so that they can be used in unions. -Dave

2 7

boost::minmax_element 2D points bounding box
by Tomas Rapkauskas 20 Jul '06

20 Jul '06

Hello, I am newbie using boost library. I have tried to use boost::minmax_element to calculate bounding box of 2D Points, but I failed... Here is some code: struct S2DPoint { int x; int y; }; typedef std::vector<S2DPoint > Poly; Poly vect; ... std::pair<iterator, iterator> result = boost::minmax_element(poly.begin(), poly.end()); this code does not compiles I need to define BinaryPredicate.. My code with out boost looks: bool CalcPolyMinMax(const S2DPoint poly[], int nCount, S2DPoint &ptMin, S2DPoint &ptMax) { if ( nCount < 1 ) { return false; } ptMin = poly[0]; ptMax = poly[0]; S2DPoint pt; for ( int i1 = 0; i1 < nCount; ++i1 ) { pt = poly[i1]; UpdateMinMaxPoints(ptMin, ptMax, pt); } return true; } void UpdateMinMaxPoints(S2DPoint &ptMin, S2DPoint &ptMax, const S2DPoint &pt) { if ( ptMin.x > pt.x ) { ptMin.x = pt.x; } if ( ptMax.x < pt.x ) { ptMax.x = pt.x; } if ( ptMin.y > pt.y ) { ptMin.y = pt.y; } if ( ptMax.y < pt.y ) { ptMax.y = pt.y; } } Question: How to transform UpdateMinMaxPoints function to the BinaryPredicate? Thanks, Tomas

2 1

Valid arguments for boost::math::gcd()
by Matthias Hofmann 20 Jul '06

20 Jul '06

Hello everybody, I just learned that boost offers a function to determine the greatest common divisor of two integers, as described here: http://www.boost.org/regression-logs/cs-win32_metacomm/doc/html/boost_math/g reatest_common_divisor_and_least_common_multiple.html However, this documentation is not very clear about which arguments are valid for boost::math::gcd(). In particular, it does not specify that the first argument has to be greater than the second one, although this seems to be a requirement with the current implementation. Also, I would like to know how the greatest common divisor is defined when one or both of the arguments are zero. In that case, boost::math::gcd() seems to return the nonzero argument, if there is any - but is this mathematically correct? -- Matthias Hofmann Anvil-Soft, CEO http://www.anvil-soft.com - The Creators of Toilet Tycoon http://www.anvil-soft.de - Die Macher des Klomanagers

1 0

[bgl] Problem with topological_sort
by Loïc Joly 20 Jul '06

20 Jul '06

Hello, When I run the example code of topo_sort.cpp in the BGL examples folder, it works perfectly. However, if I replace in this example: typedef adjacency_list<vecS, vecS, directedS, property<vertex_color_t, default_color_type> > Graph; By: typedef adjacency_list<listS, listS, directedS, property<vertex_color_t, default_color_type, property<vertex_index_t, int > > > Graph; It compiles, but during execution, some assert occurs. This assert occurs deep within my std::vector implementation (msvc8.0), in the call stack, the latest reference to boost is > TopoSort.exe!boost::iterator_property_map<std::_Vector_iterator<enum boost::default_color_type,std::allocator<enum boost::default_color_type> >,boost::adj_list_vertex_property_map<boost::adjacency_list<boost::listS,boost::listS,boost::directedS,boost::property<enum boost::vertex_color_t,enum boost::default_color_type,boost::property<enum boost::vertex_index_t,int,boost::no_property> >,boost::no_property,boost::no_property,boost::listS>,int,int const &,enum boost::vertex_index_t>,enum boost::default_color_type,enum boost::default_color_type &>::operator[](void * v=0x00355eb8) Line 349 + 0x67 bytes C++ Or, for people who don't read msvc on a daily basis, it indicated the line 350 of property_map.cpp : /*template<...> class iterator_property_map : ... { */ inline R operator[](key_type v) const { return *(iter + get(index, v)) ; } Does anybody have any idea of what the problem might be ? Best regards, -- Loïc

2 2

[function]unwanted conversion
by Sean Huang 20 Jul '06

20 Jul '06

I was surprised that the following code compiled: #include <boost/function.hpp> int main(int argc, char argv[]) { typedef boost::function< void ( void ) > MyFunction; MyFunction test1 = char(); MyFunction test2 = 0; return 0; } Is there any reason why boost::function allows such conversions? Thanks, Sean

2 2

Re: [Boost-users] [function]unwanted conversion
by Sohail Somani 19 Jul '06

19 Jul '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org > [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Sean Huang > Sent: Wednesday, July 19, 2006 1:22 PM > To: boost-users(a)lists.boost.org > Subject: [Boost-users] [function]unwanted conversion > > I was surprised that the following code compiled: > > #include <boost/function.hpp> > > int main(int argc, char argv[]) > { > typedef boost::function< void ( void ) > MyFunction; > MyFunction test1 = char(); > MyFunction test2 = 0; > return 0; > } That is a funny one... I wonder if its interpreting it as a function pointer because its 0. If you try it with a non-zero value it doesn't work.

1 0

Re: [Boost-users] Asio & strand & threads
by Zola 19 Jul '06

19 Jul '06

Boosters, Ok -- turns out I really didn't know what I was doing. And I really didn't want strand at all- I'm still fuzzy on strand. But here's the code for what I really did want. And the thread group is handling the Handlers properly. Now my question is... Why is a deadline timer the way it is? What's it good for? Why isn't a delay/timer like a wrapper, that you can give a function object to, and then push that into the io_service queue with a delay? So you could do something like, io_s.post(deadline_timer(boost::posix_time::milliseconds(100), &f)); ? - Zola #include <iostream> #include <boost/asio.hpp> #include <boost/thread/thread.hpp> #include <boost/thread/mutex.hpp> #include <boost/thread/xtime.hpp> #include <boost/bind.hpp> #include <boost/function.hpp> #include <boost/ref.hpp> #include <boost/date_time/posix_time/posix_time.hpp> #include <vector> namespace asio = boost::asio; int const num_threads = 10; std::vector<boost::thread*> tv(num_threads); boost::mutex tv_mutex; boost::mutex io_mutex; class Test { public: Test(asio::io_service& io_s, unsigned int n) : io_s(io_s), n(n), f(boost::bind(&Test::print, this)) { io_s.post(f); } void print() { if (n >= 0) { { boost::mutex::scoped_lock lock(io_mutex); std::cout << "(thread=" << thread_num() << ") "; std::cout << "Timer: " << n << std::endl; } { boost::mutex::scoped_lock lock(m_mutex); io_s.post(f); --n; } boost::xtime xt; boost::xtime_get(&xt, boost::TIME_UTC); xt.sec += 1; boost::thread::sleep(xt); } } int thread_num() { boost::thread t; for (int i=0; i<num_threads; ++i) { boost::mutex::scoped_lock lock(tv_mutex); if (tv[i] && (*tv[i] == t)) { return i; } } return -1; } private: asio::io_service& io_s; unsigned int n; boost::function<void()> f; boost::mutex m_mutex; }; int main() { asio::io_service io_s; Test t(io_s, 1000); boost::thread_group threads; for (int i=0; i<num_threads; ++i) { boost::mutex::scoped_lock lock(tv_mutex); tv[i] = threads.create_thread(boost::bind(&asio::io_service::run, boost::ref(io_s))); } threads.join_all(); }

1 0

Asio & strand & threads
by Zola 19 Jul '06

19 Jul '06

Hi Boosters, I'm trying to figure out how Asio's strand works. My concept of strand is a bit blurry. Asio's Timer.5 tutorial says: "An boost::asio::strand<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00260…>guarantees that, for those handlers that are dispatched through it, an executing handler will be allowed to complete before the next one is started. This is guaranteed irrespective of the number of threads that are calling boost::asio::io_service::run()<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>. Of course, the handlers may still execute concurrently with other handlers that were *not* dispatched through an boost::asio::strand<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00260…>, or were dispatched through a different boost::asio::strand<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00260…>object." Asio's documentation for boost::asio::io_service::run says: "Run the io_service's event processing loop. The run()<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>function blocks until all work has finished and there are no more handlers to be dispatched, or until the io_service<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>has been interrupted. Multiple threads may call the run()<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>function to set up a pool of threads from which the io_service<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>may execute handlers." The key here is "Multiple threads may call the run()<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>function to set up a pool of threads from which the io_service<http://asio.sourceforge.net/boost_asio_0_3_7/libs/asio/doc/reference/a00139…>may execute handlers." -- This implies to me that it's automatic, and the thread pool will take whatever work it can. I've created a little example which has ten threads calling io_service::run, and printing out which thread is calling a handler given to the deadline timer. The issue here is that when I run this Test, the same thread is calling print() every time, no matter what I do. My understanding of this Test is as so: io_service holds a queue of asynchronous events. You fill up the io service with events, and then you call run on it. If multiple threads call run, they'll do whatever work they can. I'm calling run from 10 threads: Each thread will see what events are there, and handle them if possible. The asynchronous waits are calling a strand wrapper. The strand wrapper is like a mutex, and ensures that no two threads will call print() at the same time. What is going on, and why is the same thread calling print() every time? Is my understanding here completely wrong? - Zola // The Test: #include <iostream> #include <boost/asio.hpp> #include <boost/thread/thread.hpp> #include <boost/bind.hpp> #include <boost/function.hpp> #include <boost/date_time/posix_time/posix_time.hpp> #include <vector> namespace asio = boost::asio; int const num_threads = 10; std::vector<boost::thread*> tv(num_threads); class Test { public: Test(asio::io_service& io, int n, int ms) : n(n), ms(ms), s(io), dt(io, boost::posix_time::milliseconds(ms)), f(boost::bind(&Test::print, this)) { dt.async_wait(s.wrap(f)); } void print() { if (n >= 0) { boost::thread t; for (int i=0; i<num_threads; ++i) { if (tv[i] && (*tv[i] == t)) { std::cout << "(thread=" << i << ") "; break; } } std::cout << "Timer: " << n << std::endl; --n; dt.expires_at(dt.expires_at() + boost::posix_time::milliseconds(ms)); dt.async_wait(s.wrap(f)); } } private: int n; unsigned int ms; asio::strand s; asio::deadline_timer dt; boost::function<void (asio::error&)> f; }; int main() { asio::io_service io; Test t(io, 1000, 5); // 1000 iterations, every 5 ms. boost::thread_group threads; for (int i=0; i<num_threads; ++i) { tv[i] = threads.create_thread(boost::bind(&asio::io_service::run, &io)); } threads.join_all(); }

1 0

adding -fPIC flag for AMD64
by Scott Amort 19 Jul '06

19 Jul '06

Hi All, I am trying to statically link a shared library I am developing with the Boost Regex library. I get the following error: /usr/lib/gcc/x86_64-pc-linux-gnu/4.1.1/../../../../x86_64-pc-linux-gnu/bin/ld: /usr/lib64/libboost_regex.a(instances.o): relocation R_X86_64_32 against `boost::object_cache<boost::re_detail::cpp_regex_traits_base<char>, boost::re_detail::cpp_regex_traits_implementation<char> >::do_get(boost::re_detail::cpp_regex_traits_base<char> const&, unsigned long)::s_data' can not be used when making a shared object; recompile with -fPIC I have run across this once before, and although I am not really clear why, it appears that on my AMD64 platform, if I try to link a static library into a shared library, that static library needs to be compiled with -fPIC (this requirement does not exist for normal x86 platforms). My question is: is there an already available option that I can add when compiling the Boost libraries to enable this setting? I am not at all familiar with the BJam build system. Thanks for any assistance! Best, Scott

3 3

binary streams?
by bringiton bringiton 19 Jul '06

19 Jul '06

I'm wondering whether boost has binary streams (or whether the standard library does and i missed it). I had a quick look at boost serialisation. looks like it maybe too involved for my needs. my only needs would be something as follows: outstream s; s << (int)3; s << (float)4.0; s << std::string("hello"); void* data = s.buffer(); int size = s.size(); instream s(data, size); int a << s; float b << s; std::string b << s; I plan on sending the data over TCP. Portability would also be handy (ie endian issues). I may use asio when it is avail in the next boost release. does that support binary streams?

2 1