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

[MultiIndex] getting the index from set::lower_bound() iterator (or a replacement/extension)
by U.Mutlu 07 May '15

07 May '15

Hi, I need a container like std::set which delivers also the index when using the member function lower_bound(). IMO std::set lacks this important functionality. I cannot use std::distance(myset.begin(), it), because that is very slow as it walks again all the items from the beginning. Can boost MultiIndex be used for this, or any other alternatives/workarounds? Here's a demonstration of what I need: struct TS { /*...*/ }; set<TS> myset; TS S(/*...*/); auto it = myset.lower_bound(S); if (it != myset.end()) { size_t ix = ...howto get the index, ie. the rank?... //... } -- Thx Uenal

3 2

boost container help
by Uthpal Urubail 07 May '15

07 May '15

Experts, I am looking for help on a container where I could group the objects having similar properties Ex: struct phyProp { unsigned int color; double height; double weight; std::string location; }; Can I use boost for grouping the objects having same property? [With tolerance?] Any help is highly appreciated. Regards, UJ

3 2

error: type/value mismatch at argument 1 in template parameter
by Lane 06 May '15

06 May '15

bounded_buffer is taken from the boost circular_buffer examples which wraps operations with lock acquisition and release (more type-safe). I've changed up the calls slightly for my app and the problem is, I am having difficulty compiling. I was hoping to make the calls like below w/o the need for a template function (as in the example) since I'm in the function where I'd like to make the calls. --- bounded_buffer<MyData> bb(200); Consumer<bounded_buffer> consumer(bb); Producer<bounded_buffer> producer(bb); boost::thread* c_thread = new boost::thread(consumer); boost::thread* p_thread = new boost::thread(producer); threads.add_thread(c_thread); threads.add_thread(p_thread); --- Here are the compile errors: error: type/value mismatch at argument 1 in template parameter list for 'template<class Buffer> class Consumer' error: expected a type, got 'bounded_buffer' The example code I'm using is, http://www.boost.org/doc/libs/1_55_0/libs/circular_buffer/example/circular_…

2 1

Can a mapped_file larger than 2GB be random seeked?
by Narcélio Filho 05 May '15

05 May '15

(I'm resending this because it seems that my other message got garbled.) I'm writting on random positions on a mapped file, but it seems that if I seek past 2GB, write something, then seek to a lower position, I got a segfault. Here is the code: #include <cassert> #include <iostream> #include <fstream> #include <boost/iostreams/stream.hpp> #include <boost/iostreams/device/mapped_file.hpp> using namespace std; using boost::iostreams::stream; using boost::iostreams::mapped_file; int main(int argc, char** argv) { const size_t GIGA = 1024l*1024l*1024l; const size_t FILESIZE = 2l * GIGA + 10000; ofstream block("file", ios::binary | ios::out); assert(block.good()); block.seekp(FILESIZE); block.write("", 1); block.close(); mapped_file mappedFile("file"); stream<mapped_file> file(mappedFile, ios::binary|ios::out); assert(mappedFile.is_open()); char buffer[10] = {0}; cout << 0 << endl; file.seekp(0); file.write(buffer, 1); cout << FILESIZE - 1 << endl; file.seekp(FILESIZE - 1); file.write(buffer, 1); cout << 1 * GIGA << endl; file.seekp(1 * GIGA); file.write(buffer, 1); return 0; } I'm compiling under Ubuntu 14.10 (amd64): $ g++ boost-mmap.cpp -lboost_iostreams $ ./a.out 0 2147493647 Segmentation fault (core dumped) Everything is 64-bits: $ file a.out a.out: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.32, BuildID[sha1]=d39e3dc0c99e666631747e496bf798b92e4d4c71, not stripped I have the same issue compiling with Visual Studio and running under Windows 8 (amd64). What I'm doing wrong? -- []s, Narcélio.

3 4

boost thread + container.
by Andreas Hehn 05 May '15

05 May '15

Hello boost-users readers, I stumbled across some funny behavior of the headers of boost::thread and boost::container/boost::move for C++O3. I am unsure if this is a known shortcoming of those libraries or if it is a bug. I didn't find anything on the topic in the documentation so far. The problem is dependent on the include order of the two files: * boost/container/vector.hpp * boost/thread/thread.hpp The following code causes several compilation errors (boost-1.58.0/git, g++-4.9.2) related to the move emulation of boost::thread: #include <boost/container/vector.hpp> #include <boost/thread/thread.hpp> int main() { boost::container::vector<boost::thread> v; v.resize(10); return 0; } It prints several error messages starting with boost/container/detail/copy_move_algo.hpp:285:112: error: no matching function for call to ‘move(boost::thread&)’. The following error messages are also boost::move releated "failed" enable_ifs and so on. If I switch the order of the includes everything compiles just fine. Is there a requirement on the order of the includes that I overlooked or is this a bug? Best, Andreas

1 0

How to stage one sub libary in C++ boost
by J Ronald 05 May '15

05 May '15

1. I have been built one sub library as below successfully: b2 locale toolset=msvc address-model=64 link=static threading=multi runtime-link=shared variant=release 2. Then, I tried to run: b2 stage Unexpectedly, it'll build other sub libraries, instead of copy the built sub sub library to the stage directory as the document says. Is there any automatic way to get the final built result in one directory?

1 0

Efficiently pass data from thread to main
by Lane 05 May '15

05 May '15

Using boost, I start a thread from the main process as such. MyWorkThread newThread; newThread = boost::thread(newThread); And in the worker thread class (MyWorkThread), I have a read function which reads from a USB device as much data as it can deliver. This function can be called several hundred times a second. MyThread::read() } //... while () { MyData data; if (hasMore(dev, &data)) // lots of data here // need to get it back to the main thread } } } I'm trying to figure out how to efficiently get this data back to the main process to be placed in a buffer since it can be called so often. - Would making a callback to the main process be the way to go? If so, any pointers how using boost? - Or should I just have a global data buffer and store it directly?

2 1

[serialization] std::shared_ptr with Boost >= 1.53
by Rüdiger Berlich 04 May '15

04 May '15

Hi there, I need to support various versions of Boost, but am in the process of porting our entire code base to C++11. This involves replacing those constructs of Boost that are now in the standard with their C++11-counterparts. A prime example is std::shared_ptr. However, most of our code needs to be serializable. I know that serialization of std::shared_ptr is possible with Boost 1.58. But 1.53 (the oldest version I need to support) cannot serialize std::shared_ptr . Is there any reason not to include the 1.58-Version of #include <boost/serialization/shared_ptr.hpp> instead of the older version, if I detect a Boost-version < 1.58 ? Thanks and Best Regards, Beet

4 12

[Multiprecision] Get mantissa of a float
by Leonardo Yvens 04 May '15

04 May '15

Hello, I'm new to the Multiprecision library, and am trying to extract the mantissa from a float. Using cpp_bin_float, it's easy to get the exponent, but how can I get the mantissa in a normalized form (1.xxxxxxxxxx)? I tried messing with the limbs, they contain large powers of two, but I couldn't interpret that. Thank you very much, Leo Y.

2 2

Can a mapped_file bigger than 2GB be random seeked?
by Narcélio Filho 04 May '15

04 May '15

Hello! I'm writting on random positions on a mapped file, but it seems that if I seek past 2GB, write something, then seek to a lower position, I got a segfault. Here is the code: #include <iostream> #include <fstream> #include <boost/iostreams/stream.hpp> #include <boost/iostreams/device/mapped_file.hpp> using namespace std; using boost::iostreams::stream; using boost::iostreams::mapped_file; int main(int argc, char** argv) { const size_t GIGA = 1024l*1024l*1024l; const size_t FILESIZE = 2l * GIGA + 10000; ofstream block("file", ios::binary | ios::out); assert(block.good()); block.seekp(FILESIZE); block.write("", 1); block.close(); mapped_file mappedFile("file"); stream<mapped_file> file(mappedFile, ios::binary | ios::out); assert(mappedFile.is_open()); char buffer[10] = {0}; cout << 0 << endl; file.seekp(0); file.write(buffer, 1); cout << FILESIZE - 1 << endl; file.seekp(FILESIZE - 1); file.write(buffer, 1); cout << 1 * GIGA << endl; file.seekp(1 * GIGA); file.write(buffer, 1); return 0; } I compile and run on Ubuntu 14.10/amd64: $ g++ example.cpp -lboost_iostreams 0 2147493647 Segmentation fault (core dumped) I also get the same error on Windows 8/amd64. What I'm doing wrong? []s, Narcélio.

1 0