Search Results

Search found 1728 results on 70 pages for 'boost bjam'.

Page 50/70 | < Previous Page | 46 47 48 49 50 51 52 53 54 55 56 57  | Next Page >

  • Optimizing sorting container of objects with heap-allocated buffers - how to avoid hard-copying buff

    - by Kache4
    I was making sure I knew how to do the op= and copy constructor correctly in order to sort() properly, so I wrote up a test case. After getting it to work, I realized that the op= was hard-copying all the data_. I figure if I wanted to sort a container with this structure (its elements have heap allocated char buffer arrays), it'd be faster to just swap the pointers around. Is there a way to do that? Would I have to write my own sort/swap function? #include <deque> //#include <string> //#include <utility> //#include <cstdlib> #include <cstring> #include <iostream> //#include <algorithm> // I use sort(), so why does this still compile when commented out? #include <boost/filesystem.hpp> #include <boost/foreach.hpp> using namespace std; namespace fs = boost::filesystem; class Page { public: // constructor Page(const char* path, const char* data, int size) : path_(fs::path(path)), size_(size), data_(new char[size]) { // cout << "Creating Page..." << endl; strncpy(data_, data, size); // cout << "done creating Page..." << endl; } // copy constructor Page(const Page& other) : path_(fs::path(other.path())), size_(other.size()), data_(new char[other.size()]) { // cout << "Copying Page..." << endl; strncpy(data_, other.data(), size_); // cout << "done copying Page..." << endl; } // destructor ~Page() { delete[] data_; } // accessors const fs::path& path() const { return path_; } const char* data() const { return data_; } int size() const { return size_; } // operators Page& operator = (const Page& other) { if (this == &other) return *this; char* newImage = new char[other.size()]; strncpy(newImage, other.data(), other.size()); delete[] data_; data_ = newImage; path_ = fs::path(other.path()); size_ = other.size(); return *this; } bool operator < (const Page& other) const { return path_ < other.path(); } private: fs::path path_; int size_; char* data_; }; class Book { public: Book(const char* path) : path_(fs::path(path)) { cout << "Creating Book..." << endl; cout << "pushing back #1" << endl; pages_.push_back(Page("image1.jpg", "firstImageData", 14)); cout << "pushing back #3" << endl; pages_.push_back(Page("image3.jpg", "thirdImageData", 14)); cout << "pushing back #2" << endl; pages_.push_back(Page("image2.jpg", "secondImageData", 15)); cout << "testing operator <" << endl; cout << pages_[0].path().string() << (pages_[0] < pages_[1]? " < " : " > ") << pages_[1].path().string() << endl; cout << pages_[1].path().string() << (pages_[1] < pages_[2]? " < " : " > ") << pages_[2].path().string() << endl; cout << pages_[0].path().string() << (pages_[0] < pages_[2]? " < " : " > ") << pages_[2].path().string() << endl; cout << "sorting" << endl; BOOST_FOREACH (Page p, pages_) cout << p.path().string() << endl; sort(pages_.begin(), pages_.end()); cout << "done sorting\n"; BOOST_FOREACH (Page p, pages_) cout << p.path().string() << endl; cout << "checking datas" << endl; BOOST_FOREACH (Page p, pages_) { char data[p.size() + 1]; strncpy((char*)&data, p.data(), p.size()); data[p.size()] = '\0'; cout << p.path().string() << " " << data << endl; } cout << "done Creating Book" << endl; } private: deque<Page> pages_; fs::path path_; }; int main() { Book* book = new Book("/some/path/"); }

    Read the article

  • C++ MPL or_, and_ implementations

    - by KRao
    Hi, I am trying to read the boost headers to figure out how they managed to implement the or_<...> and and_<...> metafunctions so that: 1) They can have an arbitrary number of arguments (ok, say up to 5 arguments) 2) They have short circuit behavior, for example: or_<false_,true_,...> does not instantiate whatever is after true_ (so it can also be declared but not defined) Unfortunately the pre-processor metaprogramming is making my task impossible for me :P Thank you in advance for any help/suggestion.

    Read the article

  • How to package Qt Framework on Mac OS X?

    - by Chris Kaminski
    I'm developing an application for the Mac, and I'm wicked new at this - what would be the best way to distribute the Qt Framework so that I'm not impacting an already existing framework, and so that my App.app picks up the libraries. I'm also using boost, so I'm curious how a simple bundle works with libraries.

    Read the article

  • Controlling Solr score/sort

    - by Znarkus
    I want to filter a property within a range, but items that does not have the property should come last in the result. My solution was to set it to -1 if the property was not set. +(property:[10000000001 TO 10000000019] property:"-1"^0.5) This doesn't work, since every document with property:-1 get a very high score, for some reason. Is there a way to reliably control the sorting here? Boosting the range instead would mean I must boost every other term, which I'd rather not do.

    Read the article

  • Xcode C++ debugging

    - by mousebird
    I'm looking for a way to have the Xcode IDE pick up on the contents of my weird Boost template classes. Basically, classes complex enough that Xcode can't display their contents correctly. Is there something like the Objective C -description method or toString() in Java that Xcode will look for? At the moment I'm just implementing print() methods and invoking them in gdb, but that's likely to confuse other developers.

    Read the article

  • OpenCV performance in different languages

    - by h0b0
    I'm doing some prototyping with OpenCV for a hobby project involving processing of real time camera data. I wonder if it is worth the effort to reimplement this in C or C++ when I have it all figured out or if no significant performance boost can be expected. The program basically chains OpenCV functions, so the main part of the work should be done in native code anyway.

    Read the article

  • OpenCV compliation on linux: how to feed to it specific zlib lib?

    - by myWallJSON
    I want to compile OpenCV with same zlib as I use for compilation of Boost Iostreams (not system default one). I want to compile OpenCV as static lib, having zlib compiled as static lib. Currently I use something like : ../$CMAKE_PATH -DCMAKE_INSTALL_PREFIX=./$OPENCV_INSTALL_SUBDIR -DBUILD_WITH_STATIC_CRT=ON -DBUILD_SHARED_LIBS=OFF -DBUILD_PYTHON_SUPPORT=OFF -DOPENCV_EXTRA_C_FLAGS=-fPIC -DOPENCV_BUILD_3RDPARTY_LIBS=TRUE make make install echo Done! I wonder: having some $ZLIB_HEADERS and $ZLIB_LIB_FILES_FOLDER path strings how to feed them into cmake to get OpenCV compiled with built by me zlib?

    Read the article

  • waiting for 2 different events in a single thread

    - by João Portela
    component A (in C++) - is blocked waiting for alarm signals (not relevant) and IO signals (1 udp socket). has one handler for each of these. component B (java) - has to receive the same information the component A udp socket receives. periodicaly gives instructions that should be sent through component A udp socket. How to join both components? it is strongly desirable that: the changes to attach component B to component A are minimal (its not my code and it is not very pleasent to mess with). the time taken by the new operations (usually communicating with component B) interfere very little with the usual processing time of component A - this means that if the operations are going to take a "some" time I would rather use a thread or something to do them. note: since component A receives udp packets more frequently that it has component B instructions to forward, if necessary, it can only forward the instructions (when available) from the IO handler. my initial ideia was to develop a component C (in C++) that would sit inside the component A code (is this called an adapter?) that when instanciated starts the java process and makes the necessary connections (that not so little overhead in the initialization is not a problem). It would have 2 stacks, one for the data to give component B (lets call it Bstack) and for the data to give component A (lets call it Astack). It would sit on its thread (lets call it new-thread) waiting for data to be available in Bstack to send it over udp, and listen on the udp socket to put data on the Astack. This means that the changes to component A are only: when it receives a new UDP packet put it on the Bstack, and if there is something on the Astack sent it over its UDP socket (I decided for this because this socket would only be used in the main thread). One of the problems is that I don't know how to wait for both of these events at the same time using only one thread. so my questions are: Do I really need to use the main thread to send the data over component A socket or can I do it from the new-thread? (I think the answer is no, but I'm not sure about race conditions on sockets) how to I wait for both events? boost::condition_variable or something similar seems the solution in the case of the stack and boost::asio::io_service io_service.run() seems like the thing to use for the socket. Is there any other alternative solution for this problem that I'm not aware of? Thanks for reading this long text but I really wanted you to understand the problem.

    Read the article

  • Advice on Mocking System Calls

    - by Robert S. Barnes
    I have a class which calls getaddrinfo for DNS look ups. During testing I want to simulate various error conditions involving this system call. What's the recommended method for mocking system calls like this? I'm using Boost.Test for my unit testing.

    Read the article

  • Compare versions as strings

    - by Dmitriy
    Comparing version numbers as strings is not so easy... "1.0.0.9" "1.0.0.10", but it's not correct. Obvious way to do it properly is parse these strings, convert to numbers and compare as numbers. Is it other way to do it more "elegant"? For example, boost::string_algo...

    Read the article

  • Is there any documentation for the Cassandra Erlang interface?

    - by Zubair
    I have looked everywhere, and to use Cassandra from Erlang you end up having to download (amongst others): boost thrift : and then you have generate the erlang library by hand, and then copy lib files and beams files. Once you have the whole thing working there is absolutely zero documentation anywhere. If anyone could show me some user friendly documentation it would be much appreciated.

    Read the article

  • Debugging with cgdb -- want to debug only my code

    - by rkbang
    Hello all, I am using standard template libraries along with boost. I want to debug code using cgdb. But I just want to debug my code while doing next and step in cgdb, like right now it enters the code in other libraries whereas I want it should just return or step into the code which I have coded and wrote down. Is there any way to make this happen.

    Read the article

  • Sorting a 2 dimensional array on multiple columns

    - by Anon
    I need to sort a 2 dimensional array of doubles on multiple columns using either C or C++. Could someone point me to the algorithm that I should use or an existing library (perhaps boost?) that has this functionality? I have a feeling that writing a recursive function may be the way to go but I am too lazy to write out the algorithm or implement it myself if it has been done elsewhere. :-) Thanks

    Read the article

  • How to set configuration properties in VS once and for all?

    - by atch
    In VS 2010RC I have to specify configuration properties and specifically included path every time I'm creating new project. Is there a way to do it just once for all future projects? I'm asking this for a reason that I'm starting to use Boost libraries and I have to specify all those paths every time I'm creating project which is bit tedious.

    Read the article

< Previous Page | 46 47 48 49 50 51 52 53 54 55 56 57  | Next Page >