febird

febird implemented a serialization framework(vs boost.serialization/google.protocolbuffer), can be used in protocol parsing, big/small data serialization, even in very small object serialize, performance is good. (such as key/data serialization in BerkeleyDB), it provide fast performance(30~80 times faster than boost.binary_archive), and lower memory usage. febird.rpc is a C++ remote procedure call without an IDL supporting, it based on the serialization framework. febird.rpc provide convenient usage and fast performance, and an uniform coding style.

most stl algorithm(sort/heap/find...) for C, provide high performance(same as stl) through C-Preprocessor Tricky.

thread red-black tree implementation, kernel is written in C, and C++ interface(std::map/set interface) is supplied also, my implementation is fast(no compare function overhead) but has no code explosion, because I used a C-preprocessor tricky.

febird 实现了一个序列化框架（对比boost.serializaiton/google.protocolbuffer），可以用在协议解析，大/小数据的序列化，有极高的性能（比boost.binary_archive快30~80倍），甚至对于非常小的对象，例如只有几个字节的对象，这在序列化BerkeleyDB中key/data这么小的对象（可能只是一个整数/变长整数）时非常有用。 该库提供了对BerkeleyDB的序列化封装，可以象使用std::map一样使用它。 该库也实现了一个不需要IDL的rpc，使用几个宏，很方便的自动完成函数参数的序列化，比MFC的MessageMap还要方便。

stl算法的C版本，使用了一种C预处理器技巧，提供和stl相当的性能(例如febird_sort远快于qsort，与stl::sort相当)。

线索红黑树（Threaded Red-Black Tree, C核心+C/Cpp接口，仿std::set/map）, 在提供高性能的同时，没有C++的代码膨胀问题，因为使用了一个C预处理(C-preprocessor)技巧，详见：http://blog.csdn.net/whinah/archive/2009/05/26/4218292.aspx

使用时请尽量checkout最新版

@see http://blog.csdn.net/whinah http://blog.csdn.net/whinah/archive/2008/11/07/3248730.aspx http://blog.csdn.net/whinah/archive/2008/11/07/3248770.aspx

Tags

  boost idl protocolparsing rpc protocolbuffer serialize febird fast coroutine losertree

Share