######################################################################
# libkrb
# Kris Beevers <[email protected]>
######################################################################

This is a C++ class library with generically useful classes for doing
some fairly advanced and/or high-volume stuff.

Most of the library consists of generic data structures and
algorithms, including:

* Bloom filter
* Counting Bloom filter
* Timeout Bloom filter
* Lossy hash table
* LRU key/value memory cache
* Ring buffer
* Mersenne twister RNG
* Murmur hash function
* Discrete PMF sampling
* Online working set size estimation
* Generic resource pool with support for user-defined sizing policies
* Libevent-based thread pool for asynchronous job execution
* LC-Trie for prefix set membership

There are also a few more utilitarian classes:

* Cached time_t/timeval for high volume servers
* Synchronization of workers at a barrier
* String and strerror-based exceptions
* Mutex locking objects
* Apache CLF log entry parser
* Apache CLF log file playback
* Simple config parser that works with boost's program_options

No doubt I will add more.


######################################################################

Requirements:

* g++ 4.4 or greater
* Boost header file library
* Pthreads
* Libevent (for thread pools)

Compiling a program:

Some of the data structures in this library are implemented entirely
in header files.  However, you'll often need to link with the library
since a lot of the data structures rely on, say, murmur_hash.

You should link against -lkrb and -lpthread.  If you are using
thread_pool.hpp, you must also link against -levent.


######################################################################

Installing:

Just run (make && make install).