Updated C client, performance, etc...

[Sean Chittenden]

Howdy all.  I've updated the API and am calling this 1.0 RC1.  List of 
changes since the last release include:

*) BUG: Found a stack bug that was showing its face on FreeBSD AMD64 
(likely other 64 bit OSes) but not on i386.

*) PERFORMANCE: Removed calls to strlen(3) and replaced them with the 
compile time sizeof operator.  This resulted in about a 10sec speedup 
in a 90sec test.  Right now, using only 25% of the CPU, I'm able to get 
about fifty thousand requests a second (could probably get more, but 
stats are broken in the server on amd64 at the moment).

*) Added different hashing schemes (read more below).

*) Performed various linting activities (found nothing major).

For your downloading pleasure:

http://people.FreeBSD.org/~seanc/libmemcache/


As for future goals, I need to clean up error handling a bit and make 
some decisions about what to do if a server craps out.  I also need to 
integrate my shared memory thought droppings.  As for performance, the 
list of things that could be done for performance on the client end of 
things includes:

*) Using kqueue(2) instead of select(2) would probably be a nice 
winner, but I don't want to deal with interface portability quite yet 
(besides, one file descriptor on a select(2) call isn't too terrible).

*) Using mmap(2) for buffers instead of malloc(3)'ed memory (not having 
to copy data from kernel to user and visa versa is always good).

*) A binary protocol.  The current protocol requires a tad bit of 
searching, but it's not bad.  I don't imagine there would be much of a 
speedup to be had here, but it's an option.

...but I doubt it'll buy much to go down those routes... no app is 
going to be limited by libmemcache's performance (that I'm aware of).

-sc

-- 
Sean Chittenden