memcached for win32, new java api

Gregory Block gblock at
Mon Nov 29 14:41:14 PST 2004

On 29 Nov 2004, at 22:36, Erik Mogensen wrote:
>> Hmmm, not sure why one would want the added overhead of running a
>> windows server for a memcached server?!
> Well, some companies do have hoards of Windoze boxes.  Say no more.

The problem is going to be libevent, pretty much no matter what happens 
here.  If someone has a hard on for Win32 WaitNextEvent(), libevent on 
Win32 is at 0.7c, and has serious known faults; it should be updated to 
0.9, and ideally maintained by somebody.

Unless someone's willing to take on that maintenance, only win32-cygwin 
is a feasible alternative, unless someone tries to squeeze more into 
libevent's 0.7c stuff at

What it needs is an evangelist with free time and a penchant for 
MT-safe Win32 event polling.  That kind of geeque is a rare thing, as 
they'll undoubtedly be the first people up against the wall when the 
revolution comes.

More information about the memcached mailing list