The old version was not working well on blocking input streams, like from remote hosts where the read was not always full. And by not working well, I mean getting exactly 2 buffer lengths into large files before calling them done.
This new one works quite well, I've run a dozen GB through it so far.
The old version was not working well on blocking input streams, like from remote hosts where the read was not always full. And by not working well, I mean getting exactly 2 buffer lengths into large files before calling them done.
This new one works quite well, I've run a dozen GB through it so far.