But I suspect you could get a lot of performance wins just by formatting out to a larger buffer, say 8KB, and then writing out in bulk increments once it's filled to 4KB or something.
Have implemented this and want to send a new pull request. Just sorting out the memory buffering for the moment.
Also, have extended the -dump functionality with an L option, to express the buffer length in lines ...
Once it's in a good state i'll commit and share the code in my branch so you can have a look first before i pull request.
In reference to #70 ...
Have implemented this and want to send a new pull request. Just sorting out the memory buffering for the moment. Also, have extended the -dump functionality with an L option, to express the buffer length in lines ... Once it's in a good state i'll commit and share the code in my branch so you can have a look first before i pull request.