[Rabbit-dev] Extending Rabbit

[Robert Olofsson]

On Thu, 18 Mar 2010 20:16:58 +0000
noor nashid <noor.nashid at csebuet.org> wrote:

> I want to download a whole page in Rabbit, and then I want to transfer
> that to the other side (i.e. to the browser/client) using a single
> HTTP pipelined connection. 

Rabbit already uses persistent connections, and there has been some
testing with pipelining, but it is usually the client that opens the 
connections for the requests.So I am not sure what exactly you are planning.

Anyway, how will you know if the browser want the other parts of the page?
I think it will be very hard for you to do this. 

You can, in most advanced browsers tell the browser how many concurrent
connections it may use. Try changing it and see what happen.

> Besides this I would like to use GraphicsMagick image library inside Rabbit.

You mean by using jni and native libs?
Sure, it can be done, but it comes as a cost as well, especially in
memory usage and stability. What will happen if rabbit tries to convert
an image that crashes graphicsmagick? Currently the external process
dies and rabbit keeps working, if you use a native lib, then rabbit
will die as well.

> And later I would also like to add Non-data Prefetching (like
> Connection Prefetching) features inside Rabbit.
> 
> My intention is to make Rabbit pipelining feature enabled. I think all
> these features will make Rabbit much stronger.

It will be interesting to see what you can make.

/robo