rabbit-proxy-development

[Rabbit-proxy-development] Caching for offline browsing

[Michael V. <mi...@vo...>]

Hello,

I was looking around for a proxy server implemented in Java, as basis
for an experimental project. RabbIT seems to be the most actively
developed, as far as I could find at least -- so first of all,
congratulations on a great project! How are you all using it? As a
"classic" proxy, like Squid etc.? I am interested in use it for
building a proxy server that would run locally on workstations, mostly
laptops that are frequently offline (network disconnected).

How could we get RabbIT to cache all visited pages for this purpose,
even if the server (or upstream proxy) say they can/should not??=20
Based on that base functionality (which, to be honest, I was hoping to
find, not implement), I am thinking about some extensions, such as
"queuing" requests for not-yet-cached pages requested during offline
operation, then "batch-prefetching" when online, and some more ideas
in that direction... interested?

BTW: The classic http://www.gedanken.demon.co.uk/wwwoffle/ has such
offline functionality, but I was hoping to find something in Java to
extend it more quickly for a POC... At
http://www.proxy-offline-browser.com/ there is something like that
too, but commercial, without sources. The
http://www.almaden.ibm.com/cs/wbi/ could probably also be used as
basis for stuff like this, but I haven't looked more closely yet
mostly because of the license. Maybe somebody here has seen other
things in this direction?

Regards,
Michael

PS: Some quick feedback on the built-in GUI that may be of interest: I
at first kept trying all sorts of things with http://localhost:9666/
until I realized that it had to be http://MYMACHINENAME:9666 - maybe
the doc could state this more clearly, or even better, instead of the
error message that currently shows up when accessing as localhost it
could say (or even just redirect?!) to use the real hostname? Also,
minor really, the LogRotator link on top goes "Couldnt find
class:rabbit.meta.LogRotator, java.lang.ClassNotFoundException:
rabbit.meta.LogRotator" and Config says "File 'config\index.html' not
found." (All this was on Rabbit 3.0.)

Re: [Rabbit-proxy-development] Caching for offline browsing

[Samat J. <li...@sa...>]

The way I believe Rabbit is meant to be used is in _front_ of a dialup
connection, not behind it. That is, Rabbit would run on some fast server
somewhere, on a fast connection, at your ISP or a colocated server
(which is what I do).

Rabbit does its magic on this remote server, which then gets sent over
your dialup link, hopefully loading much faster in the process.

So, given that... I don't see how offline caching is useful? I think you
want to look at something which would be run locally, like Squid.

Samat

Michael Vorburger wrote:
> Hello,
>
> I was looking around for a proxy server implemented in Java, as basis
> for an experimental project. RabbIT seems to be the most actively
> developed, as far as I could find at least -- so first of all,
> congratulations on a great project! How are you all using it? As a
> "classic" proxy, like Squid etc.? I am interested in use it for
> building a proxy server that would run locally on workstations, mostly
> laptops that are frequently offline (network disconnected).
>
> How could we get RabbIT to cache all visited pages for this purpose,
> even if the server (or upstream proxy) say they can/should not?? 
> Based on that base functionality (which, to be honest, I was hoping to
> find, not implement), I am thinking about some extensions, such as
> "queuing" requests for not-yet-cached pages requested during offline
> operation, then "batch-prefetching" when online, and some more ideas
> in that direction... interested?
>
> BTW: The classic http://www.gedanken.demon.co.uk/wwwoffle/ has such
> offline functionality, but I was hoping to find something in Java to
> extend it more quickly for a POC... At
> http://www.proxy-offline-browser.com/ there is something like that
> too, but commercial, without sources. The
> http://www.almaden.ibm.com/cs/wbi/ could probably also be used as
> basis for stuff like this, but I haven't looked more closely yet
> mostly because of the license. Maybe somebody here has seen other
> things in this direction?
>
> Regards,
> Michael
>
> PS: Some quick feedback on the built-in GUI that may be of interest: I
> at first kept trying all sorts of things with http://localhost:9666/
> until I realized that it had to be http://MYMACHINENAME:9666 - maybe
> the doc could state this more clearly, or even better, instead of the
> error message that currently shows up when accessing as localhost it
> could say (or even just redirect?!) to use the real hostname? Also,
> minor really, the LogRotator link on top goes "Couldnt find
> class:rabbit.meta.LogRotator, java.lang.ClassNotFoundException:
> rabbit.meta.LogRotator" and Config says "File 'config\index.html' not
> found." (All this was on Rabbit 3.0.)
>   
--
Samat Jain <http://www.samat.org/>

[Rabbit-proxy-development] Re: [Rabbit-proxy-users] Caching for offline browsing

[Robert O. <ro...@kh...>]

Michael Vorburger wrote:
> How could we get RabbIT to cache all visited pages for this purpose,
> even if the server (or upstream proxy) say they can/should not?? 

Write a filter and install it first in httpinfilters and httpoutfilters,
have that filter remove/modify all expires-headers (and possibly other 
headers as well). Check the interface HttpFilter and implement it, then 
configure your new filter in and see it work. Note order of filter is 
important.  Maybe you want to check what wwwoffle does.

As stated, doing this is not in the original goal of rabbit. Note 
however that I want rabbit to be modular and make it easy to implement 
other proxy solutions with it.

> Based on that base functionality (which, to be honest, I was hoping to
> find, not implement), 

I guess you will have to implement it, the coding part is easy, the hard 
part will be to test it.

> I am thinking about some extensions, such as
> "queuing" requests for not-yet-cached pages requested during offline
> operation, then "batch-prefetching" when online, and some more ideas
> in that direction... interested?

Not sure how you want them to work, but of course interested.

 > I at first kept trying all sorts of things with http://localhost:9666/
 > until I realized that it had to be http://MYMACHINENAME:9666

A well know request. Some day I will fix that.

> minor really, the LogRotator link on top goes "Couldnt find
> class:rabbit.meta.LogRotator, java.lang.ClassNotFoundException:
> rabbit.meta.LogRotator" 

Fixed in 3.1

> and Config says "File 'config\index.html' not
> found." (All this was on Rabbit 3.0.)

Removed in 3.1, the config part is not as easy to get working so it is
removed for now.

/robo