Fwd: Whats up with WebJDK

[Akash Shetye]

Thanks Roman, I actually don't mind writing the java.io.FileSystem
abstraction that you can't find now. Since you have achieved that earlier,
you can tell me how to go about it. Thanks for the reply.

On Tue, Jan 15, 2013 at 3:17 PM, Roman Kennke <roman at kennke.org> wrote:

> Hi Akesh,
>
> As I posted earlier, I made a filesystem prototype ~1 year ago,
> unfortunately I cannot find the sources anymore (and my backup HD
> crashed recently..) the only chance to recover this is that I probably
> sent the code to somebody by email .. in any case, here is what I
> remember from it:
>
> > Thanks Roman, we will need a lot of help. Currently I am going through
> > Caciocavallo's code and trying to get a bigger picture (not there
> > yet :P). To develop a new Redirected filesystem I gather that, we will
> > need to reimplement the java.io's File, Writer abstract class as well
> > as reader abstract class and probably more.
>
> As for the java.io.File, there already is a reasonable abstraction in
> place, it's the (package-private) java.io.FileSystem. You also need
> implementations for FileInputStream and FileOutputStream. Unfortunately,
> those call directly into native code, and our goal is to not call into
> native code, but implement some sort of protocol to HTML5/JS client.
> Therefore it'd be necessary/useful to change the file-stream classes to
> go through a (thin) abstraction layer, and implement that. (I hope I can
> recover my prototype from somewhere, it did implement all that.)
>
> >  I am thinking of using HTML5's File API for reading files and HTML5's
> > FileSystem API for writing down files to the client system, however
> > the writing happens to a virtual sandboxed file system, so wouldn't
> > have the flexibility of actually writing things down unless we use
> > ActiveX, Java Applets etc... Tell me what you think about it.
>
> Yeah, that's about it. My idea back then was to make it so that several
> different implementations are possible on the client side, and provide
> one (sandboxed) HTML5/JS based one and one that uses applets for full
> client filesystem access.
>
> I hope that helps!
>
> If I manage to recover the prototype, I'll let you know, then you
> probably don't need to reinvent the wheel ;-)
>
> Best regards,
> Roman
>
> >
> > On Wed, Jan 9, 2013 at 2:38 PM, Roman Kennke <roman at kennke.org> wrote:
> >         Hi Akash,
> >
> >         > Hey all on the dev team sorry for being off topic,
> >
> >
> >         It's not off-topic! Infact, it is right *on* topic here :-)
> >
> >         >  I tried Caciocavallo not long ago and found it amazing. I
> >         was
> >         > interested in knowing how the development for the WebJDK is
> >         going on
> >         > because I had planned to pursue a similar effort myself,
> >         later
> >         > stumbling across caciocavallo. Instead of reinventing the
> >         wheel I
> >         > wanted to know if I along with a few friends could help with
> >         any of
> >         > the WebJDK's subsystems like filesystem etc.
> >
> >
> >         WebJDK basically is Cacio-Web plus a few ideas. A while ago I
> >         started
> >         work on a filesystem prototype, but it never went very far.
> >         From there
> >         it basically stalled due to a lack of resources.
> >
> >         > The goal of doing this is to come up with a bachelor
> >         thesis/project.
> >         > We want to do something that has a positive impact and is
> >         opensource.
> >         > Having about four months of nothingness in our schedules, it
> >         would be
> >         > great if we could contribute substantially on the WebJDK
> >         thing. I
> >         > think the exercise will prove to be a great learning
> >         experience for us
> >         > above all.
> >
> >
> >         Absolutely! You would be very welcome in this effort. Just let
> >         us know
> >         if you need any help or more information to get started.
> >
> >         Cheers,
> >         Roman
> >
> >
> >
> >
> >
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/caciocavallo-dev/attachments/20130115/5704f4fe/attachment.html