Re: [DotGNU]pnetlib

[Gerard Toonstra]

On Sun, 2002-03-17 at 18:38, Gopal.V wrote:
> If memory serves me right, Gerard Toonstra wrote:
> > Looks like you could use a hand with the pnetlib that is being
> > developed, so I started on that today. I've submitted the first patch
> > for that as well.
>       First proof that the meet-a-thon was successful. Waiting for more
> (/me marks a new notch on my box ;)

SidNL offered to help out with sockets for the engine (Platform_ blah
methods). Says he'll probably need some support, but I've agreed to help
out as best as possible... And I've just received a file that will
implement the SocketMethods interface for me in C#. All that is required
next is implementation of Platform_ methods. I've stressed out that it
is important to consult either t3rm or rhysw to get that implementation
right, as it will eventually be at the communication core of
dotGNU!!!... Hence we'll be issueing a proposal for implementation in
the form of sources that they can both review. They just have to wait
for the email to arrive in their inbox with the sources included :).

>       
>       Helping you help yourself is our job !. And I'll provide help and
> will work on the runtime calls if you need them in short notice.

No short notice there required... I should be ok just developing on a
declared interface.

>       yes, these classes are the easiest to start on -- just imagine
> that Sockets work. Also ECMA spec is your best friend. I think we'll
> publish a csdoc2html'd file and link it to the status page to ease up
> your work. Also let me know when you finish each method. I'll mark it
> on the status page (from now on).

It may be true, but there are certain things _not_ documented there. One
specific question relates to URI's. In the docs it mentions that file
references are converted to the format of the current platform. How am I
going to discover and cater for that. Is there already (or will there
be) an internal call for that?

Another issue related to WebRequest and registration of Prefixes. There
is no containment field or property in WebRequest that can store this
information. It is required to create a new WebRequest type, which is
pretty important to get right because of protocol issues :).

> 
>       Rhys: anyway we can generate the status page from the sources ?
> 

So far I should be ok with that. It is already a little bit outdated, so
for the moment I just cvs-update and look at all the TODO things that I
find.

At the moment I'll possibly have to look into Uri, is someone still
actively working on that?

Cheers,

Gerard          <=      can hardly wait to be able to compile pnetlib
                        and crease the bugs out