On Sun, 07 Oct 2001, Ian Clarke wrote:

> On Sun, Oct 07, 2001 at 08:20:41PM -0500, thelema wrote:
> > > This will replace the current mechanism for obtaining a node's reference
> > > which is to use the -x or --export command line option.
> > > 
> > I think that the node's reference should be written to disk ad myref.ref
> > or something when the node is automatically created.  I don't think it's
> > necessary (or that useful) to have this command.
> 
> It certainly isn't nescessary, but I think it is useful, since it allows
> a client to find out the node's reference easily, without having to know
> where the node was installed.
> 
Why would the client want the noderef?  the reason you'd want a noderef
would be to send it to someone, for seeding purposes, and an
automatically created file would suffice quite nicely for that.

> > I'm *VERY* for this functionality.  This is something I've wanted for a
> > long time.  Instead of making it part of FCP, it'd be quite acceptable
> > to have it be a command-line option where you pass the reference file,
> > but I could code it up in liber2.pl to be able to do this, if we do it
> > over FCP.
> 
> Using command line arguments for this type of thing is ugly, which is
> why I want to move away from the "-x" functionality.  Nodes should be
> left running, with FCP used to extract information and perform other
> useful functions in the manner I have described.
> 
I agree the node should be left up, but you don't have to take down the
node to use such a command-line parameter.  You don't even have to start
a new one, just make the connection, attempt to auth, and then send the
"disconnect" message.

> > I might revise this so that instead of returning
> > nodes whose CP=1, you could have a paremeter to the request to specify
> > that you want references whose CP is greater than that number.
> 
> I did consider that, and wouldn't object to it, although I suspect most
> people would set that configuration option to 1 anyway.
> 
A default of 1 would seem to fit the situation.

> > Related to this would be a command that just dumps out (in some nice
> > format) some info about each node (their contact info or whatever) and
> > their CP, so we'd be able to look at the CPs for all neighbors.  That'd
> > be useful information for many reasons, and it'd have to be a FCP
> > command.
> 
> I did think about that too, but that would prevent the output from being
> directly pipable into a seedNodes file, as an option it might be useful,
> but I wouldn't make it a priority.  Lets worry about feature creep after
> the basic functionality is implemented.
> 
I just want it as a diagnostics tool for looking at how my node is doing
in terms of how successful it's been lately in contact other nodes.
This CP-dump would definitely be seperate, and in addition to, the
references dump for seeding purposes.

> > > More ambitiously, by allowing FCP connections from a specific remote
> > > host, that host could regularly harvest references from a variety of
> > > volunteer servers around the net, and even use those servers in a
> > > distributed way to test references using TestReference.
> > Interesting.  And possibly very useful.  I like the idea.  We just need
> > someone with enough skillz to code it up.
> 
> It would be simple to achieve once the functionality is in FCP, and that
> should be pretty straight-forward too.
> 
> Ian.

Yes, it'd be simple and straight-forward, but we're all lazy here.
Someone's going to have to code it, and I'm not volunteering.

Eric
-- 
E-mail: thelema314 at bigfoot.com        If you love something, set it free.
GPG 1536g/B9C5D1F7 fpr:075A A3F7 F70B 1397 345D  A67E 70AA 820B A806 F95D
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20011008/98db9e5f/attachment.pgp>

Reply via email to