On Tue, Nov 27, 2001 at 08:07:30PM -0500, Tavin Cole wrote:
> On Wed, Nov 28, 2001 at 01:41:17AM +0100, Oskar Sandberg wrote:
< > 
> > Can we use some matching of the keytype bytes instead of the length of
> > the keys to determine this? It is not ok to assume that that keys will
> > always have any length. Any distribution of the keys should be based
> > only on the first single or half byte anyways.
> 
> I just mentioned the current 23-byte length by way of explanation, not
> as a rule to depend on.  The right rule would be to have fake keys end
> in 0000, which iirc is true for those resulting from a manual --seed.
> I think it must be your announcing code that is putting the 16-byte
> fake keys in there that don't end in 0000 - ?

Yes, I think it is. Looking over the code, I'm pretty sure it won't
break the protocol to add 0000 when the key is generated, but I don't
feel confident enough to commit ATM.

> > Also, wouldn't it be nicer if we got this on the same port as the http
> > access? It really needs some sort of authorization, and so should fit
> > with http based administration access. Using the same visualization
> > stuff for the Diagnostics data would be nice as well (feel free to add
> > more diagnostics fields).
> 
> You can't cleanly put it on the same port as fproxy since fproxy sucks
> up the whole URI, thus defeating the mechanism for running multiple
> servlets on the same port.  So you'd have to hack it into fproxy itself.

Yeah, I know there is a problem with this. What would be preferable is
that if we allowed URIs starting with "?" to not be eaten by fproxy, I
don't suppose this works with servlets...

>

-- 

Oskar Sandberg
oskar at freenetproject.org

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to