Oskar Sandberg <md98-osa at nada.kth.se> wrote:
> The problem is of even bigger concern if we use several transports at
> once, like Brandon wanted to attempt. In that case I don't see any way
> around it but making it possible to read the type within the protocol
> (because you want to be able to copy fieldsets from one transport to
> another, possibly with completely different ways of representing values).
> 
> If we are against quoting strings, we could add a mark at beginning of
> fields that contain values instead (Since true and false can't be read as
> hex strings) and have the same result. Of course, then strings would not
> be allowed to start with this, but it could be escaped (as we have to
> escape newline as it is).

In that case, maybe it would be better just to take another byte for a type
indicator.  For example, S = string, N = number, B = boolean.  Then we
wouldn't have to perform escaping and it would be more flexible if we
eventually wanted to add new types.  e.g.

Field1=Ssomestring
Field2=N3a372f83
Field3=Btrue

theo


_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev

Reply via email to