D-BUS in KDE

[Mike Hearn]

On Sat, 02 Oct 2004 18:53:15 -0400, Havoc Pennington wrote:
> By "optional/ugly" what I mean is something like, right now we have
> new_reply() for sending a method call reply, we could add a 
> new_reply_with_call_serial() that took the unique number.
> 
> The optional/ugly aspect would not impact bindings such as the Qt
> bindings since they could probably handle it automatically.
> The whole low-level libdbus is ugly anyhow so it's sort of irrelevant.

What's wrong with using a TLS slot, ie

1) client makes outgoing RPC, new call serial is generated,
   TLS slot is set to the call serial (it's initialized to zero)
2) server accepts RPC, sets call serial TLS from packet and dispatches
3) server makes another RPC back to client, however TLS slot isn't
   zero so it simply copies it into the packet
4) client accepts packet, note that packet.serial == tls slot, and
   re-enters.

Is there some problem with this scheme I haven't spotted? It would allow
you to implement it with no API changes, AFAICT.

thanks -mike