kicad-developers team mailing list archive

["Lorenzo" <lomarcan@...>]

--- In kicad-devel@xxxxxxxxxxxxxxx, Dick Hollenbeck <dick@...> wrote:

> Cross probing does not work in Kicad if you have more than one schematic 
> or board open at the same time. There is conflict because the server end 
> of the connection is the same port in each process, and then you get a 
> collision on the port number when the subsequent servers try and open 
> that port number.

I can't agree anymore with that :D

> DBUS is far superior to XMLRPC. It is a binary marshalling 
> architecture, with API query capability. It is not overkill, it is as 
> simple as possible, but not simpler, and Einstein would have found it to 
> be the right mix of capability and complexity.

It requires an external daemom, that's for me it's overkill :P

Give me an in-process DBUS equivalent and I would agree with you.

> DCOP failed because DBUS succeeded and won the battle.

And because it is a big heap of junk :D (like most of KDE but I'm digressing...)

> It is everywhere in a contemporary linux distribution, dig deeper.

But only on linux and anyway, I never even installed it so it isn't maybe so pervasive as you think (said that I've got a lot of box without even libX11 so maybe my machines are not a representative sample)

> The geda guys are either using it or are poised to.  

No idea about that..

> If you want to write a *competing* library access plugin that uses 
> XMLRPC behind the scenes, and as a way to pull parts off of a website, 

Nonono I was only suggesting another kind of IPC communications. But if it's only for cross probing I still think that a system message bus is too much.