Re: [Bcusdk-list] eibd fork -> knxd

[Timo <bc...@ti...>]

yes, thanks for the summary. That's exact how I see this.

The most important thing is: Current version is stable. It is a
production quality software.

The main goals should be:
    - every release must be stable
    - there must be no regressions

At least for the core (daemon + client library) we should not simply
start changing code of the core components. Most important is some
system to assure the software quality. Changes to the build system, the
tools, packaging, language bindings, documentation, ... are no Problem.
But for the core I think changes should be reviewed before going to git.
Before changing "the core" we should think about automated testing.


Am 18.12.2014 um 11:16 schrieb SZENTE Balint:
> Hi,
>
>
> Many thanks for you guys keeping this project alive and many thanks for
> original authors for this great piece of software.
>
> eibd is a stable and production quality software. You can rely on it.
>
> I'm also on the same opinion that C(++) is the best language for such a
> project, for several reasons:
>  * embedded developers (the main target) are usually experienced with this
> language, large developer base
>  * very few dependencies
>  * you can put on any small computer, virtually any OS
>  * well proven
>  * resulting code is fast
>  * the possibility to debug remote with a simple console (ssh/telnet+gdb)
>  * no hidden "black magic" introduced with various software stacks and virtual
> machines which makes debugging a hell if/when the platform itself has bugs ->
> C is basically "what you see is what you get"
>
> So yes, please keep it plain simple, cross platform and rock solid. Of course
> all these arguments are for the daemon, I'm not talking about the client
> bindings.
>