Daemon or no daemon

[Magnus Bergman]

On Mon, 18 Apr 2005 10:40:06 -0400
Sean Middleditch <elanthis at awesomeplay.com> wrote:

> On Mon, 2005-04-18 at 16:52 +0200, Magnus Bergman wrote:
> > Some back-ends doesn't need the daemon, or can not benefit from it.
In
> 
> This proposal is identical to my original D-VFS proposal, which was
shot
> down for several good technical reasons.  There is no backend that can
> fulfill the D-VFS behavior requirements without a daemon.  Please read
> the wiki and list archives for more details; I'm not going to repeat
old
> discussions anymore.

Yes, I did read the wiki and I read it again more carefully. And I
understand and agree with almost all of it. But there is one detail I
don't understand: "The complexity in making backends support both
in-process and daemon operation would severely complicate the
development of a backend." Why should that be so complicated? If it
means that the back-ends would contain some hacks the make up for the
missing daemon, then of course it becomes complicated. But that's not
what I had in mind. If the back-end needs the daemon it could just
appear crippled (only supporting basic features not relying on the
daemon) or refuse to load if it was loaded in-process. While other
back-ends such as the posix one doesn't care if it's loaded by the
daemon or in-process.