Re: [RFC] gvfs connected servers

[Alexander Larsson <alexl redhat com>]

On Mon, 2008-10-06 at 14:38 -0400, David Zeuthen wrote:
> On Mon, 2008-10-06 at 19:48 +0200, Alexander Larsson wrote:
> > >  2. Not sure how if the various path bars (in the file manager and
> > the
> > >     file chooser) would stop working; with this setup we'd have mounts
> > >     inside mounts, e.g.
> > > 
> > >     [[icon] sftp mount at quad.local] [home] [[icon] Home dir on quad.local]
> > 
> > Yeah, but is this bad? In the shared-mount model you get no icon at all
> > for the "home dir" volume, other than the link to the activation
> > location. I'd argue that this is less useful for users.
> 
> I agree; I think multiple icons are desirable.
> 
> > On the other hand, I can imagine various pieces of code inside gvfs
> > breaking when there are mounts inside mounts... So I'm not 100% sure
> > this is a good idea either...
> 
> Actually we _need_ to support mounts inside mounts happening; for
> example consider

Yeah, this works fine. I'm talking about gvfs mounts. The general
problem is that the mapping from uri to mountspec needs to be unique,
otherwise you don't know which mount to talk to.

Hmmm, on the other hand, i guess that code won't see these shadow mounts
really so there might not be a problem. They just exist in the volume
monitor, the actual i/o code paths never see them.

> > That brings us back to multiple
> > volumes per mount then I guess. I'll have a look at your patches
> > tomorrow.
> 
> I'm actually pretty close to finishing a patchset for shadow mounts [1],
> e.g. the in-process proxy volume monitor will automagically create and
> remove GProxyShadowMount's for volumes with an activation_uri according
> to if there's a real mount for the activation_uri. E.g. we'll get

Sounds good.