Hi,
On Mon, Jun 15, 2009 at 09:01:55PM +0300, Sergiu Ivanov wrote:
> On Sat, Jun 13, 2009 at 03:53:27PM +0200, Carl Fredrik Hammar wrote:
> > On Thu, Jun 11, 2009 at 09:10:24PM +0300, Sergiu Ivanov wrote:
> > > diff --git a/unionmount.c b/unionmount.c
> > > new file mode 100644
> > > index 0000000..e4aa043
> > > --- /dev/null
> > > +++ b/unionmount.c
> >
> > Given that this is to implement the --mount option, I think mount.c
> > would be a better name. The context of unionfs establishes that this
> > means unioning.
> >
> > > @@ -0,0 +1,28 @@
> > > +/* Hurd unionmount
> > > + The core of unionmount functionality.
> >
> > Again the purpose of the file isn't really to implement unionmount,
> > but to implement the --mount option.
>
> The idea is that the ``--mount'' option is just an intermediate step
> towards a ``stand-alone'' unionmount implementation. That's why I call
> this file in a more general way than just ``mount''.
*If* we ultimately go with a standalone unionmount -- which isn't even
certain at this point -- naming it "unionmount.[ch]" would be even more
wrong than now. The files do not implement unionmount; they only
implement one specific part of it.
> I remember antrik objecting to this name, too, but I'd like to point
> out that the information about unionfs itself is declared in
> unionfs.h. This was the reason why I called the two filed
> unionmount.{c,h}.
unionfs.h contains definitions relevant for *all* of unionfs, not just
some specific aspect.
-antrik-
