Re: review of the Storage interface

[Dietmar Maurer <dietmar maurer-it com>]

Maciej Stachowiak wrote:

> However, it's not even true that we can make any of these changes
> without affecting anything else. For instance, Nautilus has a server
> that implements the Stream interface to use with Bonobo Controls and
> Embeddables which we use to display content in Nautilus. We'd have to
> make a lot of changes to that code to support all the things you want,
> and it would be very difficult to test our changes, since no Bonobo
> components out there use any of this functionality right now (since it
> doesn't exist yet).

I think we should do only the necessary things now:

 - remove eos();
 - extend the directory_list data structure

all other things can be delayed to the next release of bonobo, since they only
add functionality.

> Do you think that all the changes you suggested are important enough
> that we should delay shipping GNOME 1.4 for an unknown but possibly
> quite long length of time to get them in, and even ? Are they also worth
> delaying GNOME 2.0, the
> long-awaited and long-delayed development platform update, since many
> people will not start working on GNOME 2.0 in earnest until GNOME 1.4
> has shipped?

I don't want to delay things. But I think doing the above two changes is not
that much work. And doing it now prevents us from doing it later, when there
is an official release and many programs that depend on the interface.

Your argument "then risk shipping a buggier GNOME 1.4" is right if you think
of last minute changes. But it is not true if you consider the eos() function.
This function is simply not implemented in the efs
storage module, and the fs and vfs implementations are semantic different!

But anyway - my primary interest is the interface itself and not the release
date of GNOME.

Regards,

    Dietmar