Re: Read-write tarfs

[Joshua Judson Rosen]

On Wed, Jan 22, 2003 at 08:13:03AM -0800, James Morrison wrote:
> 
> --- "Alfred M. Szmidt" <ams@kemisten.nu> wrote:
> >    Sure. ;) But I think you missed the point: The `--volatile' options
> >    makes newly created nodes "volatile".  That is, they won't get
> >    synced to the tarball when asking the translator to go away. :)
> > 
> > Ponder, what is the point of removing the object files; isn't that a
> > bit bakwards?  Removing debugging symbols is probobly as stupid...

Well, I suppose they aren't -removed- so much as just -not written in
the first place-, yes? Er..., something like that....

>  No, it's not backwards.  It makes perfect sense.  Removing debugging symbols
> also makes sense for most binaries.  I think the volatile feature is really
> cool.  No more untarring stuff onto the filesystem when it can all stay in
> RAM :)

Hm. Maybe it'd be nice to put this buffering feature into its own
translator--a `bufferfs', to layer over other translators.

I remember seeing some nice things done with totally-buffered
file-systems--SuSE, for example, having `live trial' CDs that could be
loaded and run-with, and the whole file-system would be writable, but
it wouldn't go anywhere except RAM. No hard-disk needed, even if one
wanted to see what it was like to hack on things. It might be nice to
produce something like that :)

-- 
"In my experience, there's no such thing as a software design that's
 correct from the start.  Redesign should be built into the process,
 because it *will* be required." --Bill Gribble, guile-user

pgp3lLUwLE78l.pgp
Description: PGP signature