RE: Versioning File Systems?

Kerl, John (John.Kerl@Avnet.com)
Thu, 18 Apr 2002 09:51:13 -0700


Is it just me or is this sounding a lot like
ClearCase? In their filesystem (I don't know
if they implement it in user space or kernel
space, but I do remember ClearCase on Solaris
did do some kernel mods), file names are really
directories, e.g. foo.c is current; foo.c/main/3
is a (perhaps different) specified version.

& for recovering from editor screwups, one could
easily imagine "vi foo.c/-3" to recover the file
from 3 saves ago, etc.

By "deducing change sets", is the question, how
to associate various versions of *different* files?
I.e. recovering an editor screw-up of a single
file is easy, but how do you back out that RPM
you just installed, which might have affected
many files? Here ClearCase uses "labels",
which associates *one* name with the specified
versions of many files. So you could set your
"view" (in ClearCase terms) to /tuesday, etc.

When I used ClearCase in prior jobs, I loved
it -- it was a joy *because* it looked like
a plain old filesystem (e.g. vi foo.c) when you
wanted to think of it that way, but it also
had full-featured version control.

Is the idea being discussed to open-source
something of that nature, and make it into
a filesystem?

-----Original Message-----
From: Lars Marowsky-Bree [mailto:lmb@suse.de]
Sent: Thursday, April 18, 2002 8:28 AM
To: linux-kernel@vger.kernel.org
Subject: Re: Versioning File Systems?

On 2002-04-18T08:20:25,
Larry McVoy <lm@bitmover.com> said:

> It's certainly a fun space, file system hacking is always fun. There
> doesn't seem to be a good match between file system operations and
> SCM operations, especially stuff like checkin. write != checkin.
> But you can handle that with

Either that, or heuristics - file not written to / opened for writing in x
minutes -> commit.

That would actually be pretty interesting because it might also allow you to
back out editor screwups ;-)

However, deducing change sets is more difficult.

Sincerely,
Lars Marowsky-Brée <lmb@suse.de>

-- 
Immortality is an adequate definition of high availability for me.
	--- Gregory F. Pfister

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/