Re: No LVM in Woody (Was: writing a release announcement)
On Wed, May 01, 2002 at 11:04:59AM +0200, Robert van der Meulen wrote:
>
> Quoting Patrick Caulfield (patrick@debian.org):
> > Please /don't/ package rc2 (even for unstable) - it's a can of worms. leave it
> > as it is until 1.1 is properly released (or maybe even revert to the upcoming
> > 1.0.4)
>
> Can you point out some of the worms ?
Certainly. Heinz has completely re-written the internal memory management to try
to make the single-binary stuff work. All of this is new code, not heavily
tested. In fact when I did do some mild testing it all fell over.
There are also some new feature flags that call into the kernel for testing. He
hasn't incremented the IOP for this - I think he's relying in the ioctl failing
for the status; so, this relies on code that will never get into the stock Linux
kernel.
In fact there is so much new code in 1.1rc2 compared with 1.1rc1 that it really
should be 1.2rc1 :-(
We've been arguing over this internally at Sistina and the opinion is that
1.0.4 (out soon) should be the version that everyone is using. Whether we'll
kill the 1.1 branch is still under debate but it is my (not Sistina's) opinion
that it should never have happened in the first place. We are doing a lot of
work on a new LVM2 which we want to release in the near(ish) future. It's all
new code based on a small "device-mapper" kernel core. Andres Salomon has
already started packaging this for Debian - and has write access to our CVS.
patrick
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: