Re: [dark: READ!] CD images / Re: Formal objection: Changing how the testing of potato works would invalidate the whole test. So please don't change it.
On Mon, 15 May 2000, Richard Braakman wrote:
> On Mon, May 15, 2000 at 04:49:54PM +0200, J.A. Bezemer wrote:
> > Okay. Will a developer-signed "fake" .changes file with md5sum of the
> > upgrade-all.tar do? Does it have to be placed in Incoming (on ftp-master I
> > guess) or can you just fetch it from my website?
> Yes, that will do. I can fetch it, but that may be slow. It's best
> if it's already on the machine.
Okay, I'll see what I can do. I'll tell you when it's done; you can then use
`wget' or the like to get it directly to ftp-master, panic.et.tudelft.nl gets
>200kByte/s to about anywhere on the world (10Mbit ethernet directly to the
TUDelft ATM backbone).
> > apt 0.3.19 is something that I don't really know of, but it is supposed to fix
> > very annoying deadlocks that occur only when installing/upgrading from CDs.
> > AFAIK there's nothing on the boot floppies that can be broken by a new apt.
> Boot-floppies contain the apt binary. They will still use the old
> binary for the initial install. Installing the new apt means that
> the source for the one on the boot-floppies goes missing, and also means
> a new round of autocompiling.
I see now. That simple little explanations are often very enlightening ;-)
> > But if you are not going to install this (Please tell me!) than I can arrange
> > for a README on the CDs that points out the problem and where a new apt is
> > available.
> I am not going to install it for this test cycle.
Now this is doing business! My faith is high and rising again ;-)