[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: non-free firmware and d-i



On Mon, Sep 18, 2006 at 12:36:00PM -0400, Nathanael Nerode wrote:
> Wouter Verhelst wrote:
> > Can people please have a look and comment? I'd like to avoid some
> > specific GR by making whatever is going to be voted upon moot ;-)
> 
> We should have a way of doing the same thing, only with a custom network
> source.  Probably not a hard thing to do.
> 
> Otherwise, looks fine.
> 
> But you may be reinventing bits of d-i.

Yes; I noticed that too, in the mean time.

> Perhaps the correct structure is this:  after choosing "Customization
> modules", the user gets a menu of possible places to get them from (CD,
> USB, floppy, net, etc.).

Chickening out of the hard work, we could just limit this list to
CD-ROM, floppy, or network. The list of retrievers is limited to that
anyway.

Main problem is going to be the fact that currently, every installation
image contains just one or two retrievers, not more. As a result, you
can't download modules from the network when doing a CD-ROM based
install, nor can you install modules from CD-ROM when doing a
network-based install.

Personally, this is the only feature that boot-floppies had while d-i
doesn't that I would *love* to see again. It'd be great to be able to
mount a CD or DVD-ROM in a machine, run apache off of it, copy the
initrd images to a TFTP server, and install a bunch of machines over the
network. With boot-floppies, this was possible; with d-i, it isn't.

But I digress.

[...]
> However, your implementation has the advantage of being simpler and less
> dependent on complicated setup of the customization discs.

Not really. I always saw the CD-ROM stuff as intended for manufacturers
who want to support Debian installations with their official driver
CD-ROM; it's not unreasonable to request some more from them than from
J. Random User who downloads udebs, puts them on a floppy, crosses his
fingers, and boots the installer.

> Let's prove that Joey's six month estimate is wrong by doing it in less time
> even though we don't understand d-i very well.  :-)

I'm not so sure anymore his estimate is wrong; but, hey, let's try.

-- 
Fun will now commence
  -- Seven Of Nine, "Ashes to Ashes", stardate 53679.4



Reply to: