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

On installation media and firmware [Re: New section for firmware]

Switching from -vote to -boot and -cd.

On Wed, 24 Dec 2008, Frans Pop wrote:
> On Wednesday 24 December 2008, Don Armstrong wrote:
> > 1: -vote really is the wrong list to discuss this on; Cc'ing debian-cd
> > so knowledgeable people there can tell me I'm on crack.
> Second of all, the analysis for the impact on installations of
> firmware and why the D-I team would very much prefer to have any
> firmware included both on official CD images and in other D-I images
> (such as netboot, floppy and hd-media) has already been done and
> posted by Joey Hess ages ago and discussed repeatedly, both on the
> debian-boot list and on d-vote, and I'm quite certain also on
> d-project and/or d-devel.

I don't think there's any argument against having firmware in a
limited set of CD images and other d-i media which we distribute like
we distribute non-free. Furthermore, the major portion of what I
suggested doing has not been discussed recently on -project or -devel

> It's a pity that we seem to need to explain this over and over again
> in random threads.

It would be nice if you (or someone else) would then explain[1] why
having a limited set of d-i images + cd images which contained the
non-free firmware necessary for certain hardware to boot would be an
intractable solution. From where I sit, the major issues with it would

 1. Additional testing required

 2. Increase in cdimage mirror utilization

 3. Work to actual implement it

I submit that the additional testing for #1 will be required when the
firmware is placed on the images, regardless of whether we produce
images not containing it, so the major burden of testing here seems to
be to check that the firmware-related codepaths are only triggered
when the firmware is there.

#2 should be on the order of a few gigabytes, which is dwarfed by the
increase in the size of the installation sets from version to version.
[If our mirror operators will have problems with this, we've got far
more serious ones.]

#3 is what it is.

Don Armstrong

1: You dealt with problems with having an entire set duplicated and a
set of ancilliary issues, none of which I proposed.
This message brought to you by weapons of mass destruction related
program activities, and the letter G.

http://www.donarmstrong.com              http://rzlab.ucr.edu

Reply to: