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

Re: Kernel-package, fix version 2 ...

Hash: SHA1

On Sun, 23 Oct 2005 08:08:50 +0200
Sven Luther <sven.luther@wanadoo.fr> wrote:

> On Sun, Oct 23, 2005 at 03:00:12AM +0200, Jonas Smedegaard wrote:

> > On Sat, 22 Oct 2005 23:50:10 +0200
> > Sven Luther <sven.luther@wanadoo.fr> wrote:

> Not sure if allowing copa, colon or whitespace delimitation clashes
> with the filenames though, as some may use comas.

Agreed. Stay with the space delimitor. It is backwards compatible (I
notice now), and that is important.

> > > The second issue is not really related to this plan, but we really
> > > want to debconfify the kernel-package kernel scripts, this would
> > > mean replacing all die and printing to stdout with proper debconf
> > > messages, and inputs in the same way, and maybe redesigning some.
> > > I can do this in shell, but never tried the perl solution
> > > previously.
> > 
> > Please do not bloat the postinst code! Separate it into a helper
> > tool instead!
> No way, i think you are misunderstanding here. The current code does :
>   print STDERR <<"EOINFOMSG";
> Using $ramdisk to build the ramdisk.
> Other suitable ramdisk generating tools : @ramdisklist.
> Full list of probed ramdisk generating tools : $ramdiskorig.
>   ;
> And we will replace it with (shell version given) :

[snip - nice and dandy, but irrelevant for ramdisk enhancement!]

You will not introduce debconf, please. You will stick to same coding
style as Manoj uses.

Manoj does not use debconf, so you will not either. Not without consent
from Manoj. That would be outright stupid. Approaches like that helps
explain the tension between you and the initramfs-tools maintainers!

> Furthermore it has been since before sarge policy that all user
> questions need to use debconf, and it was only skipped because Manoj
> didn't want to do those change (probably for lack of time), and
> nobody jumped in. Now is the time though.

Then deal with that separately: File an RC bug about k-p breaking
policy regarding debconf, and do an NMU fixing that issue.

Such issue is separate from the issue of handling new ramdisk querying
interface, so please treat it as such.

I do not want the ramdisk enhancements dropped again by Manoj when he
gets back because you got carried away and "fixed" other frustrating
annoyances concurrently!

> Does that clarify it, and if not, please explain how i could put this
> in a helper tool ? 

this should clarify my stand:

Now: Add support for querying interface of ramdisk tools

Later (option1): Persuade Manoj to use debconf in k-p

Later (option2): Persuade Manoj to use new kernel-install-helper

Later (option3): Leave as is

 - Jonas

- -- 
* Jonas Smedegaard - idealist og Internet-arkitekt
* Tlf.: +45 40843136  Website: http://dr.jones.dk/

 - Enden er nær: http://www.shibumi.org/eoti.htm
Version: GnuPG v1.4.2 (GNU/Linux)


Reply to: