[Pkg-uml-devel] Preparing upload for user-mode-linux (and finishing up rootstrap)

[Mattia Dongili]

On Thu, Jun 08, 2006 at 09:27:22AM +0200, Stefano Melchior wrote:
> On Wed, May 31, 2006 at 09:08:14PM +0200, Mattia Dongili wrote:
[...]
> One more thing: I would like to understand the logic of improving
> u-m-l-x.x.x-1um -> 2um (it is not a critic): should it be when a
> "big"/"important" or whatever change is introduced? or just to avoid the
> epoching?

it was just to repackage the orig.tar.gz that previously contained the
debian/ directory (see r108 on SVN).

> > [...]
> > > > rootstrap:
> > > > - what do we gain removing the "fd:1" option from uml command line?
> > > >   Stefano, I tried your test cases and rootstrap doesn't seem to work
> > > >   differently with or without it.
> as discuss privately with Mattia (for other reason) I need to work with
> rootstrap remotely from office to my domestic server. I am connected home
> via a ssh connection; if I use "con0=fd:0,fd:1" and "con=pts" I can not see 
> anyting on the window and thus I need to manually modifie rootstrap
> script in order to have only "con0=fd:0" available as umlarg.

such options only mean that con0 will use stdin/stdout, all others will
use /dev/pts/X instead. rootstrap invocation of the UML kernel will only
use con0.
Note for example:
# ls -l /proc/$(pidof -s linux)/fd/[01]
lrwx------ 1 mattia mattia 64 2006-06-08 19:20 /proc/5231/fd/0 -> /dev/pts/4
lrwx------ 1 mattia mattia 64 2006-06-08 19:20 /proc/5231/fd/1 -> /dev/pts/4

And I can't see any difference if running UML with fd:0,fd:1 or simply
fd:0 either via an ssh connection to a headless server or other fancy
setups. Today I created a root fs image via putty too. I'm quite inclined
to think the problem lays somewhere else.
I'm debugging UML right now as it seems to do something weird anyway.
Could you try to build UML from a recent kernel and see if you still
have this problem?

> > > > kernel-patch-skas:
> > > > - please see #368539. I'm inclined to remove the "localversion-skas"
> > > >   hunk from the diffs until the bug is fixed. It's a very simple change
> > > >   but i'm wondering if I'm the only one seeing this problem.
> I agree with you.

So, am I the only one being unable to build the kernel using 
	make-kpkg  --added-patches skas ...
with the package I uploaded on people.d.o?

-- 
mattia
:wq!