[Top][All Lists]

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

Re: [Fwd: Questions on isc-dhcp]

From: Samuel Thibault
Subject: Re: [Fwd: Questions on isc-dhcp]
Date: Wed, 2 Mar 2011 00:50:00 +0100
User-agent: Mutt/1.5.12-2006-07-14


Svante Signell, le Tue 01 Mar 2011 20:50:33 +0100, a écrit :
> On Tue, 2011-03-01 at 09:03 +0100, Samuel Thibault wrote:
> > Svante Signell, le Tue 01 Mar 2011 08:50:39 +0100, a écrit :
> > > I have now working scripts dhclient-script.hurd and 
> > > dhclient-script.hurd.udeb
> > 
> > Ok, good, please post them for review.
> Attached below, with debug output by adding set -x to the scripts. 

Once again, thanks for your work, it'll be very useful!

Here are some modified version, just a couple of comments:

- I've just commented the aliases part: better do nothing that error out
- I have replaced --down with -A Actually, --down was the
  equivalent of Linux' ifconfig down, which got replaced (on my request,
  actually) by ifconfig 0, to just disable ipv4 without shutting down
  the interface.
- route add can be implemented by using
  fsysopts /servers/socket/2 $(showtrans /servers/socket/2) -g $router
  as this simply appends the -g option to the existing ones. This
  however won't handle multiple interfaces well, but that should be
  enough for now.
- I've made a few changes to stick even more to the Linux version.
- About -a vs -A vs --address: don't mix pfinet/fsysopts options with
  inetutils-ifconfig options: although they happen to be very similar,
  they are not the same, that's why -A wasn't working for fsysopts.

> One thing remains to do: The file
> /etc/dhcp/dhclient-exit-hooks.d/rfc3442-classless-routes should not be
> run on GNU/Hurd (or should be adapted to not use /sbin/route). Now it is
> called but does nothing.

Ok, that'll be fine enough for now.

> It does not make sense to use construct like, created by
> using ptom (softlink to netcfg) since inetutils does not support this
> synstax. (I downloaded the udebs netcfg_1.60,

Ah, ptom does that, ok, agreed indeed.

> Of course the scripts should be tested in other environments than under
> qemu like I did. And, the udebs within the Debian installer of course.

Yep, I'll do that once you post your patch against the .c code.

> I've spent enough effort on this for now. I'll send a patch as a bug
> report as mentioned earlier. Please test the patches and review the
> modified scripts.

Please post on this list first, I'll upload a package on debian-ports so
people can test it easily before you file the patch to the bug tracker.


Attachment: dhclient-script.hurd
Description: Text document

Attachment: dhclient-script.hurd.udeb
Description: Text document

reply via email to

[Prev in Thread] Current Thread [Next in Thread]