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

Re: Base set is coming closer...



On Fri, Aug 07, 1998 at 02:17:44AM +0200, Mark Kettenis wrote:
> 
>    sysvinit (? is there some hurd-ism to care about?),
> 
> the Hurd init is BSD style and has a lot of Hurd-isms.  Is a System V
> like init really necessary?

Well, I'm not sure, because I'm not expertised in this. I think we have to
change the boot script anyway (there are linux specific parts, like modules
and fs mounting).

And Gordon, I think this is somewhat important because Debian packages
integrate scripts in the boot process and shutdown process.

>    update (? will it run on the Hurd? Functionality important else
>    data loss), 
> 
> the traditional one (sync every ... seconds) should work, but I
> believe Hurd filesystem translators do periodic syncing themselves.
> At least there is some code in libdiskfs to do that.  The new Linux
> update uses the `bdflush' syscall which the Hurd does not have.

Great. Gordon already verified this, so we can put "update" to the
"not-needed" list.

I think I'll set up a web page with status information for each package.
What works, what doe snot work and why. So we will have some overview about
this, and can see how the Hurd and Debian GNU/Hurd improve over time.
 
>    So, you can stop/restart daemons etc. How robust are translators during such
>    an upgrade/replacement?
> 
> You can tell translators to "go away" or even force them to "go away"
> (sort if the difference between sending a SIGTERM and a SIGKILL).

Okay, so we only need detailed information about what translators have to go
during what upgrade/install steps. This will need some experience and so,
but should be okay.
 
> You probably have to install `/hurd/exec' first.  Then tell (or force)
> the old translator to go away.  There should be no reason to modify
> the file `/servers/exec' once the passive translator has been set.  I
> think, instead of including a file `/servers/exec' you should do
> 
>    # touch /servers/exec
>    # settrans ...
> 
> (writing this message under Linux so I cannot look right now what the
> exact command would be).

Yes, this would be done in the post install script, and only during a fresh
install. So, would I need to restart the translator when /hurd/exec is
upgraded (with the above procedure, go away, upgrade, settrans...)?
 
>    Or would anyone else be able to
>    provide an arla gateway until Michael is online again?).
> 
> You mean Thomas (who used to be Michael).

I meant Michel "Walken" LESPINASSE, who offered a gateway.

> I discovered that the CVS
> repository is accessable via http.  I guess this means you could use
> something like `wget' to mirror the CVS repository and then use CVS
> locally, but I haven't tried it yet. 
> 
> Have a look at:
> 
> http://www.mit.edu/afs/sipb/project/hurddev/

Yeah, but keeping it current via wget is pretty awkward (although doable). I
think a good solution (in the long run) would be to mirror the CVS
repository for anonymous CVS on the Debian cvs server. We already mirror the
Gnome development and other stuff. So, no load but the mirroring would be
put on mit.edu, anonymous access would be open on cvs.debian.org.

Would that be acceptable?

Thank you,
Marcus

-- 
"Rhubarb is no Egyptian god."        Debian GNU/Linux        finger brinkmd@ 
Marcus Brinkmann                   http://www.debian.org    master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de                        for public  PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/       PGP Key ID 36E7CD09


-- 
To UNSUBSCRIBE, email to debian-hurd-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


Reply to: