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

Re: Permission for 'dev/pmu'

On Mon, Jan 10, 2005 at 09:57:46AM +0100, Sjoerd Simons wrote:
> On Mon, Jan 10, 2005 at 09:21:08AM +0100, Sven Luther wrote:
> > On Mon, Jan 10, 2005 at 09:03:33AM +0100, Sjoerd Simons wrote:
> > > On Mon, Jan 10, 2005 at 03:17:22AM +0200, Martin-Éric Racine wrote:
> > > > On Sun, 9 Jan 2005, Christoph Hellwig wrote:
> > > > 
> > > > > On Sun, Jan 09, 2005 at 08:48:39PM +0100, Sven Luther wrote:
> > > > > just kill this warning.  Debian works just fine without backlight
> > > > > control on both my ibook and the g5.  And it's really not worth giving
> > > > > hardware control to an unadited GUI program.
> > > > 
> > > > Having been trying various ways to fix this for a number of days, I'm
> > > > slowly coming to the conclusion that the best option for Debian would be
> > > > to purge all of ACME's Powerbook-specific LCD control code out of the
> > > > way. 
> > > > 
> > > > If upstream ever gets around cleaning up this code to use safer, more
> > > > generic methods, such as doing everything via 'pmud' then we could
> > > > reintroduce it.
> > > > 
> > > > Any objection?
> > > 
> > > Just change the code to a g_warning instead of a dialog and stop whining? I
> > > personally use this code and so do many others, so removing it isn't an
> > > option.. 
> > 
> > Ah, you use it and you didn't come to fixing it ? This has been a problema
> > since >6 month.
> > 
> > I would be very interested in knowing how you did solve the error message
> > issue in a secure and acceptable for inclusion in the release way.
> The error message won't appear if your user has access to the pmu, which is the
> case on my system. But that's ofcourse wrong to do by default.

What do you respond to Christoph's critic, which was if i remember well : 

  "no unaudited gnome app should have such direct access to my hardware".

And the evident unsecurity of this process ? 

And still, that doesn't fix the main issue is that a clean new install will
result in this obnoxious error message showing up, with no apparent way to
solve it, nor an idea of where it comes from. At least the message should
provide instructions, or be removed completely, if this cannot be fixed, and i
have some doubt that this can be solved in a clean way in the sarge timeframe.

I hear that a consensus seems to be forming up that pbuttonsd offer this same
functionality, and thus acme doesn't really need to be doing it, and i am much
more inclined to get it removed completely, since the current situation is
buggy (RC buggy i would say, at least as far as the feature is concerned), and
there is no real incentive on the part of the people using the feature to fix
the issue, while not caring about it being utterly broken for other folk, and
it being a situation which would reflect badly on debian on powerpc if we
release with it.

> As said just change the current pmu patch in the package to do g_warning 
> instead of acme_error, which is trivial. No need to making a lot of fuss about

So let's do it ? Do you volunteer on writing the code ? I would rather
continue preparing the 2.6.10 powerpc kernel upload.

> it on several mailinglists, there is a reason we have a bug tracking system.

Well, a BTS is fine if people act on the bug, isn't it ? I wrote to
debian-gtk-gnome, because i was searching for input on a group which may be
used and which is modified dynamically to only contain the users with actual
console access. I remember something such being discussed about the
reboot/halt in logout dialog thingy.


Sven Luther

Reply to: